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
|
|---|---|---|---|---|---|
Predicting function from sequence is an important goal in current biological research , and although , broad functional assignment is possible when a protein is assigned to a family , predicting functional specificity with accuracy is not straightforward . If function is provided by key structural properties and the relevant properties can be computed using the sequence as the starting point , it should in principle be possible to predict function in detail . The truncated hemoglobin family presents an interesting benchmark study due to their ubiquity , sequence diversity in the context of a conserved fold and the number of characterized members . Their functions are tightly related to O2 affinity and reactivity , as determined by the association and dissociation rate constants , both of which can be predicted and analyzed using in-silico based tools . In the present work we have applied a strategy , which combines homology modeling with molecular based energy calculations , to predict and analyze function of all known truncated hemoglobins in an evolutionary context . Our results show that truncated hemoglobins present conserved family features , but that its structure is flexible enough to allow the switch from high to low affinity in a few evolutionary steps . Most proteins display moderate to high oxygen affinities and multiple ligand migration paths , which , besides some minor trends , show heterogeneous distributions throughout the phylogenetic tree , again suggesting fast functional adaptation . Our data not only deepens our comprehension of the structural basis governing ligand affinity , but they also highlight some interesting functional evolutionary trends .
Predicting function from sequence and/or structure is one of the most important goals of structural biology , especially considering the increasing number of available sequences derived from multiple sequencing projects [1] . General function assignment or annotation , typically based on similarity with sequences with known biochemical function by means of BLAST [2] or generally done through the inclusion of a given protein to a family using HMMER profiling [3] , is common practice . However , determining specific functional properties or aspects , like substrate specificity/affinity or catalytic efficiency of a given protein with accuracy and detail at the residue level , is not straightforward . Even so , assuming that protein function is determined by protein structure and the particular physicochemical properties of its residues , encoded by the protein’s primary structure , it should in principle be possible to predict such functional properties in detail based on sequences and structures only . The globin superfamily of heme proteins offers a large , diverse and thoroughly studied set of proteins , whose function is tightly related to small gaseous non polar ligand ( mainly O2 but also NO , and CO ) [4–6] affinity and reactivity . It is known that hemoglobins ( Hbs ) can have functions other than oxygen storage and transport , including enzymatic and sensing functions [7] . Globins , as well as other heme proteins with high O2 affinity such as mycobacterial truncated hemoglobins , usually function as O2 ( and other reactive oxygen and nitrogen -RNOS- species ) redox related enzymes [8–11] . Moderate O2 affinity globins , like the mammalian monomeric myoglobin ( Mb ) and tetrameric hemoglobin , usually act as oxygen carrier storage proteins [12 , 13] , while low O2 affinity globins , such as soluble guanylate cyclase or the globin coupled sensors ( GCS ) , are NO , CO or redox sensors [14 , 15] . The truncated hemoglobins ( trHbs ) , also known as 2/2 Hbs , form one of the three lineages within the globin superfamily of proteins and is the only one present in all three superkingdoms of life [16 , 17] . They are distinguished by a simplified and unique two-over-two α-helical fold ( see Fig 1A ) and corresponding smaller size , i . e . 75–80% relative to three-over-three globins [17] . trHbs are organized in a number of structural blocks , as demonstrated in Fig 1A , in order to facilitate textual description and quick identification . Briefly , the protein is folded as two paired helix sandwich , composed of the BE and GH helices layers [18] . The well defined heme ligand binding site is composed of five structural positions , denominated B10 , CD1 , E7 , E11 and G8 [17] , and depicted in Fig 1C , which form distinctive features characterizing the trHbs . It should be noted that although it is tempting to analyze each residue contribution separately , these so-called distal residues act as a group in order to define the ligand reactivity . The generally accepted classification of trHbs in groups N , O and P ( also labeled I , II and III ) is founded on this characterization performed by Wittenberg et . al . [17] and later corroborated by a phylogenetic analysis [19] . Furthermore , trHbs present three topologically different ligand entry paths , i . e . long tunnel ( LT ) , short tunnel G8 ( STG8 ) and the E7 gate ( E7G , the main ligand entry and escape route in three-over-three globins such as Mb ) , as schematically shown in Fig 1B , through which ligands can migrate from the solvent towards the protein active site . Small ligand affinity is determined by the ratio between the association ( kon ) and dissociation ( koff ) rate constants , which characterize the corresponding processes . Association involves ligand migration from the solvent bulk to the active site through the protein gates and/or tunnel cavity systems , displacement of heme bound ligands ( either external or protein residues ) and finally Fe-ligand bond formation [20–23] . Dissociation , on the other hand , involves the disruption of the protein bound ligand interactions and its escape to the solvent [22 , 24] . During the last decade our group developed and applied several in-silico methods to address both the ligand association and dissociation processes with atomic resolution [5 , 25] . Briefly , using advanced sampling techniques we computed the free energy profiles ( FEP ) for ligand migration along the protein through internal tunnels that , together with the energy required to release the water molecules in the active site , account for the ligand association process . Also , molecular oxygen binding energy calculated by using hybrid quantum mechanics / molecular mechanics ( QM/MM ) based methods were successfully used to understand , correlate and determine the corresponding oxygen dissociation rate constants [5 , 26–30] . These studies showed that these in-silico analyses , performed in the context of available , structural and kinetic data , allow for a deep understanding of how particular globins control ligand affinity , paving the way for the development and application of a prediction protocol for the whole protein family . Using as a working hypothesis that it is possible to predict the function starting solely from sequence information through the determination of structure based chemical reactivity patterns related to oxygen reactivity , we have developed an in-silico protocol in order to predict several functional properties , including the association and dissociation rate constants , for ca . 1000 trHbs sequences . This novel approach is based on the combination of homology modeling and molecular based energy calculations and further complemented with a phylogenetic analysis , with the ultimate goal to predict and analyze trHb function in an evolutionary context . Meta-analysis of the results not only deepens our comprehension of the structural basis governing ligand affinity , but it also highlights some interesting evolutionary trends in trHb function .
The trHb family phylogeny has previously been described through the well known clustering into clusters N , O and P , also referred to as I , II and III , as derived from an analysis of 111 sequences available ca . 10 years ago [19] . Since then , many new sequences have become public , which allows for a more elaborate analysis . HMMER profiling was used to screen UnitProt and the PDB database and a non-redundant selection of the obtained sequences were aligned with Promals3D , which incorporates structural information . This multiple sequence alignment ( MSA ) was used to obtain a novel , bayesian tree ( Fig 2A ) that contains 1107 sequences ( see S1 Fig to see where each protein ends-up , labeled with its corresponding UniProt ID ) . The tree largely corroborates the current classification , with the N group harboring 24% , the O group 45% and the P 27% of the sequences . However , the topology also suggests the existence of a novel , small ( 4% ) , clade of sequences labeled as Q ( or IV ) ( see S2 , S3 and S4 Figs to see where each protein ends-up for N , O , P and Q groups , respectively , labeled with its corresponding UniProt ID ) . Recently , Vinogradov and coworkers revised the phylogenetic relationships of bacterial and archeaeal globins . Their results are similar as those presented here , showing the presence of the N , O and P clusters with similar percentages of sequences and no cluster assignation discrepancies [7] . The taxonomic distribution shows as expected a broad range of phyla within the eukarya , bacteria and also archaea super kingdoms for the N group , emerging few sequences corresponding to plant sequences ( 3% ) , whereas P and Q contain only bacterial sequences , and the O group hosts bacterial as well as 4 . 2% plant sequences . Although there are proteins displaying no more than 15% of sequence identity , overall structure is well conserved ( Fig 2B ) . This structural conservation makes it feasible to develop a protocol for a complete family characterization based on a sequence–structure–function relationship . To identify key residues that determine trHbs properties and subgroup characteristics we analyzed the information derived from the corresponding sequence logos , as well as Cluster and Specificity Determining Positions ( CDP and SDPs , respectively ) and Mutual Information ( MI ) analysis ( see Methods section for details ) . The results , presented in Fig 2C and 2D and S1 Table , yield a lot of information about the relevance of each structural position . Fig 2C , for example , shows the network of all positions with a cMI higher than 65 highlighting E7 and its direct MI-neighbours ( most of which are also SDPs ) , which compose active site , heme binding and key structural positions . Hence , SDP analysis suggests that active site and heme binding residues are the main driving force of functional diversification . As expected , data conclusively shows the strictly conserved HisF8 that coordinates the heme group , whose binding is also supported by the basic E10 , and E4 , EF6 , F4 and H16 residues building the heme hydrophobic environment . Two GG motifs , the first between the A and B helices and the second at the end of the E-helix ( starting at EF1 ) are conserved in groups N , O and Q but are highly variable in group P . A highly conserved AspB16 marks the end of B-helix and is important for the typical trHb fold . Concerning residues that allow group specific characterization , structural positions E7 , with a conserved His defining the E7 gate ( see below ) , and E15 , governed by a Trp in group P , allow to discriminate P from O and N , hence , should be considered P specific hallmarks . Group O shows a characteristic basic stretch ( His-Pro-Arg-Leu-Arg-X-Arg ) located in the EF loop and the first turn of the F helix . Key characteristics of new identified Q group are the above mentioned GG motifs , Trp or Phe at G8 and an almost 100% conserved HisE7 , as well as novel defining unique characteristics such as a strictly conserved Glu at E17 , an Arg-X-Arg motif close to the G8 position and a rather distinct , conserved C terminus including the highly conserved Ser at structural position H22 , all of which should be considered Q specific hallmarks . Active site , or distal , residues show intermediate global conservation and also contribute to the group clustering . Most conserved is the commonly found PheB9-TyrB10 key for determining ligand affinity . G8 has the typical Trp in O , P and Q . CD1 is predominantly a Phe , but the O group also shows Tyr or His . E7 and E11 show higher levels of variation ( Fig 2D ) . It should be noted that there are four additional structural positions with significant MI values , i . e . E17 , EF8 , G5 and G11 , far away from active site or tunnels topologies , which molecular function is currently unknown ( highlighted in red at Fig 2D ) . They could be related with trHb fold stability , protein-protein interactions and/or post-translational modifications . A final point of notice is related to the sequence length , since many proteins show extended N or C terminal segments . Available structural data suggest that although they may adopt secondary structure , it does not alter the global protein fold . In order to be able to predict ligand affinity from the primary sequence only , reliable calculations of the various variables involved needed to be developed . The employed strategy used to compute the association rate constants ( kon ) is based on the determination of the free energy profiles ( FEP ) for ligand entry and the use of a kinetic model to obtain the corresponding kon . The kinetic model is based on the works of Olson , Viappiani and colleagues [21 , 22 , 31 , 32] , which considers two basic processes: the partition of the ligand or the equilibrium between the solvent and the tunnel cavities ( or wells ) , and the migration from the tunnel to the active site . The first is determined mainly by the depth of the free energy wells , and thus the barrier to escape back to the solvent , while the second is determined by the barrier that the ligand needs to cross to move forward and reach the heme . Qualitatively , in this model faster rates are achieved combining a deep well that increases the effective ligand concentration inside the protein , and a small barrier to reach the heme . As mentioned previously , trHbs present three potential paths for ligand entry and exit , LT , STG8 and E7G ( see Fig 3 ) . The LT runs parallel to the trHbs fold longer axis ( along the H-helix ) and perpendicular to the heme plane . It exits the protein between helices Q and H . It was described in Mt-trHbN , and always shows the presence of three wells and two barriers . The key residues defining LT topology ( i . e . barrier height or well deepness ) are H5 , B2 , H9 , E15 , E11 and G8 . The STG8 is oriented perpendicular to the LT , runs parallel to the heme towards the G-helix , and exits the protein between helices G and H . Key residues determining its FEP are H9 , G8 and G9 . It was also first described in Mt-trHbN , and shows the presence of two wells and only one barrier . The third entry path , the E7G , is topologically equivalent to the ligand entry site in Mb . It was first described in Mt-trHbO and Cj-trHbP . It runs parallel to the heme plane in the opposite direction with respect to the STG8 . It presents 2 wells separated by one barrier , and the residues determining their characteristics are B10 , CD1 , E7 and E11 . As shown by the examples in Fig 3 , size , shape and hydrophobicity of the residues lining the tunnels determine both the barrier heights and well depth along the corresponding FEP . Usually , barriers in the 1–3kcalmol−1 range are considered small and correspond to -and will be referred as-open tunnels , while large barriers in the 10–20kcalmol−1 range result -and will be described as- blocked or closed tunnels . Well depths are computed relative to the bulk solvent and can be as large as 3kcalmol−1 resulting in 150 times enhanced effective ligand concentration , which thus increases the association rate accordingly as shown for human Mb [33] . To analyze the reliability of our proposed model we first computed the FEP profiles along each tunnel for all trHbs whose kon rates were determined experimentally , thus determining for each protein all three tunnel contributions to the ligand association process , which we will refer to as kLT , kSTG8 and kE7G . The linear combination of the three tunnel rates in each protein , finally results in the corresponding trHb tunnel dependent association rate ( ktunnels ) . It is important to note that ktunnels range is usually dominated by the most open tunnel . In other words , once a tunnel is open , resulting in a high rate ( 105M−1S−1 ) , having a second similarly opened tunnel , does not result in a significant increase in kon . Interestingly , and as analyzed in detail in Bustamante et . al . [34] ktunnels shows poor correlation with the experimentally determined kon ( R2 = 0 . 47 ) and rates are significantly overestimated . Analysis of wt versus mutant pairs , where tunnel topologies are not altered but nonetheless result in over ten times difference in kon values , combined with literature data strongly suggested that water displacement , from the distal pocket on top of the heme required to allow oxygen coordination , is the missing factor [20–22 , 35] . To account for this effect , the water stabilization free energy for each protein as estimated and characterized by the corresponding equilibrium constant KH2O ( presented in S2 Table for all possible trHbs ) . To demonstrate the roles played by ktunnels and water displacement , pairs ( or groups ) of proteins where one of the contributions varies while the other remains fairly constant , can be compared . For example , single and double distal GlnE11Val/Ala and TyrB10Leu/Phe mutants of Mt-trHbN can be compared wt Mt-trHbN ( Fig 4 ) . The substitutions do not alter tunnel topologies but both wt residues are capable of establishing hydrogen bonds with the water blocking the heme iron . Exchanging each of them for hydrophobic residues results in a significant ( almost 10 times ) increase in kon . Moreover , the double mutant , where there are no more water stabilizing interactions , results in an even larger increase in kon . To show the role played by the tunnel , we can look at the ValG8Phe mutant of Mt-trHbN , which has a 10 times smaller kon compared to the wt [36] having very similar distal site residues , and thus similar KH2O . However , the ligand needs to overcome higher barriers along the tunnel to reach the active site , mainly due to the size increase of G8 residue ( see Figs 4A and 3B ) , a fact that is reflected in smaller ktunnels . Another example , of the role played by the tunnel , is obtained when comparing wt forms of Mt-trHbN and Cj-trHbP , which show that the former has about 15 times larger kon ( Fig 4 ) . In these cases , the ligand enters through different tunnels , STG8 and E7G , respectively , and which as shown in Fig 3A and 3C , present significantly different barrier heights , explaining the observed trend . In summary , the above examples clearly show the relevance of both factors in determining the overall association rate . The resulting values ( kLT , kSTG8 , kE7G and KH2O ) computed for all possible trHb sequences are presented in S2 Table and will be analyzed later . Combining both KH2O and ktunnels using eq 4 ( see Methods ) , we were able to have a better estimate of the association rate , which shows good correlation ( R2 = 0 . 78 ) with experimental data ( Fig 4 ) . As shown by previous works from our group , oxygen dissociation is mainly controlled by the strength of distal interactions [5] . Here a similar approach as described above for ktunnels was followed . First , the oxygen binding energy ( ΔΔEO2 ) was calculated for a number of trHbs with resolved structure using a QM/MM scheme and compared to actual empirical data ( see Methods ) . The corresponding plot of the predicted vs experimental determined koff values for all available wt and mutant trHbs ( and some additional cases from our previous works ) is presented in Fig 5 . This shows that trHbs koff prediction is , on average , as good as the predictions for kon ( R2 = 0 . 79 ) . The computed koff values for all trHbs are presented in S2 Table . Global analysis of the obtained values , suggests that three functional groups can be identified . A first group corresponds to proteins displaying fast dissociation rates ( 100s−1 ) , usually due to a lack of ligand stabilization by distal residues . Proteins with moderate dissociation rates , with half-lives of the oxygenated species in the seconds timescale form a second group , where those with low or very low dissociation rates , which usually result in high -or very high- oxygen affinities reside in a third group . Strikingly , ca . 70% of all analyzed cases display a low koff , as that observed for the trHbs N , O and P from Mycobacterium tuberculosis ( Mt-trHbN ) , Thermobifida fusca ( Tf-trHbO ) and Campylobacter jejuni ( Cj-trHbP ) , 25% showing a predicted moderate dissociation rate like that observed for Pc-trHbN or human Mb with the remainder showing high or very high rates . The structural reasons underlying this observation is the invariable presence of at least one ( and many times two ) strong hydrogen bonds ( to the ligand ) forming residues , like TyrB10 ( present in 80% cases ) , TrpG8 ( 70% ) , His or Tyr at CD1 ( 20 and 15% , respectively ) and GlnE11 ( 21% ) . Moreover , in many cases , like the already characterized Cj-trHbP , there are several hydrogen bonds forming residues that act cooperatively and dynamically establishing a tight multiple hydrogen bond network with the bound ligand , resulting in a very low koff . It is important to note that the contribution of the FEP along the tunnels for the ligand escape process and thus koff is negligible in trHbs , and was thus not considered ( see Methods ) . As for kon , the functional and phylogenetic implication of the predicted dissociations rates will be discussed below in the context of the other computed properties . The present work’s ultimate goal is to make a potential functional prediction for each member of the trHb protein family in an evolutionary context , based solely on sequence information . To achieve this task , we constructed simplified models of most possible trHbs , in which the particular tunnel and/or distal residues were exchanged in a group dependent reference structure ( 1IDR for group N , 2BMM for group O and 2IG3 for groups P and Q , see Methods for details ) . Analysis of all possible tunnel residue combinations , using the MSA , shows 460 , 156 and 137 different residue combinations that define respectively the LT , STG8 and E7G characteristics . Selecting the most representative combinations while combining similar residues in the same group ( see Methods ) resulted in 41 , 36 and 17 different residue combinations that cover more than 87% of all possible trHbs . The other structural aspect to be considered , are the active site residues that interact with coordinated O2 and water by means of hydrogen bond interactions . The five topological positions ( B10 , CD1 , E7 , E11 and G8 ) in the active site of the trHbs work cooperatively to define ligand stabilization . Analysis of the MSA shows that there are 158 different combinations of these key residues , which can be trimmed down to 28 combinations that cover over 75% of the trHbs active sites . Once all possible combinations that define the tunnel and active site structural characteristics for the whole trHb family were determined , the corresponding models were built and used to compute the FEP for each possible residue combination in each of the three tunnels and the number of hydrogen bonds retaining the water at the active site that determine the kon and the QM/MM obtained ΔΔEO2 that determine koff ( see Methods ) . As a control , we also built this simplified models for all trHbs of which a complete structure is available and the results for the obtained parameters and kinetic rates are equivalent , suggesting that the approach is appropriate . The resulting values for kLT , kSTG8 , kE7G , KH2O , kon , ΔΔEO2 and koff for each determined residue combination and thus each analyzed trHb are presented in S2 Table . It is important to note that in our approximation trHbs with the same key position residue combination will display exactly the same rate constants . The calculated values should be considered a first estimation that according to the presented results is sufficiently accurate -typically well within one order of magnitude- to infer structure-function relationships . Clearly , the predicted values will differ from real values mostly since the characters are further modulated by minor aspects that are not considered in the calculation . A first look analysis at the distribution of association rate values for all computed trHbs ( S5 Fig ) shows that although a wide range of values is possible , most trHbs display values in the 105–108M−1s−1 range , consistent with a tunnel which accesibility is only hampered by a water molecule . There are also a significant number of proteins that display values up to 109M−1s−1 , which is caused by the absence of blocking water . Finally , a minor group of proteins with association rates in the 103–104M−1s−1 range exist , which corresponds to those proteins where tunnels are blocked and/or tightly bound water blocks the access to the heme . The distribution of dissociation rates is more homogenous , but with a predominance of values below 1 . The range extends from values as low as 104s−1 , which corresponds to proteins binding oxygen tightly with several hydrogen bonds , to 104s−1 for those proteins displaying highly hydrophobic distal pockets . The reliable prediction of both association and dissociation rate constants for all types of trHbs finally allows us to determine properly their oxygen affinity , which is usually expressed as p50 , the oxygen pressure that results in half the protein loaded with O2 . The results ( S6 Fig ) show that most trHbs display low ( or very low ) p50 values ( < 1mmHg ) , which would indicate that the protein is oxygenated even in microaerobic environments . These proteins usually display a moderate kon ( in the range where most values are found ) and large variations in koff , although always displaying values below 1s−1 . There is a second group with moderate p50 values ( 1-5mmHg ) , which possibly reflects these proteins are involved in oxygen transport . Which is characterized by the presence of both moderate kon and koff values ( between 1s−1 and 100s−1 ) . And finally , there are a few members with very high p50 values ( displaying mostly both large kon and koff values ) , which suggest they are unable to bind oxygen at all . Correlation analysis of p50 vs kon and koff , suggest that p50 is predominantly controlled by koff ( R2 = 0 . 60 ) , with kon having little impact ( R2 = 0 . 05 ) . To analyze how the different oxygen binding properties are related to the evolutionary processes that resulted in the functional diversification of the trHb family we decided to combine the above computed functional parameters for all trHbs in a phylogenetic context . To understand the resulting pattern we analyzed first , how phylogeny results in the hierarchical clustering of the trHb sequences ( at the group and subgroup levels ) , and second , what properties co-cluster within each clade . Fig 6 , thus shows the phylogenetic tree of the whole trHb family , together with a mapping of the O2 stabilization and the openness of each tunnel . A similar analysis was performed for the other computed parameters ( S7 Fig ) , yielding similar conclusions . The emerging picture not only allows to further characterize each group , but also to identify several subgroups ( mono- or paraphyletic ) which share several key properties related to their ligand binding properties , and thus their putative functions , as will be shown in the discussion . Oxygen stabilization ( assigned as high , moderate and low ) , for example , shows a clear subgroup distribution that points to the groupwise conservation of key residues that determine oxygen stabilization in particular clades . Tunnel openness also shows clear group preferences . The STG8 , for example , is only open in the N group . On the contrary , E7G , which is mostly blocked in group N -except for a small lineage denominated Na- , is always open for trHbs in group O as well as in P , albeit with minor exceptions . It is also interesting to note that , due to conservation of the overall fold , all the tunnels are present in all proteins . However , as shown for example by E7G in group N , it can be completely blocked ( displaying free energy barrier for ligand entry of over 10kcalmol−1 ) by the presence of key residues ( see S2 Table ) . Reasons underlying blockade of STG8 in trHbs from groups O , P and Q ( and also some members at the Ia clade ) , can be traced to the ubiquitous presence of Trp or Phe at the structural position G8 . Clade N is particular in that it has trHbs that have either high or low but no intermediate O2 affinities , caused mostly by very low koff . Low affinity is correlated with the hydrophobic Leu at structural positions E7 , E11 and G8 . As such , clade N can be divided in subclades Na , Nb , Nc and Nd , of which Nb and Nd are monophyletic clades with high affinity , correlated with polar Gln at E7 and E11 . A single mutation ( CTA → CAA or CTG → CAG ) might explain for the large change in koff . Nb differs from the other subclades in that it has one instead of two open tunnels . Clade O , with a fully opened E7G as defining characteristic , can be subclustered in four groups , Oa and Od being paraphyletic and Ob and Oc being monophyletic . It consists of a number of moderate and many low koff trHbs with open E7G and LT . The main difference between O members is the dissociation rate , which together with the CD1 identity , which is otherwise occupied by a His or a Phe/Tyr . Group P , which contains mostly proteins with an open E7G , can be divided into six subgroups , having three monophyletic ( a , d and f ) and three paraphyletic ( b , c and e ) , in general almost all subgroups present moderate kinetic constants , excluding Pb , with moderate but also very low koff . Finally , Group Q presents only one monophyletic subgroup characterized by moderate kon and very low koff values .
Our working hypothesis was that determination of the proper physicochemical characteristics as derived from protein structure would allow us to infer ( or predict ) key trHb functional properties ( uptake and release of O2 ) and associated parameters ( ktunnels , KH2O and ΔΔEO2 ) . The herein presented results show that we are able to predict both rates quite accurately , thus encouraging the performance of a complete analysis of all trHbs possible structures . The global analysis taught us that trHbs show , in general , moderate to very low oxygen dissociation rates , and thus moderate to high oxygen affinity , due to the presence of at least one and usually several hydrogen bond interactions between the ligand and the protein , most commonly provided by TyrB10 , TrpG8 , His or Tyr at CD1 and GlnE11 , as was also previously found experimentally for some particular cases [22 , 26 , 37–41] . Concerning the tunnels , our results show that size and hydrophobicity of residues lining the tunnels results in the presence of deep wells along the FEP ( or secondary docking sites ) which increase the rate , while they reduce the rate through the imposition of sterical free energy “barriers” . Most important , almost all trHbs have at least one “open” tunnel , and many have two . It is important to note , that our data also suggest , in agreement with previous experimental observations on directed mutants [42 , 43] , that the presence of more than one tunnel is redundant in terms of ligand association rate , since the ligand will reach the heme ( and wait there to bind ) through the tunnel presenting the easiest access . In this scenario , the question of what is the relation between the presence of multiple tunnels and the trHbs function , must go beyond simple determination of kon and involve also other aspects or possibilities ( see below ) . Finally , we also show that association rate is significantly influenced by the presence of water molecules on top of the heme that interact with the distal residues through hydrogen bond interactions . The tighter the water is bound , the lower the kon . Having determined and analyzed the ligand binding properties of all trHbs and their phylogenetic relationships , the question now arises as to whether it is possible to infer or predict a possible function for them . The question of globin , and thus trHbs function is a controversial issue , since even for the hallmark protein Mb several functions ( O2 storage , nitrite reductase , NO dioxygenase ) have been proposed and shown to be possible [10 , 22 , 44] . The problem arises due to the vast heme reactivity that allows it to fulfill different tasks under different conditions . However , not all tasks will be performed with the same efficiency due to the differential heme reactivity , thus some functions may seem more likely than others . Moreover , as mentioned in the introduction , the ubiquitous presence of molecular oxygen in the environment and the large variation observed in heme protein's affinity towards it ( in opposition to CO or NO that bind tightly to the heme almost independently of the protein environment ) , allows to draw some general lines based on the key parameters of O2 association and dissociation , computed here , even although , to the best of our knowledge , there is no single trHb whose function has been undoubtedly established and that are many trHbs remain poorly characterized beyond basic ligand kinetics . For some of them , which will be used here as leading cases , tentative but well based functional assignment is available . Finally , it is important to note that the predicted affinities apply only to “hypothetical” monomeric isolated trHbs in vitro . A such , when predicting possible trHb functions starting from the computed properties , we do not consider several issues like quaternary and cooperative effects; protein localization and interaction with other proteins or membranes; and the particular circumstances of each organism living ( e . g . aerobic/anaerobic , type of metabolism ) , which provide the proper context . Therefore , our predictions should be taken as a starting point or working hypothesis to further study each trHb function in a biological relevant context , in a similar manner as what is done with in vitro kinetic measurements . Possibly the most studied trHb is Mt-trHbN , paradigm of the N group trHbs . This protein likely function is to detoxify NO through its oxidation to nitrate by the oxy heme . To fulfill this task , a high oxygen stabilization is required , and the presence of multiple tunnels is likely an important factor [35 , 45–48] . Most of the Nd subgroup proteins share these properties , and thus NO detoxification seems a likely function . Interestingly , two others subgroups of group N ( Na and Nc ) show a larger koff and thus reduced oxygen affinity more similar to that of Pc-trHbN or Mb cases [22 , 49] . For these cases , as well as other trHbs sharing a large koff and presence of one or two open tunnels , a role involving oxygen storage or transport seems more likely , since a moderate kon and moderate to large koff is a prerequisite to allow efficient oxygen uptake and delivery . Another case , could be represented by Tf or Mt trHbO , paradigms for group O trHb , which have been proposed to work in relation with reactive oxygen species , in catalase-peroxidase like functions [50 , 51] . Key properties of these proteins to perform these tasks are the presence of a tight distal hydrogen bond network , revealed in a low or very low oxygen koff and the presence of an open E7G that provides shorter and more polar access to the heme than STG8 and LT and could thus be particularly suited for the entry/escape of polar or charged ligands such as superoxide . Also noteworthy is that heme proteins performing these tasks usually display polar aromatic ( Trp-Tyr-Arg-His ) residues in their active sites that can participate in redox reactions stabilizing free radical species . In correspondence with this idea , Wang and coworkers [11] recently showed that a trHb from Roseobacter denitrificans -which belongs to the Oc group- has peroxidase activity . Although they did not analyze ligand binding properties , the presence of TyrB10 , TyrE7 GlnE11 and TrpG8 suggest low koff and open E7G , consistent with our proposal . These functions thus emerge as likely candidates for many ( even most ) group O ( or II ) trHbs sharing the mentioned properties . Less is known concerning members of the P group , the best characterized member being Cj-trHbP . Although its function is not clear , it displays structural and ligand binding properties that reveal a tight hydrogen bond network and the presence of E7G ( like previously described trHbO ) . These properties however are not shared by all group members and high variability in terms of ligand interactions and tunnel openness is revealed , preventing a general prediction about their function . The reason that P trHbs forms a distinct clade is explained by strict conservation of HisE7 . Given the hierarchic clustering of the trHbs and taking into account their functional key characteristics ( kon , koff , p50 and tunnel openness ) we can now analyze trHb distribution in their hosting organisms . The 1107 trHbs genes belong to over 600 different species , with most of them ( 73% ) harboring only one type of trHb , 23% displaying two different trHbs , and a few organisms more than two . Analysis of the phylogeny shows that for those organisms displaying two types of trHbs , almost half of them have an O and N types , ca 40% an O and P types of trHbs , and only about 15% N and P types together . These results are similar as those observed previously by Vuletich et . al . [19] and seem to point out that O is the ancestral group . Based on previous description a rough functional assignment of trHb was performed by defining an NO/O2 multiligand chemistry type ( Mt-trHbN type ) , an oxygen transport type ( Mb-like type ) and a catalase-peroxidase functional type ( Tf-trHbO and Mt-trHbO type ) . Analysis of type related presence in each organism , shows that those species having only one trHb show predominantly a catalase-peroxidase functional type of protein ( 64% ) , followed by oxygen transport and NO/O2 multiligand chemistry types , both with similar populationsize ( 18% ) which is the expected distribution based on the relative abundance of each functional type . For those organisms having two trHbs combination again reflects expected distribution . Thus , the available data does not show any evidence of functional diversification for coexisting trHbs . We also looked for clustering of the three major types of trHbs . In group N , 57% are predicted to work in NO/O2 multiligand chemistry while interestingly the remaining 43% is predicted to be involved in oxygen transport . Catalase-peroxidase proposed function emerges as the likely candidate for most ( 86% ) of group O trHbs , all sharing the mentioned properties , with the remainder being shared similarly between other functional types . Also in group P most trHbs ( 78% ) share structural and ligand binding properties as those previously mentioned for a catalase-peroxidase like function . The remaining 22% being assigned as oxygen transport like due to their higher dissociation rate . Finally , all members of the newly identified group Q ( IV ) are assigned as catalase-peroxidase like . In any case , it is interesting to note that different functional types are found among the same phylogenetic group and thus care should be taken in assigning functional solely based on phylogeny . Taken together our results provide a rough evolutionary pattern of possible trHbs functions , in the sense that they were determined by the properties related to ligand reactivity , which are distributed , despite some general trends mentioned above , quite heterogeneously ( or randomly ) in the phylogenetic tree . This behavior could point to either functional plasticity or to high flexibility in terms of sequence-evolution to function relationships , thus resulting in multiple events of divergent and convergent evolution in terms of the studied properties , along a given evolutionary line . In other words the structural fold of trHbs seems flexible enough to allow the switch from a high affinity ( or multiple open tunnels ) structure to that of a low ( or one/no-tunnel at all ) structure in a few evolutionary steps , thus allowing for multiple rounds of reactivity/affinity and thus functional adaptation . The presence of multiple trHb paralogs in all kingdoms of life [17 , 52] which always appeared as a strange fact which lacked an explanation , clearly substantiates the above mentioned plasticity and evolution-to-functional flexibility and diversity . This work also represents a proof of concept for the hypothesis that states that it is possible to infer protein function in detail -beyond family assignment- starting solely from sequence information , through the determination of key structure related chemical reactivity properties . In this context future extension of the developed methodology to other protein families , like the more structurally diverse and functionally complex 3-over-3 globins can be expected .
The complete model used to determine the ligand association ( and dissociation ) rates using the above computed properties is thoroughly explained and validated elsewhere ( Bustamante et . al . [34] ) and will be presented here only briefly . The small ligand association involves two main processes , ligand migration from solvent bulk to the protein heme cavity through the tunnel cavity system , and formation of the Fe-O2 bond , which may involve the displacement of a water molecule from top of the heme . To estimate the tunnel contribution , with the information derived from the tunnel FEP , we used a generic kinetic model ( Scheme 1 ) that considers the presence of several secondary docking sites ( wells in the FEP ) and their associated barriers [22 , 43] . As in the case of the small ligand association process , the dissociation also involves two main processes , breaking the ligand stabilization network and further ligand migration from protein heme cavity to the bulk solvent , however only the former contributes significantly to the dissociation and is used in the model ( Bustamante et . al . [34] ) . Previously , we showed that QM/MM computed oxygen dissociation energy provides a good estimate of the thermal barrier for oxygen release ( and thus koff ) [5 , 70 , 81] . Note that the oxygen release process is a unimolecular reaction , so the proposed model is: koffcalc=e−ΔEO2RT ( 5 ) This kind of approach was successfully previously used to study NO dissociation from porphyrins [88 , 89] . However , if ΔEO2 values are used directly in eq 5 , koff calculated values are significantly underestimated and thus further corrections need to be performed . First , it is well known that computed oxygen dissociation energies from the heme are significantly overestimated due to the fact that a low ( singlet ) to high spin ( quintet ) spin transition is involved and DFT overestimates the energy of the spin gap , favoring low spin configurations [90] . Second , ΔEO2 values are computed for the optimized , i . e . best possible conformation at 0K , and kinetic values are computed at room temperature . Last but not least , due to errors intrinsic to DFT-based QM/MM methods , the computed energies are strongly dependent on the exchange-correlation functional and basis set . This can be partially considered and corrected by estimating the oxygen binding energy relative to that of a free heme , using eq 6 . koffcalc=e−ΔΔEO2RT ( 6 ) where ΔΔEO2 corresponds to the ΔEO2 ( oxygen binding energy computed as described above ) and the difference between ΔEheme , the calculated oxygen binding energy of an isolated imidazol bound heme in vacuum ( which is 22Kcalmol−1 ) and kofffreeheme value ( 104 s−1 ) [43 , 70] . The computed koff values for all possible combinations of active site residues are presented in S2 Table .
Tables containing all computed rates for all possible trHbs . Bayesian trees for the four main trHb groups and additional graphs for the analysis of rate constants are also available . | Globins are a superfamily of widely studied and diverse globular proteins whose function is tightly related to their oxygen affinity and reactivity . Two prominent members are the well-known tetrameric hemoglobin and the monomeric myoglobin , both involved in reversible oxygen storage and transport in mammals . Truncated hemoglobins form one of the three main monophyletic branches of this superfamily , presenting an interesting paradigm for structure-function prediction studies , as a result of their well-known and conserved fold . In the present work we started from the working hypothesis that states that “it is possible to predict the function starting solely from sequence information through the determination of structure based chemical reactivity patterns related to oxygen reactivity” and predicted oxygen reactivity for over 1000 truncated hemoglobins and analyzed the results in an evolutionary context . Our results taught us many interesting and novel features of these proteins , underscoring flexibility and adaptability of the globin fold . The work also shows that it is possible to characterize protein function in greater detail if specific sequence-to-function bridges are built upon a solid structural basis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
]
| [
"sequencing",
"techniques",
"heme",
"taxonomy",
"oxygen",
"split-decomposition",
"method",
"multiple",
"alignment",
"calculation",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"molecular",
"biology",
"techniques",
"chemical",
"dissociation",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"computer",
"and",
"information",
"sciences",
"sequence",
"alignment",
"proteins",
"biological",
"databases",
"chemistry",
"evolutionary",
"systematics",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"biochemistry",
"sequence",
"databases",
"post-translational",
"modification",
"computational",
"techniques",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"evolutionary",
"biology",
"chemical",
"elements"
]
| 2016 | Evolutionary and Functional Relationships in the Truncated Hemoglobin Family |
Human T-cell leukemia virus type 1 ( HTLV-1 ) causes both a neoplastic disease and inflammatory diseases , including HTLV-1-associated myelopathy/tropical spastic paraparesis ( HAM/TSP ) . The HTLV-1 basic leucine zipper factor ( HBZ ) gene is encoded in the minus strand of the proviral DNA and is constitutively expressed in infected cells and ATL cells . HBZ increases the number of regulatory T ( Treg ) cells by inducing the Foxp3 gene transcription . Recent studies have revealed that some CD4+Foxp3+ T cells are not terminally differentiated but have a plasticity to convert to other T-cell subsets . Induced Treg ( iTreg ) cells tend to lose Foxp3 expression , and may acquire an effector phenotype accompanied by the production of inflammatory cytokines , such as interferon-γ ( IFN-γ ) . In this study , we analyzed a pathogenic mechanism of chronic inflammation related with HTLV-1 infection via focusing on HBZ and Foxp3 . Infiltration of lymphocytes was observed in the skin , lung and intestine of HBZ-Tg mice . As mechanisms , adhesion and migration of HBZ-expressing CD4+ T cells were enhanced in these mice . Foxp3−CD4+ T cells produced higher amounts of IFN-γ compared to those from non-Tg mice . Expression of Helios was reduced in Treg cells from HBZ-Tg mice and HAM/TSP patients , indicating that iTreg cells are predominant . Consistent with this finding , the conserved non-coding sequence 2 region of the Foxp3 gene was hypermethylated in Treg cells of HBZ-Tg mice , which is a characteristic of iTreg cells . Furthermore , Treg cells in the spleen of HBZ-transgenic mice tended to lose Foxp3 expression and produced an excessive amount of IFN-γ , while Foxp3 expression was stable in natural Treg cells of the thymus . HBZ enhances the generation of iTreg cells , which likely convert to Foxp3−T cells producing IFN-γ . The HBZ-mediated proinflammatory phenotype of CD4+ T cells is implicated in the pathogenesis of HTLV-1-associated inflammation .
Human T-cell leukemia virus type 1 ( HTLV-1 ) is known to be the causal agent of a neoplastic disease of CD4+ T cells , adult T-cell leukemia ( ATL ) [1] . In addition , this virus perturbs the host immune system , causing inflammatory diseases and immunodeficiency . Inflammatory diseases associated with HTLV-1 includeHTLV-1-associated myelopathy/tropical spastic paraparesis ( HAM/TSP ) [2] , [3] , uveitis [4] , [5] , alveolitis [6] , infective dermatitis [7] and myositis [8] . Increased expression of inflammatory cytokines and immune response to the Tax antigen has been proposed as mechanisms of these inflammatory diseases [9] . However , the detailed mechanisms of inflammation remain elusive . The HTLV-1 bZIP factor ( HBZ ) gene is encoded in the minus strand of the provirus and consistently expressed in ATL cases and HTLV-1-infected individuals [10] . In vitro and in vivo experiments have shown that the HBZ gene promotes the proliferation of T cells and increases their number [10] , [11] . Recently , we reported that HBZ transgenic ( HBZ-Tg ) mice develop both T-cell lymphomas and inflammatory diseases [12] . In HBZ-Tg mice , we found that the number of CD4+ T cells expressing Foxp3 , a master molecule for regulatory T ( Treg ) cells , was remarkably increased . HBZ induces transcription of the Foxp3 gene via interaction with Smad2/3 and a co-activator , p300 , resulting in an increased number of Foxp3+ T cells [13] . Concurrently , HBZ interacts with Foxp3 and decreases the immune suppressive function [12] . This interaction could be a mechanism of the inflammatory phenotype observed in HBZ-Tg mice . However , detailed mechanisms to induce inflammation by HBZ remain unsolved . Treg cells suppress excessive immune responses , and control the homeostasis of the immune system [14] . Foxp3 is considered a marker of Treg cells , yet several lines of evidence have shown that there is heterogeneity within Foxp3+cells [15] . Natural Treg ( nTreg ) cells are generated in the thymus while induced Treg ( iTreg ) cells are induced in the peripheral lymphoid organs . It has been reported that Treg cells that have lostFoxp3 expression ( exFoxp3 T cells ) produce interferon-γ ( IFN-γ ) , indicating thatFoxp3+ Treg cells are not terminally differentiated cells but susceptible to conversion into effector T cells according to their environment [16] . Recently , Miyao et al . have reported that Foxp3+ T cells induced by activation exhibit transient Foxp3 expression , and become exFoxp3 T cells [17] . Even though the plasticity of Treg cells remains controversial [18] , these reports suggest that Foxp3+ T cells possess not only suppressive function but also proinflammatory attributes . In this study , we found that iTreg cells increased in HBZ-Tg mice and that Treg cells of HBZ-Tg mice tend to lose Foxp3 expression , leading to increased IFN-γ-expressing proinflammatory cells . Cell adhesion and migration are enhanced in CD4+ T cells of HBZ-Tg mice . Thus , these HBZ-mediated abnormalities of CD4 T cells play critical roles in inflammatory diseases caused by HTLV-1 .
We have reported that HBZ-Tg mice develop both T-cell lymphoma and inflammatory diseases including dermatitis and alveolitis [12] . To further study the inflammatory changes affecting HBZ-Tg mice , we analyzed various tissues and organs in detail . In HBZ-Tg mice , moderate lymphoid cell infiltration was detected in the peri-bronchial space of the lung ( Figure 1A ) , the peri-follicular area of the skin ( Figure 1B ) , the mucosa of the small intestine ( Figure 1C ) , and the mucosa of the colon ( Figure 1D ) . Meanwhile , there was no obvious evidence of inflammation in liver , kidney or spinal cord . In non-Tg littermates , infiltration of lymphoid cells was not observed in skin , lung or intestine . These findings suggest the inflammatory involvement of multiple tissues and organs in HBZ-Tg mice . Infiltration of lymphocytes into various tissues suggests that the lymphocytes of HBZ-Tg mice have increased adhesive ability . We first studied the expression of LFA-1 , which is a heterodimer of CD11a and CD18 . As shown in Figure 2A , both CD11a and CD18 were upregulated on HBZ-Tg CD4+ T cells of spleen , lung and lymph nodes compared with CD4+ T cells from non-Tg mice . In addition , the expression of CD103 ( alpha E integrin ) on HBZ-Tg CD4+ T cells was also higher than that on non-Tg CD4+ T cells . These findings suggest an increased adhesive capability of CD4+ T cells in HBZ-Tg mice . Immunohistochemical analyses of lung and intestine of HBZ-Tg mice confirmed increased expression of these molecules , particularly CD18 ( Figure 2B , C ) . Expression of CD11a , CD18 and CD103 was also studied in HAM/TSP patients . In addition to healthy donors , we analyzed expression of these molecules on HTLV-1 infected cells that are identified using anti-Tax antibody . As shown in Figure 2D , CD11a and CD18 expression of CD4+Tax+ T cells was upregulated compared with CD4+ T cells from healthy donors and CD4+Tax− T cells of HAM/TSP patients while expression of CD103 was not different among these cells . These results show that enhanced expression of LFA-1 is also observed in HTLV-1 infected cells in HAM/TSP patients . We next investigated adhesion of CD4+ T cells to ICAM-1 , since ICAM-1 is critical for lymphocyte migration and adhesion to vascular epithelial cells in an inflammatory lesion . We isolated CD4+ T cells from non-Tg or HBZ-Tg splenocytes , placed them on ICAM-1-coated 96-well plates , and evaluated cell adhesion activity to ICAM-1 . CD4+ T cells from HBZ-Tg mice showed increased adhesion in the absence of stimulation , while no difference was found when cells were stimulated by anti-CD3 antibody ( Figure 3A ) . Furthermore , we evaluated the migration activity of CD4+ T cells on ICAM-1-coated plates . To induce cell migration , we stimulated CD4+ T cells with CCL22 as reported previously [19] . Cell migration of HBZ-Tg CD4+ T cells was also increased compared with migration of non-Tg CD4+ T cells ( Figure 3B ) . These results demonstrate an infiltrative phenotype of CD4+ T cells in HBZ-Tg mice . Infiltration of LFA-1 expressing T cells into various tissues suggests that ICAM-1 expression is enhanced . Indeed , expression of ICAM-1 was increased in intestine of HBZ-Tg mice ( Figure 2B ) . Enhanced migration of CD4+ T cells suggests involvement of chemokine ( s ) -chemokine receptor for HBZ-Tg mice . We analyzed expression of chemokine receptors on CD4+ T cells of HBZ-Tg mice . As shown in Figure 3C , CXCR3 expression of CD4+ splenocytes was increased while expression of CCR5 and CCR7 were not different compared with control mice ( Figure S1 ) . CXCR3 expression of CD4+ T cells was upregulated in both lung and lymph node ( Figure 3C ) . Although the ligands for CXCR3 , CXCL9 and CXCL10 , were not increased in the sera of HBZ-Tg mice ( Figure S1 ) , CXCR3 might be implicated in infiltration of CD4+ T cells . To elucidate the mechanism of the pro-inflammatory phenotype observed in HBZ-Tg mice , we investigated cytokine production in CD4+ T cells of the spleen . After stimulation by PMA/ionomycin , production of IFN-γ was increased in CD4+ T cells while that of TNF-α was suppressed ( Figure 4A ) . There were no significant differences between HBZ-Tg mice and non-Tg mice in IL-2 , IL-4 and IL-17 production by CD4+ T cells . We have reported that the number of Foxp3+CD4+ Treg cells is increased in HBZ-Tg mice . Therefore , we simultaneously stained both intracellular cytokines and Foxp3 to distinguish the cytokine production of CD4+Foxp3− T cells from that of CD4+Foxp3+ T cells . Production of TNF-α , IL-17 and IL-2 was slightly increased in CD4+Foxp3+ T cells of HBZ-Tg mice ( Figure 4B , C ) . Since Foxp3 suppresses production of cytokines [19] , and HBZ impairs function of Foxp3 [12] , HBZ-mediated impairment of Foxp3 function might be a mechanism of this increased expression of these cytokines . However , TNF-α production was suppressed in CD4+ Foxp3− T cells and total CD4+ T cells ( Figure 4A , C ) . In particular , IFN-γ production of splenic CD4+Foxp3− T cells from HBZ-Tg mice was remarkably increased compared with those from non-Tg mice ( Figure 4B ) . We also studied IFN-γ production in CD4+ T cells of PBMCs and lung-infiltrating lymphocytes . The production of IFN-γ was remarkably increased in PBMC and lung from HBZ-Tg mice ( Figure 4D ) . Taken together , these results suggest that increased IFN-γ production , especially in CD4+Foxp3− T cells , is related to the chronic inflammation observed in HBZ-Tg mice . Immunohistochemical analyses also showed that IFN-γ production was increased in both lung and intestine of HBZ-Tg mice ( Figure 2B , C ) . We have reported that HBZ enhances the transcription of the Foxp3 gene in cooperation with TGF-ß , leading to an increased number of Treg cells in vivo [12] , [13] . Two types of Treg cells have been reported: natural Treg ( nTreg ) cells and induced Treg ( iTreg ) cells in CD4+Foxp3+ cells . The expression of Helios , a member of the Ikaros family of transcription factors , is considered a marker of nTreg cells [20] . To determine which Treg cell population is increased in HBZ-Tg mice , we analyzed the expression of Helios . Expression of Helios in CD4+Foxp3+ T cells in HBZ-Tg mice was lower than that in non-Tg mice ( Figure 5A , C ) , suggesting that the number of iTreg cells is increased in HBZ-Tg mice . A higher proportion of CD4+Foxp3+Helioslow cells were found in the lungs of HBZ-Tg mice ( Figure S2 ) . Next , we analyzed the expression of Helios in Treg cells from HAM/TSP patients . As shown in Figure 5 B and D , Helios expression of Treg cells in HAM/TSP patients was lower than that of Treg cells in healthy controls . We also analyzed Helios expression in Foxp3+ T ( nTreg ) cells of the thymus . The level of Helios expression in nTreg cells in HBZ-Tg mice was equivalent to that of non-Tg mice ( Figure S3 ) . These data collectively suggest that the iTreg cell population is increased not only in HBZ-Tg mice , but also in HAM/TSP patients . Recent studies have reported that Helios expression is not always associated with nTreg cells [21]–[23] . A previous study reported that conserved non-coding DNA sequence ( CNS ) elements in the Foxp3 locus play an important role in the induction and maintenance of Foxp3 gene expression [24] . Among these elements , CNS2 , methylated in iTreg cells , was suggested to be responsible for the lack of stable expression of Foxp3 in these cells [24] . This region is not methylated in Helios- nTreg cells , indicating that unmethylation of this region is a suitable marker of nTreg cells [21] . Therefore , we sorted the Treg fraction from HBZ-Tg or non-Tg mice splenocytes , extracted genomic DNA , and determined the DNA methylation status in the CNS2 region of the Foxp3 gene . The results revealed that in HBZ-Tg CD4+Foxp3+ T cells , the CNS2 region had a higher methylation status than in non-Tg CD4+Foxp3+ cells ( Figure 6 ) , indicating that the increase in CD4+Foxp3+ cells in HBZ-Tg mice indeed mostly consists of iTreg cells . Recent studies have revealed that CD4+Foxp3+ T cells are not terminally differentiated but have the plasticity to convert to other T cell subsets [25] . When Treg cells lose the expression of Foxp3 ( exFoxp3 T cells ) , such cells produce pro-inflammatory cytokines [16] . It has been reported that Foxp3 expression in nTreg cells is stable but that it is not in iTreg cells [15] . These findings suggest that in HBZ-Tg mice , which have greater numbers of iTreg cells as shown in this study , Foxp3 expression in these cells tends to diminish , letting these cells acquire an effector phenotype associated with the production of pro-inflammatory cytokines such as IFN-γ . To investigate this possibility , we sorted Treg cells from the spleens of HBZ-Tg or non-Tg mice based on their expression of CD4 , CD25 and GITR; cultured them for 7 days; and analyzed Foxp3 expression by flow cytometry . After 7 days in culture , the percentage of Foxp3+ T cells diminished remarkably in HBZ-Tg mice compared with non-Tg mice ( Figure 7A , B ) . We investigated the production of IFN-γ at this point , and found that it was increased in Foxp3− T cells from HBZ-Tg mice compared with those from non-Tg mice ( Figure 7C ) . In sharp contrast to this finding , Foxp3 expression of nTreg cells did not change in CD4+ thymocytes of HBZ-Tg mice ( Figure 7D ) . Collectively , these data indicate thatFoxp3 expression in nTreg cells is stable in HBZ-Tg mice , while most of the Treg cells in the periphery are iTreg cells . The enhanced generation of exFoxp3 T cells in the periphery is a possible mechanism of the increase in IFN-γ -producing Foxp3− T cells in HBZ-Tg mice . We reported that HBZ induced the Foxp3 gene transcription via interaction with activation of TGF-β/Smad pathway [13] . Reduced expression of Foxp3 in HBZ-Tg CD4+Foxp3− T cells might be caused by low HBZ expression in that cell population . To investigate this possibility , we analyzed the relationship between HBZ and Foxp3 expression in CD4+ T cells of HBZ-Tg mice . We isolated CD4+CD25+GITRhigh T cells as Foxp3+ T cells , and CD4+CD25−GITRlow T cells as Foxp3− T cells from HBZ-Tg mice . Although Foxp3+ T cells are contaminated in CD4+CD25−GITRlow T cells , level of the Foxp3 gene transcript was much higher in CD4+CD25highGITRhigh T cells ( Figure S4 ) . However , level of HBZ transcript was no different among these cells , indicating that level of HBZ expression is not associated with reduced Foxp3 expression .
HTLV-1 is a unique human retrovirus with respect to its pathogenesis , since it causes not only a neoplastic disorder , but also various inflammatory diseases . For most viruses , tissue-damaging inflammation associated with chronic viral infection is generally triggered by the immune response against infected cells , which involves both antigen specific and non-specific T cells that produce pro-inflammatory cytokines , chemokines , and other chemical mediators that promote tissue inflammation [26] . However , this study shows that HTLV-1 can induce inflammation by a different mechanism that does not involve an immune response against infected cells , but instead , involves deregulation of CD4+ T-cell differentiation mediated by HBZ . Since transgenic expression of HBZ does not induce an immune response to HBZ protein itself , the inflammation observed in this study is attributed to an intrinsic property of HBZ-expressing cells . Studies of the pathogenesis of inflammatory diseases related to HTLV-1 are usually focused on HAM/TSP , since it is the most common inflammatory disease caused by this virus [9] . Two different mechanisms of HAM/TSP pathogenesis have been reported: one mechanism involves the immune response to viral antigens , and another mechanism implicates the proinflammatory attributes of HTLV-1-infected cells themselves . Previous studies reported a strong immune response to Tax in HTLV-1-infected individuals [9] , [27] . In lesions of the spinal cord , CD4+ T cells expressing viral gene transcripts were identified by in situ hybridization [28] . The presence of CTLs targeting Tax in cerebrospinal fluid and lesions in the spinal cord suggest an important role of the immune response and the cytokines produced by CTLs in the pathogenesis of HAM/TSP by HTLV-1 [29] . Those studies showed the involvement of the immune response to Tax in the pathogenesis of HAM/TSP . In addition , cell-autonomous production of proinflammatory cytokines by HTLV-1-infected cells has been reported . HTLV-1-transformed cells produce a variety of cytokines , including IFN-γ , IL-6 , TGF-ß , and IL-1α [30] . It was speculated that Tax was responsible for the enhanced production of these cytokines . In this study , we have shown a new role of HBZ in inflammatory diseases . CTLs against HBZ have been reported in HTLV-1 carriers and HAM/TSP patients; this immune response might be involved in inflammation caused by HTLV-1 [31] . However , an immune response to HBZ does not occur in HBZ-Tg mice , indicating that the proinflammatory phenotype of HBZ expressing T cells is sufficient to cause the inflammation . Does HBZ induce IFN-γ production in CD4+ T cells ? HBZ and Tax have contradictory effects on many pathways . For example , Tax activates both the canonical and non-canonical NF-κB pathways , while HBZ suppresses the canonical pathway [32] , [33] . Conversely , HBZ activates TGF-ß/Smad pathway , while Tax inhibits it [13] , [34] , [35] . Tax activates the IFN-γ gene promoter , whereas HBZ suppresses the transcription of the IFN-γ gene through inhibition of AP-1 and NFAT , which are critical for IFN-γ gene transcription [36] . These findings collectively suggest that the enhanced production of IFN-γ is not due to a direct effect of HBZ , but may be attributed to the increased presence exFoxp3 T cells triggered by HBZ as shown in this study . Recent studies reported that exFoxp3 T cells produce higher amount of IFN-γ [17] , [37] . This indicates that increased production of IFN-γ in exFoxp3 T cells surpasses the suppressive function by HBZ . In this study , HBZ inhibited the production of TNF-α as we reported [36] , indicating that enhanced production is specific to IFN-γ . However , it remains unknown how the production of IFN-γ is enhanced in exFoxp3 T cells . We have shown that the Foxp3+ T cells of HBZ-Tg mice tend to lose Foxp3 expression and change into IFN-γ-producing proinflammatory cells . This observation makes sense in the light of several other studies on Treg cells . It was reported thatFoxp3+ T cells convert to Foxp3− T cells [37]–[39] . Recently , Miyao et al . reported that Foxp3 expression of peripheral T cells induced by activation is promiscuous and unstable , leading to conversion to exFoxp3 T cells [17] . Peripheral induced Foxp3+ T cells show lower expression of CD25 and Helios , which corresponds to the phenotype we observed in the Foxp3+ T cells of HBZ-Tg mice . Thus it is likely that HBZ induces unstable Foxp3 expression and generates iTreg cells , which then convert to exFoxp3 T cells with enhanced production of IFN-γ as shown in this study . It has recently been reported that CD4+CD25+CCR4+ T cells in HAM/TSP patients were producing extraordinarily high levels of IFN-γ , when compared to cells of healthy donors . These findings are consistent with those of this study . Importantly , the frequency of these IFN-γ-producing CD4+CD25+CCR4+Foxp3− T cells was increased and found to be correlated with disease severity in HAM/TSP patients [40] . In addition , it has been reported that HBZ expression is correlated with the severity of HAM/TSP [41] . Thus , the presence of abnormal HBZ-induced IFN-γ-producing cells is a plausible mechanism that leads to inflammation in HAM/TSP patients . FOXP3 expression is detected in two thirds of ATL cases , suggesting that ATL cells originate from Treg cells in these cases [42] , [43] . Human FOXP3+ T cells have been divided into three subgroups based on their functions and surface makers: resting Treg cells ( rTreg ) , activated Treg ( aTreg ) cells , and FOXP3lownon-suppressive T cells [44] . Recently , we reported that HTLV-1 infection is frequently detected in Treg cells , which include FOXP3low non-suppressive T cells and FOXP3high activated Treg cells , and concordantly , some ATL cells also belong to the population of FOXP3low non-suppressive T cells [44] , [45] . This suggests that HTLV-1 increases the population of aTreg and FOXP3low non-suppressive T cells and induces leukemia/lymphoma of these cells . It is thought that most of nTreg are resting and activated Treg cells and iTreg cells contain both aTreg cells and Foxp3low non-suppressive T cells in human . The CNS2 region in the Foxp3 locus is highly methylated in FOXP3low non-suppressive T cells [44] , like we report for the iTreg cells of HBZ-Tg mice . It is likely that a fraction of FOXP3low non-suppressive T cells lose FOXP3 expression and change to FOXP3− proinflammatory T cells as reported in HAM/TSP patients [40] , suggesting that the finding of this study is indeed the case in HTLV-1 infection . It has been widely believed that nTreg cells represent a highly stable lineage in which few cells lose Foxp3 expression under normal homeostatic conditions [46] . In contrast , small subsets of CD25−Foxp3+ Treg cells have recently been reported to be unstable and to rapidly lose Foxp3 expression after transfer into a lymphopenic host [16] . The CNS2 sequence is methylated in iTreg cells [24] . Consistent with this finding , CNS2 was heavily methylated in Treg cells of HBZ-Tg mice , indicating that Treg cells in HBZ-Tg mice largely belong to the iTreg cell subset . Foxp3 expression of CD4+ thymocytes in HBZ-Tg mice did not decrease after in vitro culture , a fact which shows that loss of Foxp3 expression is not a direct effect of HBZ , but is due to the increased number of iTreg cells converting to exFoxp3 cells . Recently , it was reported that Foxp3+ T cells without suppressive function convert to exFoxp3 T cells [17] . We recently reported that HBZ enhances Foxp3 gene transcription by activating the TGF-ß/Smad pathway [13] . Collectively , it is likely that HBZ increases Foxp3+ T cells in HBZ-Tg mice and most of Foxp3+ T cells are iTreg and/or non-suppressive Foxp3+ T cells . Foxp3 expression in HBZ-Tg mice is unstable as shown in this study , and such cells easily convert to exFoxp3 T cells , which produce excess amounts of IFN-γ , leading to inflammation . Helios expression has been reported to be high in nTreg cells , and low in iTreg cells [20] . This study showed that Helios expression in CD4+Foxp3+ cells of HBZ-Tg mice was low although it was higher than control iTreg cells . Recently , it has been reported that stimulation enhances Helios expression of iTreg cells , which might account for increased Helios expression in CD4+Foxp3+ cells of HBZ-Tg mice compared with control iTreg cells [22] . In particular , inflammation caused by HBZ expression might increase Helios expression of iTreg cells of HBZ-Tg mice . In addition , it has been reported that Helios is not expressed in a part of nTreg cells and its expression is induced in iTreg cells , indicating that only Helios expression cannot discriminate nTreg cells from iTreg cells [21]–[23] . However , CNS2 is not methylated in Helios− nTreg cells , which shows that the methylation status of CNS2 is critical [21] . In this study , analysis of DNA methylation of CNS2 confirms that most of CD4+Foxp3+ cells in HBZ-Tg mice are iTreg cells . Importantly , the similar pattern of Heilos expression was observed in HAM/TSP patients . The present study has demonstrated that HBZ-Tg mice develop inflammation in the intestines , skin and lungs . These tissues are always exposed to extrinsic antigens and commensal microbes , where Treg cells are critical for maintaining the homeostasis of the host immune system . In addition to the increased production of IFN-γ by HBZ-expressing cells , it is likely that the cell adhesion attributes of these cells also play a role in their pro-inflammatory phenotype . Treg cells express a variety of molecules that are important for cell adhesion , including LFA-1 , CCR4 , and CD103 [12] . We have shown that these molecules are also present on HBZ-expressing CD4+ T cells . In this study , we showed that HBZ increases the number of iTreg cells , which subsequently convert into exFoxp3 T cells . The proinflammatory phenotype of HBZ-expressing T cells indicates that HBZ plays an important role in the inflammatory diseases caused by HTLV-1 . In conclusion , HBZ-Tg mice developed chronic inflammation accompanied with hyper IFN-γ production , which is consistent with the findings in HAM/TSP patients . CD4+Foxp3+ T cells , especially iTreg cells , were increased in HBZ-Tg mice . The expression of Foxp3 was not stable and tended to be lost , which resulted in the enhanced generation of exFoxp3 cells producing IFN-γ . This could be a mechanism for the development of chronic inflammation in HBZ-Tg mice and HTLV-1-infected individuals .
Transgenic mice expressing HBZ under the murine CD4 promoter have been previously described [12] . Genotypes were determined by means of PCR on mouse ear genomic DNA . All the mice were used at 10–20 weeks of age . Animal experimentation was performed in strict accordance with the Japanese animal welfare bodies ( Law No . 105 dated 19 October 1973 modified on 2 June 2006 ) , and the Regulation on Animal Experimentation at Kyoto University . The protocol was approved by the Institutional Animal Research Committee of Kyoto University ( permit number: D13-02 ) . All efforts were made to minimize suffering . A total of 10 HAM/TSP patients and 10 healthy donors participated in this study . Written informed consents were obtained from all the subjects in accordance with the Declaration of Helsinki as part of a clinical protocol reviewed and approved by the Institutional Ethics Committee of Kyoto University ( approval number: 844 ) . Blood samples were collected from the subjects and peripheral blood mononuclear cells ( PBMC ) were isolated by Ficoll-Paque Plus ( GE Healthcare Bio-Sciences ) density gradient centrifugation . Production of recombinant mouse ICAM-1 was performed as described previously [47] . A 96-well plate was coated with 100 µl/well of 0 . 25 µg/ml mouse mICAM-1-Ig ( R&D Systems ) at 4°C overnight , followed by blocking with 1% BSA for 30 min . Mouse CD4+ cells were labeled with 2′ , 7′-bis- ( 2-carboxyethyl ) -5- ( and-6 ) carboxyfluorescein ( Molecular Probes , Inc . ) , suspended in RPMI 1640 containing 10 mM HEPES ( pH 7 . 4 ) and 10% FBS , transferred into the coated wells at 5×104 cells/well and then incubated at 37°C for 30 min . Non-adherent cells were removed by aspiration . Input and bound cells were quantitated in the 96-well plate using a fluorescence concentration analyzer ( IDEXX Corp . ) . Random cell migration was recorded at 37°C with a culture dish system for live-cell microscopy ( DT culture dish system; Bioptechs ) . Thermoglass-based dishes ( Bioptechs ) were coated with 0 . 1 µg/ml mouse ICAM-1 . CD4+ mouse splenocytes were loaded in the ICAM-1-coated dish , and the dish was mounted on an inverted confocal laser microscope ( model LSM510 , Carl Zeiss MicroImaging , Inc . ) Phase-contrast images were taken every 15 s for 10 min . The cells were traced and velocity was calculated using ImageProR Plus software ( Media Cybernetics ) . Single-cell suspensions of mouse spleen , lung or PBMC or human PBMC were made in RPMI 1640 medium supplemented with 10% FBS . To detect Tax , CD8+ cells were depleted from human PBMC using the BD IMAG cell separation system with the anti-human CD8 Particles-DM ( BD Pharmingen ) according to the manufacturer's directions and then the cells were cultured for 6 hours . Surface antigen expression was analyzed by staining with the following antibodies: anti-mouse CD4 ( RM4-5 ) , CD11a ( 2D7 ) , CD18 ( C71/16 ) or CD103 ( M290 ) ( all purchased from BD Pharmingen ) or anti-human CD4 ( RPA-T4 ) , CD11a ( HI111 ) , CXCR3 ( G025H7 ) ( all purchased from BioLegend ) , CD18 ( 6 . 7 ) , CD103 ( Ber-ACT8 ) ( all purchased from BD Pharmingen ) . For intracellular cytokine staining , cells were pre-stimulated with 20 ng/ml phorbolmyristate acetate ( PMA , NacalaiTesque ) , 1 µM ionomycin ( NacalaiTesque ) and Golgi plug ( BD Pharmingen ) for 4 h prior to surface antigen staining . After this stimulation period , cells were fixed and permeabilized with Fixation/Permeabilization working solution ( eBioscience ) for 30 min on ice and incubated with antibodies specific for the following cytokines: IFN-γ ( XMG 1 . 2 ) , IL-17 ( TC11-18H10 ) , IL-2 ( JES6-5H4 ) ( all BD Pharmingen ) , TNF-α ( MP6-XT22 , eBioscience ) and IL-4 ( 11B11 , eBioscience ) . Intracellular expression of mouse Foxp3 ( FJK-16s , eBioscience ) , human FoxP3 ( PCH101 , eBioscience ) , Tax ( MI73 ) , human IFN-γ ( 4SB3 , BD Pharmingen ) and Helios ( 22F6 , BioLegend ) was detected following the protocol for cytokine staining . Dead cells were detected by pre-staining the cells with the Live/dead fixable dead cell staining kit ( Invitrogen ) . Subsequently , the cells were washed twice , and analyzed by FACS CantoII with Diva software ( BD Biosciences ) . Mouse tissue samples were either fixed in 10% formalin in phosphate buffer and then embedded in paraffin or frozen in embedding medium Optimal Tissue-TeK ( SAKURA Finetek Japan ) . Hematoxylin and eosin staining was performed according to standard procedures . Tissue sections prepared from the frozen samples were also stained with anti-mouse IFN-γ ( RMMG-1 , Abcam ) , CD11a ( M17/4 , BioLegend ) , CD18 ( N18/2 , BioLegend ) , CD103 ( M290 , BD Pharmingen ) and CD54 ( ICAM-1 ) ( YN1/1 . 7 . 4 , BioLegend ) . Images were captured using a Provis AX80 microscope ( Olympus ) equipped with an OLYMPUS DP70 digital camera , and detected using a DP manager system ( Olympus ) . The α chemokines CXCL9 and CXCL10 were analyzed using an enzyme linked immunosorbent assay ( ELISA ) . For α chemokines , capture and detection antibody concentrations were optimized using recombinant chemokines from R&D Systems Inc . ( Minneapolis , MN , U . S . A . ) according to the manufacturer's guidelines . Genomic DNA was extracted from sorted Treg cells as described below . One mg of genomic DNA ( 10 µl ) was denatured by the addition of an equal volume of 0 . 6 N NaOH for 15 min , and then 208 µl of 3 . 6 M sodium bisulfite and 12 µl of 1 mM hydroxyquinone were added . This mixture was incubated at 55°C for 16 hours to convert cytosine to uracil . Treated genomic DNA was subsequently purified using the Wizard clean-up system ( Promega ) , precipitated with ethanol , and resuspended in 100 µml of dH2O . Sodium bisulfite-treated genomic DNAs ( 50 ng ) were amplified with primers targeting the specified DNA regions , and then PCR products were subcloned into the pGEM-T Easy vector ( Promega ) for sequencing . Sequences of 10 clones were determined for each region using Big Dye Terminator ( Perkin Elmer Applied Biosystems ) with an ABI 3100 autosequencer . The primers used for nested PCR were as follows: for the mouse Foxp3 promoter: mproF , 5′-GTGAGGGGAAGAAATTATATTTTTAGATG-3′; mproR , 5′-ATACTAATAAACTCCTAACACCCACC-3′; mproF2 , 5′-TATATTTTTAGATGATTTGTAAAGGGTAAA-3′; mproR2 , 5′-ATCAACCTAACTTATAAAAAACTACCACAT-3′ . For mouse Foxp3 intronic CpG: mintF , 5′-TATTTTTTTGGGTTTTGGGATATTA-3′; mintR , 5′-AACCAACCAACTTCCTACACTATCTAT-3′; mintF2 , 5′-TTTTGGGTTTTTTTGGTATTTAAGA-3′; mintR2 , 5′-TTAACCAAATTTTTCTACCATTAAC-3′ . To sort Treg cells , we isolated mouse splenocytes and resuspended them in FACS buffer for subsequent staining with the following antibodies purchased from BD Pharmingen: anti-mouse CD4 ( RM4-5 ) , GITR ( DTA-1 ) , CD25 ( PC61 ) . CD4+CD25+GITRhigh cells and CD4+CD25−GITRlowcells were sorted as Foxp3+ or Foxp3−cells using FACS AriaII with Diva software ( BD Biosciences ) . To confirm the purity of the sorted Treg cells , we measured the percentage of Foxp3 expression by intracellular staining , as described above . Sorted Treg cells were cultured in RPMI1640 containing 10% FBS , antibiotics , and 50 µM 2-mercaptoethanol ( Invitrogen ) . Total RNA of sorted cells was extracted with TRIZOL reagent ( Invitrogen ) according to the manufacturer's instructions . Approximately 200 ng of RNA were used to prepare cDNA using the SuperScript III enzyme ( Invitrogen ) . Levels of HBZ and Foxp3 transcripts were determined with FastStart Universal SYBR Green Master reagent ( Roche ) in a StepOnePlus real time PCR system ( Apllied Biosystems ) . Data was analyzed by the delta Ct method . The sequence of the primers used were as follows: HBZ Forward: 5′-GGACGCAGTTCAGGAGGCAC-3′ , Reverse: 5′-CCTCCAAGGATAATAGCCCG-3′; Foxp3 Forward: 5′-CCCATCCCCAGGAGTCTTG-3′ , Reverse: 5′-ACCATGACTAGGGGCACTGTA-3′; 18S rRNA Forward: 5′-GTAACCCGTTGAACCCCATT-3′ , Reverse: 5′- CCATCCAATCGGTAGTAGCG -3′ . | Viral infection frequently induces tissue inflammation in the host . HTLV-1 infection is associated with chronic inflammation in the CNS , skin , and lung , but the inflammatory mechanism is not fully understood yet . Since HTLV-1 directly infects CD4+ T cells , central player of the host immune regulation , HTLV-1 should modulate the host immune response not only via viral antigen stimulation but also via CD4+ T-cell-mediated immune deregulation . It has been reported that Foxp3+CD4+ T cells are increased in HTLV-1 infection . It remains a central question in HTLV-1 pathogenesis why HTLV-1 induces inflammation despite of increase of FoxP3+ cells , which generally possess immune suppressive function . We have elucidated here that most of the increased Foxp3+ cells in HBZ-Tg mice or HAM/TSP patients is not thymus-derived naturally occurring Treg cells but induced Treg cells . Since the iTreg cells are prone to lose FoxP3 expression and then become cytokine-producing cells , the increase of iTreg cells could serve as a source of proinflammatory CD4+ T cells . Thus HTLV-1 causes abnormal CD4+ T-cell differentiation by expressing HBZ , which should play a crucial role in chronic inflammation related with HTLV-1 . This study has provided new insights into the mechanism of chronic inflammation accompanied with viral infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2013 | HTLV-1 bZIP Factor Induces Inflammation through Labile Foxp3 Expression |
Reciprocating exchange with other humans requires individuals to infer the intentions of their partners . Despite the importance of this ability in healthy cognition and its impact in disease , the dimensions employed and computations involved in such inferences are not clear . We used a computational theory-of-mind model to classify styles of interaction in 195 pairs of subjects playing a multi-round economic exchange game . This classification produces an estimate of a subject's depth-of-thought in the game ( low , medium , high ) , a parameter that governs the richness of the models they build of their partner . Subjects in each category showed distinct neural correlates of learning signals associated with different depths-of-thought . The model also detected differences in depth-of-thought between two groups of healthy subjects: one playing patients with psychiatric disease and the other playing healthy controls . The neural response categories identified by this computational characterization of theory-of-mind may yield objective biomarkers useful in the identification and characterization of pathologies that perturb the capacity to model and interact with other humans .
Here we write the model for how player forms an estimate of optimal play at each round t by calculating the values of their possible actions . The actions are the amounts to invest or to return . The values are the expected summed utilities over the next two rounds ( as a simplification , players are assumed to look at the current round and the round after ) . The utility for player i depends on the actions of player j , which in turn depends on the type of player j , and the reasoning that player j does about player i . Player i does not know player j's type , but can learn about it from the history of their interactions , which , up to round t , is . Formally , player i maintains beliefs , in the form of a probability distribution over the type of player j , and computes expected utilities by averaging over these beliefs . Bayes theorem is used to update the beliefs based on evidence . The value on round t is a sum of two expectations: ( 2 ) The first is the utility of the exchange on that round . This iswhere , for convenience , we write as a function of the possible actions a of player j rather than the money this player earns . The second term in equation ( 2 ) concerns the value of future 2 rounds in the exchange ( except in the last round , where this term is 0 ) . This is thus an average over values on round t+1 , where the new beliefs take account of the action being considered by player i , and all the possible actions of player j . Equation ( 2 ) is a form of Bellman evaluation equation . The players can calculate the values , including updating the beliefs , by simulating the course of play with their partners . This simulation is a central feature of the model with players adopting higher levels of depth-of-thought requiring more simulation ( see belief updates in Supporting Information ) .
The model described above constitutes a full generative account of a subject playing the multi-round trust game , and incorporates several key cognitive mechanisms engaged by such a staged interpersonal interaction . Player i is characterized by their private type , their depth-of-thought level , and the prior beliefs . Player j is characterized in just the same way . We estimated the parameters of both players in each dyad by maximizing the log likelihood of their choices over the 10 rounds of the game . The averaged maximal log likelihood of all 195 investors was −11 . 92±0 . 27 . In our model , we assume that players take one of five possible actions . If all the five possible actions were chosen with equal probability , the log likelihood would take the value . Our computational theory-of-mind model fitted the behavior significantly better than assuming that players act randomly ( one-sample test , P = 1 . 51×10−35 ) . For the purposes of comparison , we also built a reinforcement learning ( RL ) model incorporating inequality aversion ( details in Supporting Information ) . We found that the RL model performed poorly; when we optimized the learning rate in the model , the optimum was degenerate in the sense that no learning occurred , and all actions were selected with equal probability ( random choices ) . Figure 2A shows the frequency histogram of depth-of-thought classification achieved by inverting the generative model described above . About half of all the 195 investors are classified as strategic level 0 . The remaining investors are almost equally divided into level 1 and level 2 players . There are significant dynamic behavioral features that correlate with the depth-of-thought levels that we estimate using our model . The style of play across rounds of the game is different and correlates well with the intuition that players with higher depths-of-thought are sensitive to richer features of the game than those possessing lower levels . In Figure 2B , of all 195 investors , levels 1 and 2 start the game with high offers and maintain throughout the game , except that the highest depth-of-thought players decrease their offers towards the end of the game ( which is strategic ) . Moreover , level 0 investors open low and stay low throughout the game , a strategy that tends to break cooperation with the trustee . Lastly , level 1 and 2 players make significantly more money overall than level 0 players ( Figure 2C ) . According to the generative model , players make predictions about the likely course of events through the game . These predictions lead naturally to prediction errors , which can be used to generate control signals to guide choices . In games against nature , prediction errors associated with rewarding outcomes have frequently been observed in the BOLD signal measured in striatal regions [17]–[19] . Games against other players offer much richer possibilities for neural responses since players have a range of interpersonal signals that they can model ( e . g . Figure 1C ) . We here focus on the investor side of the interaction because this role has proved to be particularly sensitive for classifying styles of play in prior work [20] . Two types of interpersonal prediction errors emerge naturally in the reciprocating interactions of the multi-round trust game . The first order prediction error in the investor is a comparison between the investor's current model of what the trustee will return and the amount actually returned . This error is computed at the time that the repayment from the trustee is revealed to the investor . This error requires information sent back from the trustee . By contrast , the second order prediction error in the investor requires a comparison between the investor's offer and the investor's internal model of what the trustee expects from the investor , that is , information that is exclusively internal to the investor . This information is available to the investor before any immediate feedback from the trustee , and is potentially available during the entire epoch , starting from the time of the cue and up until the time when the actual investment is made . In this paper , we choose the time the investor submits as a natural trigger for this signal , but with the understanding that it might have been computed and thus available earlier . Thus , the first order error can be evaluated at the time the repayment from the trustee is revealed . In a similar spirit , the second order error is defined at the time the investor's offer is submitted since it is at this time that the investor brain can compare their actual offer to their ( internal ) model of what the trustee expects . Our hypothesis for the first order inter-personal prediction error was that players classified as level 0 would display a large response to this error , while the higher levels would not , since this signal is not a critical component of the high level players' planning . We divided the first order interpersonal prediction error of all 195 healthy investors classified within a certain cognitive level into quintiles , performed separate GLM analysis at individual rounds , and then generated contrasts between rounds with high 1st order prediction errors ( >60% ) and rounds with low 1st order prediction errors ( ≤40% ) on the beta images of the events of interest . The contrast analysis at the revelation of the trustee's repayment showed that level 0 investors ( n = 102 ) had robust activations in bilateral striatal regions ( Figure 3A , whole-brain FDR corrected at P<0 . 05; peak MNI coordinates: right caudate ( 8 , 12 , 0 ) , t = 4 . 49 , 57 voxels; left caudate ( −12 , 12 , 4 ) , t = 3 . 74 , 73 voxels; right putamen ( 24 , 4 , 0 ) , t = 4 . 02 , 88 voxels; left putamen ( −24 , 4 , 4 ) , t = 4 . 10 , 72 voxels ) . These striatal activations were not observed in investors with level 1 ( n = 49 ) or level 2 ( n = 44 ) depth-of-thought . We also performed a direct comparison among investors with different depth-of-thought levels on the 1st order interpersonal prediction errors using ANOVA . The group contrast results showed that the level 0 investors had higher caudate activation than level 1 investors ( Figure 3B left , P<0 . 001 , uncorrected; peak MNI coordinates: ( 4 , 16 , 0 ) , t = 4 . 04 , FWE corrected at P<0 . 05 with small volume correction applying the anatomical mask of bilateral caudate ) . We also found that level 2 investors had higher right temporal-parietal junction ( TPJ ) activation than level 0 investors associated with the 1st order interpersonal prediction errors ( Figure 3B right , whole-brain FDR corrected at P<0 . 05; peak MNI coordinates: ( 52 , −48 , 28 ) , t = 4 . 70 , 7 voxels ) . Our hypothesis for the second order inter-personal prediction error was that players classified as level 0 would display no response to this higher order interpersonal error ( since their model of the other's model of themselves is impoverished ) , whereas players classified as higher level would . We divided the second order inter-personal prediction error of all 195 healthy investors classified within a certain cognitive level into quintiles , performed separate GLM analysis at individual rounds , and then generated contrasts between rounds with high 2nd order prediction errors ( >60% ) and rounds with low 2nd order prediction errors ( ≤40% ) on the beta images of the events of interest . The contrast at the submission of the investor's decisions revealed that level 2 investors had significant activations in bilateral putamen ( Figure 3C , whole-brain FDR corrected at P<0 . 05; peak MNI coordinates: right putamen ( 24 , 8 , −4 ) , t = 3 . 79 , 23 voxels; left putamen ( −20 , 8 , −4 ) , t = 3 . 11 , 7 voxels ) . We did not observe any striatal activations in level 0 and level 1 investors for the 2nd order prediction errors . We also performed an ANOVA analysis on the three depth-of-thought levels of investors . The group contrast analysis found that level 2 investors had higher ventral striatal activation than level 0 investors when computing the 2nd order interpersonal prediction errors ( Figure 3D , P<0 . 005 uncorrected; peak MNI coordinates ( 12 , 8 , −12 ) , t = 3 . 41 , FWE corrected at P<0 . 05 with small volume correction applying the anatomical mask of bilateral caudate ) . It is possible that when grouping the rounds according to the high or low quintiles of prediction errors , some subjects might be exclusively included in the high group , or in the low group . This raised the concern that the contrast results above might be biased by those distinct subjects . We therefore counted the number of subjects only present in the high group , or in the low group for the 1st and 2nd interpersonal prediction errors , respectively . We showed that the vast majority of subjects made contributions to all quintiles of prediction errors , with only an extremely small number of subjects contributing to just the high or low quintiles ( Table S1 ) . We also plotted the magnitudes of the interpersonal prediction errors divided into high or low quintiles across the depth-of-thought levels . We did this to rule out the possibility that a few subjects were dominating the observed results . The differences between the high and low quintiles were comparable across all the three levels of investors for both the 1st and 2nd order interpersonal prediction errors ( Figure 4 ) . Thus , the differential neural activations to the prediction errors observed here cannot be attributed to the differences in the magnitudes of prediction errors per se . Earlier work [9] found that trustees diagnosed with Borderline Personality Disorder ( BPD ) played uncooperatively to an extent that they could not maintain the cooperation of their partner investor . In that work , the impact of the trustee behavior was ‘read out’ through the willingness of the investor to sustain high offer levels throughout the rounds of the game . Figure 5 shows two distributions of estimated investor depth-of-thought levels as a function of distinct trustee types . Panel A shows the distribution when healthy investors play anonymous healthy trustees ( n = 48 pairs ) . In this exchange , healthy subjects never meet their partner before the game and do not see or meet them after the game . They arrive at the lab and are randomly assigned roles in separate rooms . Panel B shows the distribution when healthy investors play subjects diagnosed with borderline personality disorder . There is a more dramatic shift toward lower depth-of-thought levels despite the fact that these subjects play the healthy investor anonymously . The distributions in panels A and B are statistically different ( see legend Figure 5 ) . We also recruited 38 trustee matched for lower socio-economic scale ( SES ) as a SES match for the Borderline personality disorder trustees . These trustees also played anonymously and induced a similar lower depth-of-mind distribution in the investors ( Figure S2 ) suggesting that lower SES may be one source of influence for the incapacity of the Borderline subjects to sustain cooperation with their investor partners .
In this paper , we used a Bayesian computational model that involves an explicit representation of theory of mind to classify a large number of subjects playing an economic exchange game . We used the model to assess their level of depth-of-thought . Our classification produces three levels of players whose behaviour correlates with important measures of performance through the task . Neuroimaging results based on the model classification showed a differential response to depth-of-thought . Additionally we found a significant difference for investor depth-of-thought distributions when comparing play with healthy trustees to play with subjects diagnosed with borderline personality disorder ( BPD ) , a disorder known to disrupt inter-personal interactions . BPD subjects are characterized by their unstable relationships , and when they have played this game , they have tended to break cooperation . Indeed , it has been shown that , for this group , the anterior insula failed to sense the opponent's low offers [8] . The striatum has long been shown to encode reward prediction error signals in both passive and instrumental conditioning tasks [17] , [21]–[23] . Recently striatal activation has also been observed in social learning tasks [24] and tasks requiring mentalizing a partner's intention [3] . Here we found that striatum activity correlated with two types of interpersonal prediction errors evoked in a repeated social exchange game , and that these signals were modulated by players' depth-of-thought levels . Level 0 players , but not level 2 players , had robust activations in the striatum to high 1st order interpersonal prediction errors suggesting the naïve players were particularly sensitive to opponent's actions and mainly used this type of errors to adjust their own action policy . However , the striatum in level 2 players responded only to the 2nd order interpersonal errors suggesting that these relatively sophisticated players discounted the direct influence of opponent's actions and rather put more emphasis on simulating and manipulating opponent's beliefs and actions . Other imaging experiments requiring subjects to model others' intentions have also reported activations in frontoparietal regions [3] , [5] , [24] . It is not clear why frontoparietal regions were not observed in our paradigm . However , there is a clear path from known error signaling in the striatum to our observations here of 2nd order inter-personal prediction errors , since a 2nd order prediction error can be seen as a direct proxy for future returns to the investor . In this reciprocation game , we have previously reported that deviations from neutral reciprocity or tit-for-tat behavior cause players to change their behavior [7] , [9] . Therefore , an investment that deviates positively from what the trustee expects ( based on their model of the investor ) should generate a positive error signal in the trustee's brain , which would itself lead to the investor expecting an increased return . Under this interpretation , the signal is exactly analogous to the range of prediction error signals that show up encoded in BOLD responses in the striatum . These neural results are congruent with our behavioral observations . The most sophisticated level 2 investors invested high at the beginning to cultivate trust and promote cooperation with their partners . But towards the end of the exchange , they responded to the horizon of the game and risked less money , reflecting their manipulative maneuver in the beginning . Furthermore , we found that the sophisticated level 2 investors had higher activations in the right TPJ in response to the 1st and 2nd order interpersonal prediction errors than the naïve level 0 investors . Right TPJ has been demonstrated to play a critical role in belief reasoning tasks involving “theory of mind” [25] , [26] . Right TPJ has also been found to be specifically modulated in people with higher strategic levels [27] . Furthermore the coordinates of the peak voxel of this activation place it in a recently designated posterior region of the TPJ ( TPJp ) that is well-connected to “areas identified with social cognition” [28] . The TPJ activation and its specific location within TPJ is consistent with the idea that level 2 investors build more sophisticated models of their opponents . Computational accounts developed in the framework of Markov Decision Processes ( MDP ) , and in particular reinforcement learning models [29] , have been successful in representing behavior and illuminating neural substrates in situations where agents interact with nature , and in which the environmental states are fully observable . Such models have furthered our understanding of the role of dopamine and related neural structures in reward learning and decision-making [30] , [31] . However , those models are limited in the typical social situations where agents interact and effectively create an ever-changing , adapting landscape , which are plausibly a raison d'etre for sophisticated cognition . Recently , some progress has been made in establishing model-based approaches to social interaction [3] , [4] , [32] , [33] . Our approach makes a commitment to an explicit , generative model of higher-order thinking about other social actors , some aspects of which are in common with the recent work by Yoshida et al . ( who also use their models to compare autistic and healthy subjects ) [4]–[6] . The space of such models is vast , and explicit choices must be made at many steps [4] , [10] . Nonetheless , our model is able to capture striking heterogeneity in the behavior which we are then able to connect to differences in neural activity . Further developments of this approach also incorporating genetic data promise to help uncover the genetic underpinnings of social heterogeneity .
Informed consent was obtained for all research involving human participants , and all clinical investigation was conducted according to the principles expressed in the Declaration of Helsinki . All procedures were approved by the Institutional Review Board of the Baylor College of Medicine . Data from four groups , total 195 pairs of subjects ( 18–64 yrs ) who played the trust game previously [5]–[8] were examined , including an Impersonal group ( 48 pairs ) , a Personal group ( 54 pairs ) , a BPD group ( 55 pairs ) , and a BPD control group ( 38 pairs ) . Subject pairs from the Impersonal , BPD , and BPD control groups never met each other throughout the experiment . Subject pairs in the Personal group were introduced to each other before playing the task . Trustees in the BPD group were diagnosed with borderline personality disorder ( BPD ) , and were matched to trustees in the BPD control group on socioeconomic status ( SES ) . In addition , investors in the BPD and BPD control groups were recruited with socioeconomic status matched to trustees . Investors in the Impersonal groups were students from Caltech and Baylor College of Medicine . All scans were carried out on 3 . 0 Tesla Siemens Allegra scanners . High-resolution T1-weighted scans ( 1 . 0 mm×1 . 0 mm×1 . 0 mm ) were acquired using an MP-RAGE sequence ( Siemens ) . Subjects then played the iterated trust game for 10 rounds while undergoing whole-brain functional imaging . The detailed settings for the functional run were as follows: echo-planar imaging , gradient recalled echo; repetition time ( TR ) = 2000 ms; echo time ( TE ) = 40 ms; flip angle = 90°; 64×64 matrix , 26 4-mm axial slices angled parallel to the anteroposterior commissural line , yielding functional 3 . 3 mm×3 . 3 mm×4 . 0 mm voxels . Images were analyzed using SPM2 ( http://www . fil . ion . ucl . ac . uk/spm/software/spm2/ ) . Slice timing correction was first applied to temporally align all the images . Motion correction to the first functional image was performed using a 6-parameter rigid-body transformation . The average of the motion-corrected images was co-registered to each subject's structural images using a 12-parameter affine transformation . Images were subsequently spatially normalized to the Montreal Neurological Institute ( MNI ) template by applying a 12-parameter affine transformation , followed by nonlinear warping using standard basis functions . Finally , images were smoothed with an 8 mm isotropic Gaussian kernel and then high-pass filtered ( 128 s width ) in the temporal domain . Separate general linear models were specified for individual rounds of each subject ( 6 ) . All visual stimuli , motor responses and motion parameters were entered as separate regressors that were constructed by convolving each event onset with a canonical hemodynamic response function in SPM2 . Beta maps were estimated for regressors of interest . The SPM images shown in Figure 3 was generated as follows: both the first order and second order interpersonal prediction errors of subjects classified with the same depth-of-thought were divided into quintiles . For the 1st order interpersonal prediction errors , beta images associated with the event when the repayments were revealed were sorted according to the prediction error quintiles . Contrast analysis between the beta images from top two quintiles ( >60% ) and images from the bottom two quintiles ( ≤40% ) were performed . Similarly , contrasts for the 2nd interpersonal prediction errors were generated from beta images associated with the event when the investments were submitted . See Text S1 for detailed descriptions . We also include a reinforcement learning model in Text S1 for comparison . | Human social interactions are extraordinarily rich and complex . The ability to infer the intentions of others is essential for successful social interactions . Although most of our inferences about others are silent and subtle , traces of their effects can be found in the behavior we exhibit in various tasks , notably repeated economic exchange games . In this study , we use a computational model that uses an explicit form of other-modeling to classify styles of play in a large cohort of subjects engaging in such a game . We classify players according to their depth of recursive reasoning ( depth-of-thought ) , finding three groups whose performance throughout the task differed according to several measures . Neuroimaging results based on the model classification show a differential neural response to depth-of-thought . The model also detected differences in depth-of-thought between two groups of healthy subjects: one playing patients with psychiatric disease and the other playing healthy controls . These results demonstrate the power of a quantitative approach to examining behavioral heterogeneity during social exchange , and may provide useful biomarkers to characterize mental disorders when the capacity to make inferences about others is impaired . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"biology",
"neuroscience"
]
| 2012 | Computational Phenotyping of Two-Person Interactions Reveals Differential Neural Response to Depth-of-Thought |
Phylogeographic methods aim to infer migration trends and the history of sampled lineages from genetic data . Applications of phylogeography are broad , and in the context of pathogens include the reconstruction of transmission histories and the origin and emergence of outbreaks . Phylogeographic inference based on bottom-up population genetics models is computationally expensive , and as a result faster alternatives based on the evolution of discrete traits have become popular . In this paper , we show that inference of migration rates and root locations based on discrete trait models is extremely unreliable and sensitive to biased sampling . To address this problem , we introduce BASTA ( BAyesian STructured coalescent Approximation ) , a new approach implemented in BEAST2 that combines the accuracy of methods based on the structured coalescent with the computational efficiency required to handle more than just few populations . We illustrate the potentially severe implications of poor model choice for phylogeographic analyses by investigating the zoonotic transmission of Ebola virus . Whereas the structured coalescent analysis correctly infers that successive human Ebola outbreaks have been seeded by a large unsampled non-human reservoir population , the discrete trait analysis implausibly concludes that undetected human-to-human transmission has allowed the virus to persist over the past four decades . As genomics takes on an increasingly prominent role informing the control and prevention of infectious diseases , it will be vital that phylogeographic inference provides robust insights into transmission history .
Phylogeographic methods aim to infer many aspects of population evolution from genetic data . The phylogeography term often encompasses methods that infer changes in population size ( phylodynamics ) and population divergence events ( see [1] ) . In this work , we focus on the inference of migration between distinct subpopulations ( such as in the structured coalescent , see [2 , 3] ) . For many years , nested clade phylogeographic analysis ( NCPA , see e . g . [4 , 5] ) was the leading method to test for isolation and migration ( reviewed in [1 , 6] ) . More recently , model-based inference for phylogeography has flourished and has replaced NCPA as the new standard approach ( reviewed in [7 , 8] ) . Probabilistic model-based inference for phylogeography has widely been used to study the spread of pathogens between geographic locations and to identify their original source [9–12] , and they are commonly applied to study the migration history of animals [13–15] , plants [16 , 17] , and even languages [18] . Phylogeographic methods are useful for addressing a wide range of questions in epidemiology , for example in studying transmission of pathogens between body compartments within a host [19] , between individual hosts [20] , between host social groups [21] , and between host species [22] . One major class of modelling approaches comprise likelihood-based methods implementing the structured coalescent [23–27] , which corresponds to the classic migration matrix model [28] , a generalization of Wright’s Island model [29] . These approaches use the structured coalescent to infer migration rates and effective population sizes . However , they are impractical in scenarios with large numbers of subpopulations and migration events due to their computational demand . This is because they explore not only the parameters of primary interest ( such as migration rates , population sizes , and phylogeny ) but also all possible migration histories , vastly increasing the computational complexity . Recently , an alternative phylogeographic approach has risen in prominence , which treats the migration of lineages between locations as if the location were a discrete trait , evolving in a manner analogous to the substitution of alleles at a genetic locus [9 , 10 , 15] . Since migration is modelled like mutation , this approach is referred to as “Mugration” by [30 , 31] , or “discrete trait analysis” ( DTA in the following ) . The gained popularity of this approach ( see e . g . [1] ) is at least partly attributable to its computational efficiency and user-friendly software . However , the DTA model inherits a set of assumptions appropriate for the independent mutation of loci within lineages , but profoundly at odds with classical population genetics models of migration ( see e . g . [3 , 32 , 33] ) , as summarised in Table A in S1 Text . While methods based on the structured coalescent , which explicitly accounts for the effects of migration on the shape and branch lengths of the genealogy , are in theory often preferable to DTA , the latter is frequently chosen due to the computational demands of current implementations of the structured coalescent . DTA is also commonly used to describe the evolution of discrete phenotypes . In many such cases , DTA is appropriate [34 , 35] . However , use of the DTA entails a number of assumptions that are unusual or inappropriate when applied to the migration of lineages between geographic locations , for example ( i ) the relative size of subpopulations drifts over time , such that subpopulations can become lost ( extinct ) or fixed ( the sole remaining subpopulation ) instead of being constrained , e . g . by local competition , ( ii ) sample sizes across subpopulations are proportional to their relative size . There is a scarcity of studies in the scientific literature assessing the accuracy of DTA and comparing different phylogeographic approaches , but concerns have been raised because , among other issues , DTA is thought to be sensitive to local sampling intensity [12 , 36] . Further , the conceptual separation of coalescent process and migration process made by DTA is expected to lead to suboptimal use of information . Here we demonstrate that these concerns are well founded , in that DTA suffers from various biases and statistical inefficiency despite its computational speed . To address the problems with DTA we introduce a new model-based approach that achieves a close approximation to the structured coalescent ( similar in spirit to [37 , 38] ) . The idea behind this approximation is to efficiently integrate over all possible migration histories , therefore reducing the computational effort needed to explore the parameters of primary interest . We implement this approach , called BASTA ( BAyesian STructured coalescent Approximation ) , in the Bayesian phylogenetic package BEAST2 [31] . We compare its performance to DTA and MultiTypeTree ( MTT , a recent Bayesian structured coalescent software , see [27] ) using simulations based on the structured coalescent . We illustrate the use of the method to reconstruct transmission dynamics in human , animal and plant viruses . We demonstrate the important influence of model choice on study conclusions through an analysis of genomic data from previous and ongoing Ebola epidemics [39] using different phylogeographic approaches to interpret the role of zoonotic events in the origin of human outbreaks . Our results show , based both on simulations and real data analyses , that DTA and structured coalescent methods can lead to different conclusions , and that DTA is often inaccurate .
In this section we define the structured coalescent before describing the DTA model and introducing BASTA . The structured coalescent is a statistical model describing the genealogy of individuals sampled from a structured population that evolves according to the migration matrix model [28] . For simplicity , here we assume individuals are haploid , but the model applies more generally . The key assumptions are: ( i ) The subpopulations , or demes , are stable in size over time , with their effective sizes defined by the vector θ . ( ii ) Migration occurs at a constant rate over time , defined by the migration matrix f , such that fa , b is the total rate of migration of individuals from deme a to b , divided by the effective number of individuals in deme a . ( iii ) There is no substructure within demes . ( iv ) There are no differences in fitness between individuals . ( v ) Within demes , individuals are sampled at random . However , no assumptions are made about the total sample size nor the relative sample sizes per deme . A potential source of confusion arises from the convention of using the backwards-in-time migration rate matrix m in the structured coalescent , defined such that mb , a is the total rate of migration of individuals from deme a to b , divided by the effective number of individuals in deme b . Mathematically , mb , a = fa , b θa/θb . The backwards migration matrix m is considered convenient because it provides the rate at which a lineage appears to move between demes backwards in time . For this reason , we refer to f as the forwards-in-time migration rate matrix . In the notation of [27] , the demes are represented by a set D , sampled individuals are represented by the set I , the aligned sequences by the set S = {si∣i ∈ I} , the sampling dates by the set tI = {ti∣i ∈ I} and the sampling locations by the set L = {li∣i ∈ I} . In addition to the parameters of primary interest , m and θ , T represents the genealogy , μ the nucleotide substitution rate matrix and M the migration history of lineages in the tree , i . e . the timing , source , sink and lineage involved in each migration event . MultiTypeTree ( MTT ) is a method implemented in BEAST2 for estimating the parameters of the structured coalescent by Bayesian inference [27] . Formally , the target of inference is the posterior distribution of the parameters given the data: P ( T , M , μ , m , θ | S , t I , L ) ∝ P ( S | T , t I , μ ) P ( T , M | t I , L , m , θ ) P ( μ , m , θ ) . ( 1 ) The posterior consists of several components . The first term on the right is the likelihood of the sequences given the genealogy and substitution model , which is computed using Felsenstein’s pruning algorithm [40] . The second term is the probability density of the genealogy and migration history under the structured coalescent given the migration matrix and effective population sizes . The third term represents the prior distribution assumed for the parameters , and might be factored into independent priors for the separate parameters , P ( μ ) P ( m ) P ( θ ) . To calculate P ( T , M∣tI , L , m , θ ) under the structured coalescent , the sequence of B time intervals between successive events ( coalescence , sampling , or migration ) is considered , starting from the most recent sample and going back to the root of the genealogy . Suppose that the vector τ records the duration of each time interval . For a haploid population , P ( T , M | t I , L , m , θ ) = ∏ i = 1 B L i , ( 2 ) where L i = exp [ - τ i ∑ d ∈ D ( ( k i , d 2 ) 1 θ d + k i , d ∑ d ′ ∈ D , d ′ ≠ d m d d ′ ) ] E i , ( 3 ) ki , d is the number of lineages in deme d in interval i , and Ei is the contribution of the event that ends interval i: E i = { 1 if it is a sampling event m d d ′ if it is a migration event from d to d ′ 1 θ d if it is a coalescence event in deme d . ( 4 ) In the structured coalescent , migration events affecting lineages in the genealogy are explicitly parameterized and estimated ( Fig 1 ) . In the context of migration , DTA [9 , 15] is a model that achieves much greater computational efficiency by integrating over all possible migration histories using the pruning algorithm [40] , which is widely-used in phylogenetics to integrate over all possible mutation histories . However , treating the migration process as if it were analogous to a mutation process implies a set of assumptions that differ substantially from standard migration models . Namely: ( i ) The total effective population size is fixed to θ , but demes can change in relative size over time due to drift ( the chance birth and death of individuals ) ; the rate of drift is the same across demes and determined by the total effective population size . ( ii ) Demes can also change in relative size due to migration , which occurs at a constant rate per lineage , defined by the forwards-in-time migration matrix f . ( iii ) Individuals are sampled at random from demes in proportion to their relative size . There are some unusual consequences of the DTA modelling assumptions: ( i ) Demes can be lost , and they can be resurrected . ( ii ) The relative sampling intensities of the demes are treated as data , informative about the migration parameters , even before any sequence data is analysed . ( iii ) It is unclear what the relationship is between the effective population size parameter θ of the DTA and the vector of effective population sizes θ of the structured coalescent , hindering interpretation . Formally , the target of inference in a Bayesian implementation of DTA is the posterior distribution of the parameters given the data . This differs considerably from Eq 1 of MTT: P ( T , μ , f , θ | S , t I , L ) ∝ P ( L | T , t I , f ) P ( S | T , t I , μ ) P ( T | t I , θ ) P ( μ , f , θ ) . ( 5 ) The sampling locations L are treated as informative data , rather than uninformative auxiliary variables . The first term on the right is the likelihood of the sampling locations given the genealogy and migration matrix , calculated by integrating over all possible migration histories using the pruning algorithm . The second term is the likelihood of the sequences integrated over all possible mutation histories using the pruning algorithm , as in MTT . The third term is the probability density of the genealogy , approximated by a standard neutral coalescent prior for an unstructured population [41] . The fourth term represents the prior distribution . A different prior is required for the effective population size parameter θ in DTA because , differently from MTT , θ is the same for all demes . In essence , the assumptions of DTA are well motivated when employed to analyse randomly sampled alleles or discrete phenotypes which evolve independently across individuals . But they are questionable when employed to analyse the migration of individuals between subpopulations , whose relative frequencies are maintained by external forces such as resource availability , and for which the sampling frame might not be related to the relative sizes of the subpopulations . The consequences of the various approximations that the DTA represents have not been thoroughly explored in the literature , despite the popularity of the approach . One concern is that the assumption that sampling intensity is proportional to subpopulation size leads to biased estimates of migration rates when this assumption is not met [12 , 36] . Second , ignoring the population structure when calculating the probability of the coalescent tree could lead to bias or lost power . For example , when migration rates are very low , one expects very long branches close to the root . This interdependency between the shape and branch lengths of the genealogy and the migration process is ignored by DTA , which could reduce accuracy . We test these concerns using simulations , described later in the Methods . As an alternative to DTA , we pursue an approximation to the structured coalescent that is both accurate and computationally efficient , which we have developed in a Bayesian statistical framework and implemented in BEAST 2 , a software package for Bayesian evolutionary analysis . Like DTA , we gain computational efficiency by integrating over all possible migration histories with an approximation . Unlike DTA , we treat the genealogy as informative about the migration process and the sampling locations as uninformative a priori . In our approximation to the structured coalescent , we split each interval between successive coalescence events into two sub-intervals , within which the migration of lineages ( backwards-in-time ) is independent of one another ( similarly to [37 , 38] ) . This approximation differs from the assumptions of DTA because it ensures that the rate of coalescence between lineages depends on the probability they are in the same deme at the same time . It is approximate because ( i ) within each interval we model locations of lineages independently of each other , and ( ii ) we update the probability distribution of lineages among demes only at the beginning and end of each interval instead of continuously in time . Formally , we seek to approximate the same posterior distribution as MTT , but integrated over all possible migration histories: P ( T , μ , m , θ | S , t I , L ) ∝ P ( S | T , t I , μ ) P ( T | t I , L , m , θ ) P ( μ , m , θ ) . ( 6 ) The first term on the right is the likelihood of the sequences given the genealogy and substitution model , as in Eq 1 . The second term is the probability density of the genealogy under the structured coalescent , integrated over migration histories . This must be approximated in BASTA because no exact form is known . The third term represents the prior distribution for the parameters , as in Eq 1 . To approximate P ( T∣tI , L , m , θ ) in BASTA , we consider the probability density of each time interval between successive events ( coalescence or sampling ) . Denoting each interval Ai = [αi − 1 , αi] , where αi is the older event time of Ai and αi − 1 the more recent one , the probability density of interval Ai can be written as L i ′ = exp [ - ∫ α i - 1 α i ∑ d ∈ D ∑ l ∈ Λ ∑ l ′ ∈ Λ , l ′ ≠ l P ( d l = d , d l ′ = d | t ) 1 θ d d t ] E i ′ , ( 7 ) where Λ is the set of all extant lineages during interval Ai , dl is the deme to which lineage l belongs , and P ( dl = d , dl′ = d∣t ) is the probability that lineages l and l′ are in the same deme d at time t . Ei′ is the contribution of the coalescent or sampling event: E i = { 1 if it is a sampling event , ∑ d ∈ D P l , α i , d P l ′ , α i , d 1 θ d if it is a coalescence between l and l ′ . ( 8 ) To approximate Li′ we first substitute P ( dl = d , dl′ = d∣t ) with P ( dl = d∣t ) P ( dl′ = d∣t ) , which treats the migration of lineages as if they were independent of one another . As shorthand , we define Pl , t to be the vector whose dth element is Pl , t , d = P ( dl = d∣t ) . Next , we split each interval Ai into two sub-intervals of equal length Ai1 = [αi − 1 , ( αi+αi − 1 ) /2] and Ai2 = [ ( αi+αi − 1 ) /2 , αi] , and replace Pl , t with Pl , αi − 1 for all t in Ai1 and Pl , αi for all t in Ai2 . The approximated probability density contributions of Ai1 and Ai2 become: L ˜ i 1 = exp [ - τ i 2 ∑ d ∈ D ∑ l ∈ Λ ∑ l ′ ∈ Λ , l ′ ≠ l P l , α i - 1 , d P l ′ , α i - 1 , d 1 θ d ] ( 9 ) and L ˜ i 2 = exp [ - τ i 2 ∑ d ∈ D ∑ l ∈ Λ ∑ l ′ ∈ Λ , l ′ ≠ l P l , α i , d P l ′ , α i , d 1 θ d ] E i ′ . ( 10 ) Further improvements to the approximation could be obtained by considering more sub-intervals , albeit at increased computational cost . Between intervals , the probability distribution of lineages among demes is updated as P l , α i = P l , α i - 1 exp ( τ i · m ) , ( 11 ) where time is scaled in Ne = ∑d ∈ D θd generations , the exponential is a matrix exponential , and m is the backwards-in-time migration rate matrix , whose diagonal elements are defined such that the rows sum to zero . For a lineage l sampled from deme d at time t , Pl , t is a vector whose dth element equals one and all other entries equal zero . If lineages l1 and l2 coalesce to an ancestral lineage l at time t , then P l , t = ( P l 1 , t , 1 P l 2 , t , 1 θ 1 , … , P l 1 , t , | D | P l 2 , t , | D | θ | D | ) ∑ d = 1 | D | P l 1 , t , d P l 2 , t , d θ d , ( 12 ) which is the normalised entrywise product ( element by element product ) of the distributions of the coalescing lineages . The probability density of the genealogy under the structured coalescent , integrated over migration histories , is finally approximated as P ( T | t I , L , m , θ ) = ∏ i = 1 B L ˜ i 1 L ˜ i 2 . ( 13 ) Details of how we efficiently compute these quantities , in particular Eq 10 , are given in S1 Text . The software implementing BASTA can be freely downloaded from https://bitbucket . org/nicofmay/basta-bayesian-structured-coalescent-approximation , including the source code . The software can alternatively be installed from the graphical user interface BEAUti [42] of BEAST2 . Example files and data from the analyses described hereby can be found in Supplementary S1 Dataset . To assess the adequacy of the approximations in BASTA , and to compare its performance to MTT and DTA , we performed simulations under the structured coalescent [2 , 3] with the software msms [43] . We quantified the performance of the methods by analysing a large number of datasets simulated from a range of migration rates . By comparing the simulated ( “true” ) and estimated parameters , we could assess performance using a number of statistics: Bias: mean difference between the simulated and estimated parameter . RMSE: square root of the mean squared difference between the simulated and estimated parameter . Correlation: Pearson’s correlation coefficient between the simulated and estimated parameter . Calibration: proportion of datasets for which the simulated parameter lay within the 95% credible interval . In all cases , point estimates were taken to be the estimated posterior median and 95% credible intervals were taken to be the 95% region of the estimated posterior distribution with the highest density . The theoretically optimal values for the bias , RMSE and correlation are 0 , 0 and 1 respectively . The theoretically optimal value for the calibration is 0 . 95 when the parameters are simulated under the same prior distribution as that used for analysis . Values greater than 0 . 95 are considered conservative . Since we expect the information content of the sequences ( including sequence length and diversity ) to have a strong effect on the analysis , we investigated three levels of genetic information: Fixed tree: abundant genetic data so that the genealogical topology and branch lengths are essentially known ( up to a scaling factor ) without error , achieved by providing BEAST2 with the simulated genealogy . Even in this scenario , we still expect uncertainty in parameter estimates due to inherent stochasticity in the migration process . Variable tree: limited genetic data so that there is uncertainty in the genealogy . For this we simulated an alignment of 2000 bp using SeqGen [44] with a transition/transversion ratio of κ = 3 and mutation rate of 0 . 01 in units of Ne generations , and we estimated the genealogy in BEAST2 along with the other parameters . No data: to test for susceptibility to sampling bias , we took the unusual step of analysing sequence data that were completely uninformative about the genealogy by providing a single ambiguous base ( ‘N’ ) for each individual . Unless the method is biased , the posterior produced by BEAST2 in this case should equal the prior . We simulated under two scenarios , a “Continental” model with two subpopulations , and an “Archipelago” model with eight subpopulations , and investigated the performance of the methods under different sampling strategies ( even versus uneven ) and mean migration rate ( fast versus slow ) . In the Continental model , we considered two subpopulations , with different rates of migration between the two , and a total sample size of 200 . We compared even sampling , in which 100 individuals were sampled per subpopulation , to uneven sampling , in which 10 individuals were sampled from one subpopulation and 190 from the other . We sample migration rates used in simulations from the DTA prior distribution , that is , relative migration rates r1 , 2 and r2 , 1 were simulated from independent Γ ( 1 . 0 , 1 . 0 ) distributions . This mildly favours DTA because MTT and BASTA use log-normal priors with σ = 4 instead . The relative migration rates were then rescaled so that the mean migration rate f ‾ was equal to a value simulated from an exponential distribution with mean 0 . 1 ( very slow ) , 0 . 5 ( slow ) , 2 . 0 ( moderate ) or 5 . 0 ( fast ) . After rescaling , f1 , 2 = c r1 , 2 and f2 , 1 = c r2 , 1 , where c = f ‾ ( r 1 , 2 + r 2 , 1 ) / ( 2 r 1 , 2 r 2 , 1 ) . Since DTA assumes that the rate of drift is the same in every deme , we fixed all effective population sizes in the simulations and in BEAST2 to be equal to one in order to reduce the disparity in modelling assumptions . This has the effect of ( i ) simplifying interconversion between forwards-in-time and backwards-in-time migration rates , so that fi , j = mj , i and ( ii ) scaling migration rates in “coalescent time units” . However , the interpretation of effective population size ( and hence coalescent time units ) differs between the structured coalescent and DTA models , so we based model comparison on the relative migration rate f1 , 2/f2 , 1 . In the Archipelago model , we considered two groups ( archipelagos ) of four subpopulations ( islands ) , with two migration rates: a faster rate between islands in the same archipelago and a slower rate between islands not in the same archipelago . Forty individuals were sampled from each subpopulation . We fixed the rate of migration within ( fw ) and between ( fb ) archipelagos to fw/fb = 10 and simulated fb from an exponential distribution with mean 0 . 5 . From all simulations , migration rates and root location were then estimated using DTA [9 , 15] , MTT and BASTA , all as implemented in BEAST2 . For the “No data” scenario , the posteriors from ten independent chains were merged , each of 5 × 106 iterations . For the “Fixed tree” scenario , a single chain of respectively 106 , 2 × 105 , and 105 iterations for DTA , MTT and BASTA was used . For the “Variable tree” scenario , we used a single chain of respectively 2 × 107 , 2 × 107 , and 107 iterations for DTA , MTT and BASTA . Finally , for the “Archipelago” scenario we used a single chain of 2 × 106 iterations for both MTT and BASTA . We applied DTA , MTT and BASTA to two datasets with moderately high numbers of subpopulations , one consisting of a collection of Avian Influenza Virus ( AIV ) haemagglutinin ( HA ) segments collected from different avian hosts and different locations [45] , and one of a collection of Tomato Yellow Leaf Curl Virus ( TYLCV ) sequences ( the CP dataset free of detectable recombination from [46] ) . For the AIV data we use two distinct subdivisions of samples into discrete host species classes , following the classifications in [45] . The first involves five groups , and the second ten groups . For the TYLCV dataset we used a single subdivision of samples into eight geographical classes , obtained following [46] . In these analyses , the effective population sizes of all demes were set equal in both MTT and BASTA . To study changes of host type in Ebola we used whole genome Ebola sequences from 78 patients recently obtained and aligned with sequences from previous outbreaks [39] . The authors of this study investigated the phylogenetic relationship of samples within or between Ebola outbreaks . We applied the three phylogeographic methods presented above to infer the contribution of zoonotic events to Ebola spread . We used the same alignment provided in [39] for the BEAST2 analysis , including sampling dates , but we also added information regarding host type . We defined two subpopulations , human and animal reservoir , and we allowed lineages to transmit forwards in time from the animal reservoir to a human host , but not vice-versa . So our phylogeographic model had two locations ( respectively human and animal reservoir ) but migration was only assumed to occur in one direction . This results in a structured coalescent model with three phylogeographic parameters for MTT and BASTA ( one migration rate and two effective population sizes ) , but only two parameters for DTA , as only a single general effective population size can be defined in that model . A peculiarity of these analyses is that no samples from one of the two considered populations were available . While this might seem an impassable limitation , previous studies have shown that the structured coalescent can provide meaningful estimates even in the absence of samples from one populations ( i . e . “ghost deme” , see [47] ) , suggesting that it is possible to perform statistical inference of zoonosis rates in this scenario . Since the inclusion of no animal samples is unusual , we considered a second , more typical , analysis in which we included genetic sequences from bats . Relatively little sequencing has been performed in potential animal reservoirs , so we were able to include only partial Ebola virus sequences from a 265 bp region of the polymerase ( L ) gene from seven bats collected in [48] . In this analysis , it was necessary to allow a small but non-zero rate of migration from humans to the animal reservoir to avoid predetermining inference of the ancestral location of the root . Therefore we constrained the migration rate from humans to animals at a rate 105 times lower than the animal to human rate . This preserves the ability of the model to infer ancestral locations in either of the two subpopulations , once samples from the animal reservoir have been included .
To test for susceptibility to biases associated with sampling strategy in DTA , MTT and BASTA , we analysed datasets completely lacking any genetic information ( the “No data” scenario ) , and containing only the sampling locations of 200 individuals from two populations in our Continental model . We compared two sampling strategies . In the first , individuals were sampled evenly ( 100 per subpopulation ) and in the second , unevenly ( 10 from one and 190 from the other ) . All three methods are Bayesian , so in the absence of information we expect the posterior distribution of the parameter of interest to be unchanged from the prior . For comparability across methods , the parameter we analysed was f1 , 2/f2 , 1 , the ratio of migration rates between the two subpopulations . We found that for DTA the posterior distribution was substantially different to the prior , exhibiting a bias that depended on sampling , and a reduction in parameter uncertainty , unlike the structured coalescent methods ( MTT and BASTA ) . Particularly with high migration rates ( mean f ‾ = 5 . 0 ) DTA posteriors showed large biases ( posterior median of rates log-ratio 1 . 7 with standard deviation 0 . 94 , Fig 2a ) , indicating that the sampling strategy significantly influenced the result . The posterior distributions for MTT and BASTA were unbiased , centred on the prior mean of 0 . 0 , but noticeably less smooth ( Fig 2b and 2c ) , indicating that they need running for longer than DTA . Even when migration rates were low ( mean f ‾ = 0 . 1 ) DTA substantially over-estimated them ( Fig . Aa in S1 Text ) . This is because the DTA model expects that , at low migration rates , one subpopulation will drift to high frequency , and that samples are collected proportionally to subpopulation size , so a random sample would be unlikely to capture multiple locations . The presence of multiple locations therefore suggests to DTA an appreciable migration rate . In contrast , the structured coalescent allows arbitrary sampling schemes and accounts for the fact that there must be at least D − 1 migration events when D subpopulations are sampled , regardless of migration rates . Next we assessed the accuracy of the 95% credibility intervals produced by the three methods . Again employing the Continental model , this time we quantified the performance of the methods in the favourable situation of highly informative sequences . Methods are expected to perform best when genetic data is so informative that the phylogenetic tree can be estimated with little error . We investigated this scenario by providing the true tree topology and relative branch lengths as input , and estimating only the tree height together with the migration rate parameters ( the “Fixed tree” scenario ) . As before , we analysed f1 , 2/f2 , 1 the ratio of migration rates between the two subpopulations . DTA exhibited generally poor performance ( Fig 3 , and Fig . B in S1 Text ) , with overly narrow credible intervals . The 95% credibility intervals were not well calibrated , including the true parameter between 56%-81% of the time , compared to 80%-96% for MTT , 84%-97% for BASTA , and the theoretical target of 95% ( Table 1 ) . Furthermore , the point estimates ( posterior median ) were much less well correlated with the true parameter values for DTA ( 0 . 33–0 . 64 ) than for BASTA ( 0 . 51–0 . 85 ) and MTT ( 0 . 42–0 . 77 ) , indicating poorer statistical efficiency . Poor performance was not restricted to estimating relative migration rates . The accuracy with which the location of the root ( the most recent common ancestor ) was estimated was 54% for DTA , compared to 68% for MTT and 77% for BASTA ( Fig 4 and Fig . C in S1 Text ) . Earlier methods for estimating parameters of the structured coalescent exhibited disproportionately increased computational demands with elevated migration rates due to the need to explore a larger parameter space of possible migration histories [25] . Here we found that MTT performed similarly well under different total migration rates , supporting the view that its new proposal functions represent a very considerable improvement over previous approaches [27] . We went on to assess the relative performance of the methods in a more realistic setting , when there is both phylogenetic signal and phylogenetic uncertainty ( the “variable tree” scenario ) . This scenario is more complex as phylogenetic uncertainty makes inference more computationally demanding . All three methods account for phylogenetic uncertainty by exploring possible trees using MCMC . Again we simulated under the Continental model , this time with a 2000 bp alignment , a mutation rate of 0 . 01 per base per Ne generations , 50 samples per subpopulation and a mean migration rate of f ‾ = 2 . 0 . All methods reported greater uncertainty in this setting , as expected , with DTA continuing to show weaker correlation between point estimates and the truth and severely underestimating posterior uncertainty compared to BASTA . While MTT most faithfully captured posterior uncertainty , it showed the worst correlation between point estimates and the truth , possibly reflecting a need to run it for longer than the other methods in the presence of phylogenetic uncertainty ( Fig . D in S1 Text and Table 2 ) . The over-confidence of phylogeographic inference made by DTA appears to affect analyses of real datasets as well as simulations . We compared the results of DTA and BASTA applied to a collection of Avian Influenza Virus ( AIV ) sequences sampled from different avian hosts [45] and a collection of Tomato Yellow Leaf Curl Virus ( TYLCV ) sequences [46] sampled from different locations worldwide . For both the AIV dataset ( Fig 5 ) and the TYLCV dataset ( Fig 6 ) , DTA reported very high confidence throughout the tree in the reconstructed ancestral subpopulations , representing host species and geographic location respectively . DTA reported posterior probabilities above 90% for ancestral reconstruction of most subpopulations ( 135 out of 145 internal nodes in Fig 6 and all 132 in Fig 5 ) , even deep within the tree . In contrast , BASTA placed high confidence on ancestral subpopulation reconstruction only for internal nodes close to samples , and only the minority had subpopulation posterior probability above 90% ( 63 out of 145 internal nodes in Fig 6 and 61 out of 132 in Fig 5 ) . Although we do not know the true host species and geographic locations of ancestors in these real datasets , the results of the simulations suggest that the high posterior probabilities reported by DTA could be poorly calibrated and overly confident , and that the results of BASTA are more reliable . So far , we have mostly considered scenarios with just two subpopulations , for which structured coalescent methods are expected to work in a reasonable time . However , with more populations , they may be too computationally demanding for practical inference . To compare the performance of BASTA to MTT in such a scenario , we simulated an Archipelago model with eight subpopulations arranged in two clusters of four islands , with 40 samples from each island . Migration between islands in the same archipelago was assumed fast ( mean 5 . 0 ) while migration between archipelagoes was 10-fold lower . To assist inference under MTT , we fixed the tree . Both methods reported considerable uncertainty in their estimates of the migration rates and root location ( Fig . G in S1 Text ) . However , for BASTA the MCMC algorithm reached convergence more quickly and more satisfactorily ( measured by the effective sample size , ESS > 200 [49] ) and in reasonable time ( 2 × 106 MCMC steps over 1 . 3 × 104 seconds per chain ) . With similar computational effort for MTT , the MCMC algorithm was far from convergence ( see e . g . Figs . E and F in S1 Text for some randomly sampled replicates ) with unsatisfactory estimates of the posterior distribution for most parameters ( ESS < 20 ) . These results show that not only does BASTA produce a modest but consistent improvement in calibration and statistical efficiency over MTT ( see also Table 3 ) but it has also broader applicability to scenarios with more populations . These results are important for the analysis of real datasets with more than just a few subpopulations , where BASTA currently offers the only practical alternative to DTA . The number of subpopulations in the AIV and TYLCV examples are moderately high , with 5–10 host species in the former [45] ( depending on pooling ) and eight global locations in the latter [46] . We found that this many subpopulations challenged or exceeded the range of applicability of MTT . In the analysis of the AIV dataset , MTT required a large number of MCMC iterations to achieve convergence ( Fig . I in S1 Text ) , while the analysis of the TYLCV data proved infeasible ( Fig . J in S1 Text ) . In contrast , we were able to run BASTA on both datasets in less than a day ( Figs 5 , 6 and Fig K in S1 Text ) . While we write , the most deadly known outbreak of Ebola virus is ongoing in West Africa . In recent work , Gire et al . [39] have collected and whole genome sequenced 99 Ebola virus samples from 78 patients . Using these and previous data , the authors have shown that all available sequences within each outbreak since 1976 cluster together phylogenetically; furthermore , divergence of lineages leading to different outbreaks usually considerably pre-dates the older outbreak . This fact and the shape of the inferred phylogeny suggest that independent zoonotic transmissions are the source of different Ebola outbreaks in humans . Ebola infections in different animals have been directly observed more than 50 times , with bats thought to be the main reservoir [50] . We addressed this subject in order to explore the potential impact of modelling considerations on epidemiological conclusions based on genetic data . We defined a highly simplified phylogeographic model with two subpopulations: the first representing human hosts , the second representing an animal reservoir . In this model , coalescence events within the human population originate from human-to-human transmission; similarly coalescence events in the animal reservoir originate from transmission between animal hosts . Migration from the animal reservoir to the human population corresponds to a zoonotic transmission . Migration from human to animal was assumed sufficiently rare to be ignored ( see [50] ) . Using this phylogeographic model , we investigated the effect of model choice—DTA versus structured coalescent—on the epidemiological conclusions concerning the role of zoonotic transmission in seeding human outbreaks of Ebola . We found that the two models gave diametrically opposed results . Consistent with general understanding of the emergence of Ebola outbreaks in humans , BASTA inferred that outbreaks were seeded by independent zoonosis events from the Ebola reservoir population ( Fig 7b ) . In keeping with this , the effective population size in the animal reservoir was inferred to be larger than in humans ( median of 29 . 4 times larger , with 95% CI [15 . 7 , 58 . 1] ) . The most recent common ancestor of all sampled human outbreaks was inferred to have originated in the animal reservoir population with 100% posterior probability . These results were also supported by MTT . In direct contrast , the DTA painted a very different picture of Ebola outbreak emergence that does not accord with scientific understanding . With high confidence , no zoonotic transmissions from animals to humans were inferred in the history of the sampled outbreaks ( 100% posterior probability , with the most recent common ancestor inferred to have occurred in the human population ( Fig 7a ) ) . Despite the implausibility of undetected human outbreaks having sustained Ebola virus in humans over four decades , DTA supported this scenario with high confidence . To test the robustness of this result , we performed a second analysis in which we incorporated limited available Ebola sequences from bats comprising seven 265 bp partial polymerase sequences [48] . By including animal samples it was necessary to permit a very low but non-zero rate of human-to-animal migration otherwise the ancestral location of lineages ancestral to the bats would be predetermined as occurring in the animal reservoir . With the addition of samples from bats , the results were largely consistent ( Fig . H in S1 Text ) : BASTA still inferred human outbreaks to be preceded by zoonotic transmission events from animals , with the root of the tree occurring in the animal reservoir with high probability ( 95% ) . DTA continued to erroneously infer that the majority of ancestral lineages occurred in the human population , but its confidence in this result was substantially reduced ( 59% ) . These results illustrate the strong influence of model choice on phylogeographic inference . They demonstrate the possibility of obtaining implausible results with DTA , which may be accompanied by high posterior probabilities . Although in the case of Ebola the strength of evidence concerning the epidemiology of the disease is more than sufficient to disregard the discrete trait analysis out of hand , it demonstrates the potential to produce highly misleading inference when independent epidemiological understanding is scarce .
Phylogeography has rapidly gained prominence in a wide range of settings where it can quantify historical patterns of migration from just genetic data and sampling locations . In the context of infectious disease epidemics , phylogeographic methods have been used to infer transmission rates and patterns of spread even in the complete absence of reliable epidemiological information ( see e . g . [9–11 , 51] ) . Yet , these methods have only been partially tested and compared . Here , through a combination of simulations based on explicit process-driven population genetics models and real data analysis , we showed that different methods exhibit dramatic differences in their inference properties , and these differences have a direct influence on biological interpretation . While discrete trait analysis ( DTA ) is extremely fast and accounts for phylogenetic uncertainty , it has difficulty accurately estimating migration rates even with as few as two subpopulations . In particular , DTA is sensitive to the relative sampling intensity of subpopulations , such that the sampling strategy adopted can influence the results , particularly when migration rates are high and genetic data are sparse . We reiterate that we have assessed the performance of DTA as a model of migration , and not in the context of the evolution of discrete traits ( such as genetic or phenotypic traits ) , for which DTA was originally developed . MTT , on the other hand , was robust to sampling strategy , produced less biased and less noisy parameter estimates , and produced well-calibrated reports of parameter uncertainty . Together with other methods based on the structured coalescent , MTT has the additional advantage over DTA of explicitly modelling , and therefore being able to estimate and account for , differences in the sizes of subpopulations . MTT proved useful even when migration rates were elevated , where previous structured coalescent-based methods showed convergence problems . However , we found that when moderately many subpopulations were analysed ( we simulated eight , but [27] suggest not to exceed four ) , MTT can suffer convergence issues . To deal with this problem , we proposed a new approach , BASTA , based on an approximation to the structured coalescent similar to those of [37] and [38] . BASTA approximately integrates over all possible migration histories rather than explicitly parameterizing them and exploring them with MCMC , thereby considerably reducing the computational requirements of the method . Not only did this approach show appreciable improvements in accuracy with respect to MTT with just two populations , but it was easily able to analyse eight subpopulations in 3–4 hours , whereas analysis of this many subpopulations was beyond the reach of MTT in feasible time . In the future , we will explore possible extensions of the model to cases with many demes , for example in patient-to-patient transmission inference , or in a stepping-stone island model [52] . In these scenarios , the technique of matrix exponentiation might prove too computationally demanding , and approaches based on shorter time subintervals , as in [37 , 38] , could be more efficient , particularly when the migration matrix is sparse . In applications to real AIV and TYLCV datasets , we showed that BASTA could be used in cases of up to ten sub-populations , where MTT struggles to converge . We found that DTA reported much more confidence—and on the basis of simulations , over-confidence—in the inferred reconstruction of ancestral subpopulations than BASTA , which simulations found to be well calibrated , indicating that the methods produce substantial differences of interpretation in phylogeography studies . Finally , analysing real data from Ebola outbreaks in humans we underlined the importance of model choice , by showing that different models can lead in practice to completely different results . In fact , diametrically opposite phylogeographic patterns were estimated using DTA versus structured coalescent-based methods . We recommend that users exercise caution in choosing phylogeographic models , and we point out that methods based on the structured coalescent are in general more reliable in modelling migration , although also more computationally demanding . The fact that the three approaches considered here are all implemented in the same phylogenetic package ( BEAST2 ) is a considerable advantage , as it is possible to run and compare different methods while installing a single piece of software and using similar formats . | When studying infectious diseases it is often important to understand how germs spread from location-to-location , person-to-person , or even one part of the body to another . Using phylogeographic methods , it is possible to recover the history of spread of pathogens ( or other organisms ) by studying their genetic material . Here we reveal that some popular , fast phylogeographic methods are inaccurate , and we introduce a new more reliable method to address the problem . By comparing different phylogeographic methods based on principled population models and fast alternatives , we found that different approaches can give diametrically opposed results , and we offer concrete examples in the context of the ongoing Ebola outbreak in West Africa and the world-wide outbreaks of Avian Influenza Virus and Tomato Yellow Leaf Curl Virus . We found that the most popular phylogeographic method often produces completely inaccurate conclusions . One of the reasons for its popularity has been its computational speed , which has allowed users to analyse large genetic datasets with complex models . More accurate approaches have until now been considerably slower , and therefore we propose a new method called BASTA that achieves good accuracy in a reasonable time . We are relying more and more on genetic sequencing to learn about the origin and spread of infections , and as this role continues to grow , it will be essential to use accurate phylogeographic methods when designing policies to prevent or curb the spread of disease . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | New Routes to Phylogeography: A Bayesian Structured Coalescent Approximation |
Artificial Intelligence is exponentially increasing its impact on healthcare . As deep learning is mastering computer vision tasks , its application to digital pathology is natural , with the promise of aiding in routine reporting and standardizing results across trials . Deep learning features inferred from digital pathology scans can improve validity and robustness of current clinico-pathological features , up to identifying novel histological patterns , e . g . , from tumor infiltrating lymphocytes . In this study , we examine the issue of evaluating accuracy of predictive models from deep learning features in digital pathology , as an hallmark of reproducibility . We introduce the DAPPER framework for validation based on a rigorous Data Analysis Plan derived from the FDA’s MAQC project , designed to analyze causes of variability in predictive biomarkers . We apply the framework on models that identify tissue of origin on 787 Whole Slide Images from the Genotype-Tissue Expression ( GTEx ) project . We test three different deep learning architectures ( VGG , ResNet , Inception ) as feature extractors and three classifiers ( a fully connected multilayer , Support Vector Machine and Random Forests ) and work with four datasets ( 5 , 10 , 20 or 30 classes ) , for a total of 53 , 000 tiles at 512 × 512 resolution . We analyze accuracy and feature stability of the machine learning classifiers , also demonstrating the need for diagnostic tests ( e . g . , random labels ) to identify selection bias and risks for reproducibility . Further , we use the deep features from the VGG model from GTEx on the KIMIA24 dataset for identification of slide of origin ( 24 classes ) to train a classifier on 1 , 060 annotated tiles and validated on 265 unseen ones . The DAPPER software , including its deep learning pipeline and the Histological Imaging—Newsy Tiles ( HINT ) benchmark dataset derived from GTEx , is released as a basis for standardization and validation initiatives in AI for digital pathology .
Artificial Intelligence ( AI ) methods for health data hold great promise but still have to deal with disease complexity: patient cohorts are most frequently an heterogeneous group of subtypes diverse for disease trajectories , with highly variable characteristics in terms of phenotypes ( e . g . bioimages in radiology or pathology ) , response to therapy , clinical course , thus a challenge for machine-learning based prognoses . Nevertheless , the increased availability of massive annotated medical data from health systems and a rapid progress of machine learning frameworks has led to high expectations about the impact of AI on challenging biomedical problems [1] . In particular , Deep Learning ( DL ) is now surpassing pattern recognition methods in the most complex medical images challenges such as those proposed by the Medical Image Computing & Computer Assisted Intervention conferences ( MICCAI , https://www . miccai2018 . org/en/WORKSHOP---CHALLENGE---TUTORIAL . html ) , and it is comparable to expert accuracy in the diagnosis of skin lesions [2] , classification of colon polyps [3 , 4] , ophthalmology [5] , radiomics [6] and other areas [7] . However , the reliable comparison of DL with other baseline ML models and human experts is not a diffuse practice yet [8] , and also the reproducibility and interpretation of the challenges’ outcome have been recently criticized [9 , 10] . DL refers to a class of machine learning methods that model high-level abstractions in data through the use of modular architectures , typically composed by multiple nonlinear transformations estimated by training procedures . Notably , deep learning architectures based on Convolutional Neural Networks ( CNNs ) hold state-of-the-art accuracy in numerous image classification tasks without prior feature selection . Further , intermediate steps in the pipeline of transformations implemented by CNNs or other deep learning architectures can provide a mapping ( embedding ) from the original feature space into a deep feature space . Of interest for medical diagnosis , deep features can be used for interpretation of the model and can be directly employed as inputs to other machine learning models . Deep learning methods have been applied to analysis of histological images for diagnosis and prognosis . Mobadersany and colleagues [11] combine in the Survival Convolutional Neural Network ( SCNN ) architecture a CNN with traditional survival models to learn survival-related patterns from histology images , predicting overall survival of patients diagnosed with gliomas . Predictive accuracy of SCNN is comparable with manual histologic grading by neuropathologists . Further , by incorporation of genomic variables for gliomas in the model , the extended model significantly outperforms the WHO paradigm based on genomic subtype and histologic grading . Similarly , deep learning models have been successfully applied to histology for colorectal cancer [12] , gastric cancer [13] , breast cancer [14] and lung cancer [15 , 16] . As human assessments of histology are subjective and hard to repeat , computational analysis of histology imaging within the information environment generated from a digital slide ( digital pathology ) and advances in scanning microscopes have already allowed pathologists to gain a much more effective diagnosis capability and dramatically reduce time for information sharing . Starting from the principle that underlying differences in the molecular expressions of the disease may manifest as tissue architecture and nuclear morphological alterations [17] , it is clear that automatic evaluation of disease aggressiveness level and patient subtyping has a key role aiding therapy in cancer and other diseases . Digital pathology is in particular a key tool for the immunotherapy approach , which stands on the characterization of tumor-infiltrating lymphocytes ( TILs ) [18] . Indeed , quantitative analysis of the immune microenvironment by histology is crucial for personalized treatment of cancer [19 , 20] , with high clinical utility of TILs assessment for risk prediction models , adjuvant , and neoadjuvant chemotherapy decisions , and for developing the potential of immunotherapy [21 , 22] . Digital pathology is thus a natural application domain for machine learning , with the promise of accelerating routine reporting and standardizing results across trials . Notably , deep learning features learned from digital pathology scans can improve validity and robustness of current clinico-pathological features , up to identifying novel histological patterns , e . g . from TILs . On the technical side , usually deep learning models for digital pathology are built upon imaging architectures originally aimed at tasks in other domains and trained on non-medical datasets . This is a foundational approach in machine learning , known as transfer learning . Given domain data and a network pretrained to classify on huge generic databases ( e . g . ImageNet , with over 14 million items and 20 thousand categories [23] ) , there are three basic options for transfer learning , i . e . to adapt the classifier to the new domain: a ) train a new machine learning model on the features preprocessed by the pretrained network from the domain data; b ) retrain only the deeper final layers ( the domain layers ) of the pretrained network; c ) retrain the whole network starting from the pretrained state . A consensus about the best strategy to use for medical images is still missing [24 , 25] . In this study we aim to address the issue of reproducibility and validation of machine learning models for digital pathology . Reproducibility is a paramount concern in biomarker research [26] , and in science in general [27] , with scientific communities , institutions , industry , and publishers struggling to foster adoption of best practices , with initiatives ranging from enhancing reproducibility of high-throughput technologies [28] to improving the overall reuse of scholarly data and analytics solutions ( e . g . the FAIR Data Principles [29] ) . As an example , the MAQC initiatives [30 , 31] , led by the US FDA , investigate best practices and causes of variability in the development of biomarkers and predictive classifiers from massive omics data ( e . g . microarrays , RNA-Seq or DNA-Seq data ) for precision medicine . The MAQC projects adopt a Data Analysis Plan ( DAP ) that forces bioinformatics teams to submit classification models , top features ranked for importance and performance estimates all built on training data only , before testing on unseen external validation data . The DAP approach is methodologically more robust than a simple cross validation ( CV ) [30] as the internal CV and model selection phase is replicated multiple times ( e . g . , 10 times ) to smooth the impact of a single training/test split; the performance metrics is thus evaluated on a much larger statistics . Also , features are analyzed and ranked multiple times , averaging the impact of a small round of partitions . The ranked feature lists are fused in a single ranked list using the Borda method [32] and the bootstrap method is applied to compute the confidence intervals . This approach helps mitigating the risk of selection bias in complex learning pipelines [33] , where the bias can stem in one of many preprocessing steps as well as in the downstream machine learning model . Further , it clarifies that increasing task difficulty is often linked to a decrease not only in accuracy measures but also of stability of the biomarker lists [32] , i . e . the consistency in the selection of the top discriminating features across all repeated cross validation runs . Although openness in sharing algorithms and benchmark data is a solid attitude of the machine learning community , the reliable estimation on a given training dataset of predictive accuracy and stability of deep learning models ( in terms of performance range as a function of variations of training data ) and the stability of deep features used by external models ( as the limited difference of top ranking variables selected by different models ) is still a gray area . The underlying risk is that of overfitting the training data , or worse to overfit the validation data if the labels are visible , which is typical when datasets are fully released at the end of a data science challenge on medical image data . As the number of DL-based studies in digital pathology is exponentially growing , we suggest that the progress of this field needs environments ( e . g . , DAPs ) to prevent such pitfalls , especially if features distilled by the network are used as radiomics biomarkers to inform medical decision . Further , given an appropriate DAP , alternative model choices should be benchmarked on publicly available datasets , as usual in the general computer vision domain ( e . g . , ImageNet [23] or COCO [34] ) . This study provides three main practical contributions to controlling for algorithmic bias and improving reproducibility of machine learning algorithms for digital pathology: We first apply DAPPER to a set of classification experiments on 787 Whole Slide Images ( WSIs ) from GTEx . The framework ( see Fig 1 ) is composed by ( A ) a preprocessing component to derive patches from WSIs; ( B ) a 3-step machine learning pipeline with a data augmentation preprocessor , a backend deep learning model , and an adapter extracting the deep features; ( C ) a downstream machine learning/deep learning head , i . e . the task specific predictor . In our experiments , we evaluate the accuracy and the feature stability in a multiclass setting for the combination of three different deep learning architectures , namely VGG , ResNet and Inception , used as feature extractors , and three classifiers , a fully connected multilayer network , Support Vector Machine ( SVM ) [36] and Random Forest ( RF ) [37] . This component is endowed with the DAP , i . e . , a 10 × 5 CV ( 5-fold cross validation iterated 10 times ) . The 50 internal validation sets are used to estimate a vector of metrics ( with confidence intervals ) that are then used for model selection . In the fourth component ( D ) we finally provide an unsupervised data analysis based on the UMAP projection method , and methods for feature exploration . The DAPPER software is available together with the Python scripts and the instructions to generate the HINT benchmark dataset as a collection of Jupyter notebooks at gitlab . fbk . eu/mpba-histology/dapper , released under the GNU General Public License v3 . Notably , the DAP estimates are provided in this paper only for the downstream machine learning/deep learning head in component ( C ) ; whenever computational resources are available , the DAP can be expanded also to component ( B ) . Here we kept as a separate problem the model selection exercise on the backend deep learning architecture in order to clarify the change of perspective with respect to optimization of machine learning models in the usual training-validation setting . As a second experiment , in order to study the DAPPER framework in a transfer learning condition , we use the deep features from the VGG model trained on a subset of HINT on the 1 , 300 annotated tiles of the KIMIA Path24 dataset [38] to identify in this case the slide of origin ( 24 classes ) . Previous work on classifying WSIs by means of neural networks was introduced by [38 , 39] , also with the purpose of distributing the two original datasets KIMIA Path960 ( KIMIA960 ) and KIMIA Path24 ( KIMIA24 ) . KIMIA24 consists of 24 WSIs chosen on purely visual distinctions . Babaie and coauthors [38] manually selected a total of 1 , 325 binary patches with 40% overlap . On this dataset , in addition to two models based on Local Binary Patterns ( LBP ) and Bag-of-Visual-Words ( BoVW ) , they applied two shallow CNNs , achieving at most 41 . 8% accuracy . On the other hand , KIMIA960 contains 960 histopathological images belonging to 20 different WSIs that , again on visual clues , were used to represent different texture/pattern/staining types . The very same experimental settings as the one for KIMIA24 , i . e . , LBP , BoVW and CNN , has been replicated on this dataset by Kumar and coauthors [39] . In particular , the authors applied AlexNet or VGG16 , both pretrained on ImageNet , to extract deep features; instead of a classifier , accuracy was established by computing similarity distances between the 4 , 096 features extracted . Also , Kieffer and coauthors in [25] explored the use of deep features from several pretrained structures on KIMIA24 , controlling for the impact of transfer learning and finding an advantage of pretrained networks against training from scratch . Conversely , Alhindi and coworkers [40] analyzed KIMIA960 for slide of origin ( 20 slides preselected by visual inspection ) , and similarly to our study they compared alternative classifiers as well as feature extraction models in a 3-fold CV setup . Considering the importance of clinical validation of predictive results [8] , we finally compared the performance of the DAPPER framework with an expert pathologist . DAPPER outperforms the pathologist in classifying tissues at tile level , while at WSI level performance are similar . DAPPER represents an advancement over previous studies , due to the DAP structure and its application to the large HINT dataset free of any visual preselection .
The images used to train the models were derived from the Genotype-Tissue Expression ( GTEx ) Study [35] . The study collects gene expression profiles and whole-slide images ( WSIs ) of 53 human tissues histologies used to investigate the relationship between genetic variation and tissue-specific gene expression in healthy individuals . To ensure that the collected tissues meet prescribed standard criteria , a Pathology Resource Center validated each sample origin , content , integrity and target tissue ( https://biospecimens . cancer . gov/resources/sops/ ) . After sectioning and Haemotoxylin and Eosin staining ( H&E ) , tissue samples were scanned using a digital whole slide imaging system ( Aperio ) and stored in . svs format [41] . A custom Python script was used to download 787 WSIs through the Biospecimen Research Database ( total size: 192 GB , average 22 WSIs for each tissue ) . The list of the downloaded WSIs is available in S1 Table . A data preprocessing pipeline was developed to prepare the WSIs as training data ( see Fig 2 ) . The WSIs have a resolution of 0 . 275 μm/pixel ( Magnification 40X ) and variable dimensions . Further , the region interested by the tissue is only a portion of the WSI and it varies across the samples . Hence first we identified the region of the tissue in the image ( see Fig 2 ) , then we extracted at most 100 tiles ( 512 × 512 pixel ) from the WSIs , by randomly sampling the tissue region . We applied the algorithm for the detection of the tissue region ( see Fig 2 ) on each tile and rejected those where the portion of the tissue was below 85% . A total number of 53 , 727 tiles was extracted , with a number of tiles per tissue varying between 59 ( for Adipose—Visceral ( Omentum ) ) and 2 , 689 ( for Heart—Left Ventricle ) . Four datasets ( HINT5 , HINT10 , HINT20 , HINT30 ) have been derived with increasing number of tissues for a total of 52 , 991 tiles; the full number of tiles per anatomical zone , for each dataset , is available in S2 Table and summarized in Table 1 . We refer to the four sets as the HINT collection , or the HINT dataset in brief . We choose the five tissues composing HINT5 based on exploratory experiments , while the other three datasets were composed including the tissues with higher number of tiles . The class imbalance is accounted for by weighting the error on predictions . In detail , the weight w of the class i used in the cross entropy function is computed as: wi = nmax/ni , where nmax is the number of tiles in the class with more tiles and ni is the number of tiles in the class i . Since image orientation should not be relevant for the tissue recognition , the tiles are randomly flipped ( horizontally and vertically ) and scaled , following a common practice in deep learning known as data augmentation . Data augmentation consists of different techniques ( such as cropping , flipping , rotating images ) performed each time a sample is loaded , so that the resulting input image is different at each epoch . Augmentation has proven effective in multiple problems , increasing the generalization capabilities of the network , preventing overfitting and improving models performance [42–44] . Such randomized transformations were found to provide more comparable performance between the prognostic accuracy of the deep learning SCNN architecture and that of standard models ( i . e . , Support Vector Machine , Random Forest ) based on combined molecular subtype and histologic grade [11] . In addition , each tile is cropped to a fixed size , which is dependent on the type of network used to extract the deep features . We exploited three backend architectures commonly used in computer vision tasks: These architectures have reached highest accuracy in multiclass classification problems over the last 4 years [48] and differ in resource utilization ( see Table 2 ) . The feature extraction layer of each backend network is obtained as the output of an end-to-end pipeline composed of the following main blocks ( see panel B in Fig 1 ) : The 1 , 000 Adapter features are then used as input for a classifier providing predicted tissue labels as output . As predictive models , we used a linear SVM with regularization parameter C set to 1 , a RF classifier with 500 trees ( both implemented in scikit-learn , v0 . 19 . 1 ) and a fully connected head ( FCH ) , namely a series of fully connected layers ( see panel C in Fig 1 ) . Inspired by [11] and [49] , our FCH consists of four dense layers with 1 , 000 , 1 , 000 , 256 and number of tissue classes nodes , respectively . The feature extraction block was initialized with the weights already trained on the ImageNet dataset [23] , provided by PyTorch ( v0 . 4 . 0 ) and frozen . The Adapter block is trained together with the FCH as a one network . Training also the weights of the feature extraction block improves accuracy ( see S3 Table ) . However , these results were not validated rigorously within the DAP and therefore they not are not claimed as generalized in this study . For the optimization of the other weights ( Adapter and FCH ) we used the Adam algorithm [50] with the learning rate set to 10−5 and fixed for the whole training . We used the cross entropy as the loss function , which is appropriate for multiclass models . The strategy to optimize the learning rate was selected based on results of a preparatory study with the VGG network and HINT5 . The strategy approach with fixed learning rate achieved the best results ( see S4 Table ) and was therefore adopted in the rest of the study . Following the rigorous model validation techniques proposed by the MAQC projects [30 , 31] , we adopted a DAP to assess the validity of the features extracted by the networks , namely a 10 × 5-fold cross validation ( CV ) schema . The input dataset is first partitioned in two separate datasets , the training set and the test set , also referred as external validation set as reported in [30 , 31] . The external validation set will be kept completely unseen to the model , and it will be only used in the very last step of the DAP for the final model evaluation . In our experimental settings , we used 80% of the total samples for the training set , and the remaining 20% for the external validation set . A stratification strategy upon the classes of tiles , i . e . , 5 , 10 , or 20 , has been adopted in the partitioning . The training set further undergoes a 5-fold CV iterated 10 times , resulting in 50 separated internal validation sets used for model evaluation within the DAP . The same stratification strategy is used in the creation of the folds . At each CV iteration , features are ranked by KBest , with ANOVA F-score as the scoring function [51] , and four separate models are trained on sets of increasing number of ranked features ( namely: 10% , 25% , 50% , 100% of the total number of features ) . A list of top-ranked features is obtained by Borda aggregation of the ranked lists , for which we also compute the Canberra stability with a computational framework designed for sets of ranked biomarker lists [32] . As for model evaluation , we considered the accuracy ( ACC ) , and the Matthews Correlation Coefficient ( MCC ) in their multiclass generalization [52–54]: ACC = ∑ k = 1 N C k k ∑ i , j = 1 N C i j , 0 ≤ ACC ≤ 1 ( 1 ) MCC = ∑ k , l , m = 1 N ( C k k C m l - C l k C k m ) ∑ k = 1 N [ ∑ l = 1 N C l k ∑ f , g = 1f ≠ k N C g f ] ∑ k = 1 N [ ∑ l = 1 N C k l ∑ f , g = 1 f ≠ k N C f g ] , - 1 ≤ MCC ≤ 1 ( 2 ) where N is the number of classes and Cst is the number of elements of true class s that have been predicted as class t . MCC is widely used in Machine Learning as a performance metric , especially for unbalanced sets , for which ACC can be misleading [55] . In particular , MCC gives an indication of prediction robustness among classes: MCC = 1 is perfect classification , MCC = −1 is extreme misclassification , and MCC = 0 corresponds to random prediction . Finally , the overall performance of the model is evaluated across all the iterations ( i . e . , internal validation sets ) , in terms of average MCC and ACC with 95% Studentized bootstrap confidence intervals ( CI ) [56] , and then on the external validation set . As a sanity check to avoid unwanted selection bias effects , the DAP is repeated stochastically scrambling the training set labels ( random labels mode ) or by randomly ranking features before building models ( random ranking mode: in presence of pools of highly correlated variables , top features can be interchanged with others , possibly of higher biological interest ) . In both modes , a procedure unaffected by selection bias should achieve an average MCC close to 0 . We designed a set of experiments reported in Table 3 to provide indications about the optimal architecture for deep feature extraction , while keeping fixed the other hyper-parameters . In particular we set batch size ( 32 ) and number of epochs ( 50 ) , large enough to let the network converge: we explored increasing numbers of epochs ( 10 , 30 , 50 , 100 ) and , since the loss stabilizes after about 35 epochs , we set the number of epochs to 50 . First , we compared the three backend architectures on the smallest dataset HINT5 , with fixed learning rate . Both VGG and ResNet architectures achieved good results , outperforming Inception as shown in Tables 4 and 5 . In successive analyses we thus restricted to use VGG and ResNet as feature extractors and validated performance and features with the DAP . The same process was adopted on HINT10 and HINT20 . An experiment with 30 tissues has also been performed . Results are listed in S5 Table . In the second experiment , we used VGG on the KIMIA24 dataset with the deep features extracted by VGG on GTEx; the task is the identification of the slide of origin ( 24 classes ) . In the DAPPER framework , classifiers were trained on 1 , 060 annotated tiles and validated on 265 unseen ones . In order to perform an unsupervised exploration of the features extracted by the Feature Extractor module , we projected the deep features onto a bi-dimensional space by using the Uniform Manifold Approximation and Projection ( UMAP ) multidimensional projection method . This dimension reduction technique , which relies on topological descriptors , has proven competitive with state-of-the-art visualization algorithms such as t-SNE [57] , preserving both global and local structure of the data [58 , 59] . We used the R umap package with following parameters: n_neighbors = 40 , min_dist = 0 . 01 , n_components = 2 , and Euclidean metric . All the code of the DAPPER framework is written in Python ( v3 . 6 ) and R ( v3 . 4 . 4 ) . In addition to the general scientific libraries for Python , the scripts for the creation and training of the networks are based on PyTorch; the backend networks are implemented in torchvision . The library for processing histological images ( available at gitlab . fbk . eu/mpba-histology/histolib ) is based on OpenSlide and scikit-image . The computations were performed on Microsoft Azure Virtual Machines with 4 NVIDIA K80 GPUs , 24 Intel Xeon E5-2690 cores and 256 GB RAM .
Regardless of difference in image types , VGG-KIMIA24 with both RF and SVM heads with ACC = 43 . 4% ( see Table 6 ) , improving on published results ( ACC = 41 . 8% [38] ) . It is worth noting that transfer learning from ImageNet to HINT restricts training to the Adapter and Fully Connected Head blocks . In one-shot experiments , MCC further improves when the whole feature extraction block is retrained ( see S3 Table ) . However , the result still needs to be consolidated by extending the DAP also to the training or retraining of the deep learning backend networks to check for actual generalization . The Canberra stability indicator was also computed for all the experiments , with minimal median stability for ResNet-20 ( Fig 5 ) . We evaluated the performance of DAPPER at WSI-level on the HINT20 external validation set , with the ResNet+SVM model . In particular , all the predictions for the tiles are aggregated by WSI , and the resulting tissue will be the most common one among those predicted on the corresponding tiles . However , it is worth noting that the number of tiles per WSI in the HINT20 external validation set varies ( min 1 , max 31 ) due to a stratification strategy only considering the tissues-per-sample distribution ( see Section Data Analysis Plan ) . Therefore , we restricted our evaluation to a subset of 15 WSI per class ( 300 WSI in total ) , each of which associated to 10 tiles randomly selected . This value represents a reasonable number of Regions of Interest ( ROIs ) a human pathologist would likely consider in his/her evaluations . In this regard , we further investigate how the DAPPER framework performs on an increasing number of tiles per WSI , namely 3 , 5 , 7 , and 10 . As expected , the overall accuracy improves as the number of tiles per WSI increases , reaching 98 . 3% when considering all 10 tiles per WSI . Notably , the accuracy is high even when reducing to 3 tiles per WSI ( see Table 7 ) . We tested the performance of DAPPER against an expert pathologist on about 25% of the HINT20 external validation set , 2 , 000 tiles out of 8 , 103 , with 100 randomly selected tiles for each class . We asked the pathologist to classify each tile by choosing among the 20 classes of the HINT20 dataset , without imposing any time constraint . The confusion matrix resulting from the evaluation of tiles as produced by the pathologist is shown in Fig 6 . Predictions produced by the DAPPER framework for comparative results are then collected on the same data . The best-performing model on the HINT20 dataset , namely the ResNet+SVM model , has been considered for this experiment . As reported in Table 8 , DAPPER outperforms the pathologist in the prediction of tissues at a tile-level . To provide an unbiased estimation of the performance of DAPPER , we repeated the same evaluation on 10 other randomly generated subsets of 2 , 000 tiles extracted from the HINT20 external validation set . The obtained average MCC and ACC with 95% CI are 0 . 786 ( 0 . 783 , 0 . 789 ) , and 79 . 6 ( 79 . 3 , 79 . 9 ) , respectively . Finally , since the classification at tile-level is an unusual task for a pathologist , who is instead trained on examining the whole context of a tissue scan , as a second task we asked the pathologist to classify 200 randomly chosen WSIs ( 10 for each class of HINT20 ) . As expected , the results in this case are better than those at tile-level , i . e . , MCC = 0 . 788 , and ACC = 79 . 5% , to be compared with the DAPPER performances reported in Table 7 . As a second contribution of this study , we are making available the HINT dataset , generated by the first component of tools in the DAPPER framework , as a benchmark dataset for validating machine learning models in digital pathology . The HINT dataset is currently composed of 53 , 727 tiles at 512 × 512 resolution , based on histology from GTEx . HINT can be easily expanded to over 78 , 000 tiles , as for this study we used a fraction of the GTEx images and at most 100 tiles from each WSI were extracted . Digital pathology still misses a universally adopted dataset to compare deep learning models as already established in vision ( e . g . , ImageNet for image classification , COCO for image and instance segmentation ) . Several initiatives for a “BioImageNet” will eventually improve this scenario . Histology data are available in the generalist repository Image Data Resource ( IDR ) [60 , 61] . Further , the International Immuno-Oncology Biomarker Working Group in Breast Cancer and the MAQC Society have launched a collaborative project to develop data resources and quality control schemes on Machine Learning algorithms to assess TILs in Breast Cancer . HINT is conceptually similar to KIMIA24 . However , HINT inherits from GTEx more variability in terms of sample characteristics , validation of donors and additional access to molecular data . Further , we used a random sampling approach to process tiles excluding background and minimize human intervention in the choice and preparation of the images . We applied an unsupervised projection on all the features extracted by VGG and ResNet networks on all tissues tasks . In the following , we discuss an example for features extracted by VGG on the HINT20 task , displayed as UMAP projection ( Fig 7 ) , points are coloured for 20 tissue labels . The UMAP displays for the other tasks are available in S1–S4 Figs . The UMAP display is in agreement with the count distributions in the confusion matrix ( Fig 4 ) . The deep learning embedding separates well a set of histology types , including Muscle-Skeletal , Spleen , Pancreas , Brain-Cortex and Cerebellum , Heart-Left Ventricle and Atrial Appendage which group into distinct clusters ( See Fig 7 and Table 9 ) . The distributions of the activations for the top-3 deep features of the VGG backend network on the HINT10 dataset are displayed in S5 Fig; the top ranked deep feature ( #668 ) is clearly selective for Spleen . The UMAP projection also shows an overlapping for tissues such as Ovary and Uterus , or Vagina and Esophagus-Mucosa , or the two Esophagus histotypes , consistently with the confusion matrix ( Fig 4 ) . Examples of five tiles from two well separated clusters , Muscle-Skeletal ( ACC: 93 . 4% ) and Spleen ( ACC: 94 . 6% ) , are displayed in panel A of Fig 8 . Tiles from three clusters partially overlapping in the neural embedding and mislabeled in both the VGG-20 and ResNet-20 embeddings with SVM ( Esophagus- Mucosa ACC = 53 . 2% , Esophagus-Muscularis ACC = 72 . 1% , Vagina ACC = 59 . 0% ) are similarly visualized in Fig 8B . While the aim of this paper is to introduce a framework for honest comparison of models that will be used for clinical purposes rather than fine-tuning accuracy in this experiment , it is evident that these tiles have morphologies that are hard to classify . This challenge requires more complex models ( e . g . ensembles ) and a structured output labeling , already applied in dermatology [2] . Further , we are exploring the combination of DAPPER with image analysis packages , such as HistomicsTK ( https://digitalslidearchive . github . io/HistomicsTK/ ) or CellProfiler [62] , to extract features useful for interpretation and feedback from pathologists .
Digital pathology would greatly benefit from the adoption of machine learning , shifting human assessment of histology to higher quality , non-repetitive tasks . Unfortunately , there is no fast , easy route to improve reproducibility of automated analysis . The adoption of the DAP clearly sets in a computational aggravation not usually considered for image processing exercises . However , this is an established practice with massive omics data [28] , and reproducibility by design can handle secondary results useful for diagnostics and for interpretation . We designed the DAPPER framework as a tool for evaluating accuracy and stability of deep learning models , currently only backend elements in a sequence of processing steps , and possibly in the future end-to-end solutions . We choose as test domain H&E stained WSIs for prediction of tissue of origin , which is not a primary task for trained pathologists , but a reasonable benchmark for machine learning methods . Also , we are aware that tissue classification is only a step in real digital pathology applications . Mobadersany and colleagues [11] used a deep learning classifier to score and visualize risk on the WSIs . Similarly , deep learning tile classification may be applied to quantify histological differences in association to a genomic pattern , e . g . , a specific mutation or a high-dimensional protein expression signature . In this vision , the attention to model selection supported by our framework is a prerequisite for developing novel AI algorithms for digital pathology , e . g . , for analytics over TILs . Although we are building on deep learning architectures known for applications on generic images , they adapted well to WSIs in combination with established machine learning models ( SVM , RF ) ; we expect that large scale bioimaging resources will give the chance of improving the characterization of deep features , as already emerged with the HINT dataset that we are providing as public resource . In this direction , we plan to release the network weights of the backend DAPPER models that are optimized for histopathology as alternative pretrained weights for digital pathology , similarly to those for the ImageNet dataset and available in torchvision . | In this study , we examine the issue of evaluating accuracy of predictive models from deep learning features in digital pathology , as an hallmark of reproducibility . It is indeed a top priority that reproducibility-by-design gets adopted as standard practice in building and validating AI methods in the healthcare domain . Here we introduce DAPPER , a first framework to evaluate deep features and classifiers in digital pathology , based on a rigorous data analysis plan originally developed in the FDA’s MAQC initiative for predictive biomarkers from massive omics data . We apply DAPPER on models trained to identify tissue of origin from the HINT benchmark dataset of 53 , 000 tiles from 787 Whole Slide Images in the Genotype-Tissue Expression ( GTEx ) project , available at the web address https://gtexportal . org . We analyze accuracy and feature stability of different deep learning architectures ( VGG , ResNet and Inception ) as feature extractors and classifiers ( a fully connected multilayer , Support Vector Machine and Random Forests ) on up to 20 classes . Further , we use the deep features from the VGG model ( trained on HINT ) on the 1 , 300 annotated tiles of the KIMIA24 dataset for identification of slide of origin ( 24 classes ) . The DAPPER software is available together with the scripts to generate the HINT benchmark dataset . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
]
| [
"medicine",
"and",
"health",
"sciences",
"medical",
"doctors",
"medical",
"personnel",
"biomarkers",
"health",
"care",
"histology",
"health",
"care",
"providers",
"artificial",
"intelligence",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"imaging",
"techniques",
"deep",
"learning",
"support",
"vector",
"machines",
"people",
"and",
"places",
"biochemistry",
"professions",
"anatomy",
"computer",
"architecture",
"pathologists",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"machine",
"learning"
]
| 2019 | Evaluating reproducibility of AI algorithms in digital pathology with DAPPER |
Humans and animals face decision tasks in an uncertain multi-agent environment where an agent's strategy may change in time due to the co-adaptation of others strategies . The neuronal substrate and the computational algorithms underlying such adaptive decision making , however , is largely unknown . We propose a population coding model of spiking neurons with a policy gradient procedure that successfully acquires optimal strategies for classical game-theoretical tasks . The suggested population reinforcement learning reproduces data from human behavioral experiments for the blackjack and the inspector game . It performs optimally according to a pure ( deterministic ) and mixed ( stochastic ) Nash equilibrium , respectively . In contrast , temporal-difference ( TD ) -learning , covariance-learning , and basic reinforcement learning fail to perform optimally for the stochastic strategy . Spike-based population reinforcement learning , shown to follow the stochastic reward gradient , is therefore a viable candidate to explain automated decision learning of a Nash equilibrium in two-player games .
Neuroeconomics is an interdisciplinary research field that tries to explain human decision making in neuronal terms . Behavioral outcomes are construed as results of brain activity and the neuronal correlates of the quantities relevant for the decision making process are identified . Humans , as economic agents , attempt to optimize some reward function by participating in the production , exchange and maintenance of goods . Reward for the individuals will depend in general not merely upon their own actions but also on those of the other players and , furthermore , these will adapt their own strategies . Classical models in neuroeconomics are based on temporal difference ( TD ) learning [1] , an algorithm to maximize the total expected reward [2] with potential neuronal implementations [3] , [4] . It assumes that the environment can be described as a Markov decision process ( MDP ) , i . e . by a finite number of states with fixed transition probabilities [5] . Multi-agent games , however , are not Markovian as the evolution of the environment typically does not only depend on the current state , but also on the history and on the adaptation of the other agents . Such games can be described as partially observable Markov decision processes ( POMDP , [6] ) by embedding the sequences and the learning strategies of the other agents into a large state space . We have presented a policy gradient method for population reinforcement learning which , unlike TD-learning , can cope with POMDPs and can be implemented in neuronal terms [7] . Yet , since a human learner would need to successfully explore the large state space of the POMDP , this appears to be an unrealistic scenario for explaining decision making in a multi-agent environment . A more realistic learning scenario is that humans transiently conceive the other players to follow a fixed strategy , and try to find their optimal counter strategy under this stationarity approximation . Maximizing one's own payoff while assuming stationarity in the opponents strategy is called a fictitious play and conditions are studied when this play effectively converges to a stationary ( Nash ) equilibrium [8] . Here we show that for classical two-player games [9] a simplified population reinforcement learning approach [7] , which is policy gradient under the stationarity approximation , can reproduce human data . We consider two games , blackjack [10] and the inspector game [11] , as examples for which the optimal strategy is either deterministic or stochastic , respectively . Optimality is expressed in terms of the Nash equilibrium , a solution concept for games involving two or more players . It is reached when no player has anything to gain by changing its strategy unilaterally . Each player is making the best decision it can , taking into account the decisions of the other ( s ) , hence the Nash equilibrium constitutes an optimum . Our algorithm is consistent with behavioral experiments for these games [10] , [11] while performing optimally according to the Nash equilibrium . We also show that TD-learning as well as covariance learning fail to find the stochastic Nash equilibrium for the inspector game . The current paper follows a long tradition of explaining human and animal behavior by simple models of reward-based learning , starting from Thorndike's law of effect [12] and Pavlovian conditioning paradigms [13] , [14] up to more recent theories of reinforcement learning [1] , [2] , [15] , [16] . Basic reinforcement learning with simple models of a few free parameters have also been applied to games . It has been shown for a wide set of two-player games that these simple algorithms well approximate human performance [17] . Yet , we show that basic reinforcement learning does not follow the reward gradient , and in fact it does not fit human data on the inspector game as well as our gradient rule . Obviously , playing games involves cognitive reasoning , as for instance captured by the theory of ‘adaptive control of thought–rational’ ( ACT-R , [18] ) . Within such a theory , our model represents a neuronal implementation of a ‘production rule’ which initiates a behavioral pattern in response to sensory and possibly cognitive input .
The network representing a player consists of a population of Spike-Response-Model ( SRM ) neurons with escape noise [19] , driven by a common presynaptic stimulus encoding the current state of the game ( the player's hand value in blackjack , and a fixed stimulus for the inspector game ) . Each input spike pattern ( ) is composed of afferent spike trains generated once by independent Poisson processes with duration , and then repeatedly presented with the same fixed spike timings . The population neurons integrate the afferent presynaptic input spike trains and produce an output spike pattern ( , see Fig . 1 ) . The decision of an individual postsynaptic neuron is denoted by , with , if the considered neuron does not spike , otherwise . Behavioral decisions are stochastically made based on the population activity defined as sum of the individual decisions across the population neurons: if is small , the population decision is likely , and the larger is , the more likely is . At the end of a game involving either a single decision ( like in the inspector game ) or a sequence of decisions ( like in blackjack ) , a reward signal is delivered by an external critic which either informs about winning or losing ( with like in blackjack ) or delivers a specific payoff ( like in the inspector game ) . The synapses feeding the stimulating spike pattern to the population neurons are updated according to a multi-factor plasticity rule involving the reward , the behavioral decision , the single neuron decision and the eligibility trace which depends on the post- and pre-synaptic activity: ( 1 ) Here , is the reward signal encoding the reward prediction error [20] ( see Eq . 5 in Methods ) , is the global feedback signal informing the synapses about the population decision weighted by the population activity ( Eq . 6 in Methods ) , and is the neuronal decision ( spike/no spike ) . The eligibility trace is a synaptic buffer roughly encoding the covariance between the past pre- and postsynaptic activity relevant for learning ( Eq . 8 in Methods ) . Technically , is the derivative of the log-likelihood of producing the postsynaptic spike train . The learning rule can be shown to perform gradient ascent in the expected reward ( Supporting Text S2 ) . While most of the terms in Eq . 1 may have their standard biological counterpart [16] , there is less experimental evidence for assigning the decision feedback ' to one specific neuromodulator . Yet , be the population neurons recurrently connected [21] or not , decision learning based on a population always requires that a global population signal is fed back to the individual neurons , as otherwise learning would quickly degrade with increasing population size [7] , [22] . By the same performance reasons it is not possible to replace the other factors ‘’ in Eq . 1 by a classical spike-timing dependent plasticity ( STDP ) implementation endowed with the multiplicative reward signal ‘’ [23] . In fact , reward-modulated STDP is only able to learn multiple stimulus-response associations when the reward factor averages out to zero for each stimulus individually , requiring an additional reward-prediction network [16] . Our neuronal implementation is as simple as possible to provide the required computational properties . The lack of feedback connectivity avoids issues relating to population spike correlations [24] , and the neural mechanisms supporting the readout of the decision and the population feedback signal are not considered here . Similarly , the fixed spike trains representing an input pattern is a biological simplification which does not fundamentally restrict the suggested approach . The simplified version of blackjack considered here was played in 18th century France and is the precursor of the version played in casinos nowadays . The card decks used consist of cards . Ace counts eleven , jack , queen and king ten points and the numbers two to ten according to their written value . The player ( gambler ) draws one card after the other , starting with an initial two card hand , with the object of bringing the hand value ( total across drawn cards ) as close as possible to , but stopping early enough so that it does not exceed this number , in which case he immediately loses . Afterwards the croupier does the same for the bank . The player wins if its score is higher than that of the croupier or if the croupier exceeds , otherwise the croupier wins . The winner's payoff is , the loser's . We assume that both player and croupier base their decision whether to draw another card or not only on their current hand value . Player and bank follow a strategy defined by the hand value and , respectively , from which on they stop to draw another card . The described rules of the game result in the payoff-matrix ( Table 1 ) comprising the average payoff of the bank as a function of the strategies and of the player and bank , respectively ( Methods ) . The gambler loses whatever the bank wins , therefore the game is an example of a zero sum game . For zero sum games a Nash equilibrium corresponds to a minimax solution [25] . If the pay-off matrix has a saddle point ( an entry which is the maximum in its row and the minimum in its column ) the corresponding strategy pair is a minimax solution which represents a pure Nash equilibrium . In blackjack there is a unique such pair , , and hence there is a unique Nash equilibrium at all . For this optimal strategy pair the gambler stops drawing another card as soon as he has points or more , while the croupier stops at or more . The entry represents the lowest loss for the gambler given the strategy of the bank ( minimum in the column ) , and the maximal payoff obtainable by the bank given the strategy of the gambler ( maximum in the row ) . The Nash equilibrium is asymmetric because in the case of a standoff ( equal final hand values ) the croupier always obtains reward and the player . For hand values smaller than it is safe to draw another card whereas for more than drawing another card leads to certain loss due to exceeding . While we do not model these trivial actions , we address the learning problem for hand values between and . The inspector game [27] has been widely studied in neuroeconomics [28] . The economic story surrounding the game is that a lazy employee prefers not to work . An employer knows this and sometimes ‘inspects’ , but has to pay some cost ‘’for inspection . The payoffs for employee and employer are shown in Table 2 . The inspector game shows only a mixed Nash equilibrium in which decisions are taken stochastically with a fixed probability . At the equilibrium , the players mix pure strategies , each with the same payoff: had these pure strategies different payoffs , then it would be better to just follow the pure strategy with the highest expected payoff . For each value of the inspection cost , there is a unique mixed Nash equilibrium in which the probability with which the employee is shirking just corresponds to the inspection cost , , and the probability with which the employer is inspecting is . In this case , neither player can improve its expected payoff by only unilaterally changing its strategy . In fact , using , the expected payoff for the employer is alwaysindependently of . Likewise , if , the expected payoff for the employee is always , independently of .
We considered a population of spiking neurons which represent an adaptive agent in a dynamic environment including other adaptive agents . The agent's adaptation was implemented as population reinforcement learning algorithm ( pRL ) which was previously shown to perform stochastic gradient ascent in the reward for partially observable Markov decision processes ( POMDPs ) [7] . Here we showed with blackjack and the inspector game that pRL can also cope with a dynamic multi-agent environment and that the performance is comparable to human data in both these games . In fact , when two neuronal populations play against each other , they learn to behave according to the optimal ( but unstable ) Nash equilibrium . By definition , no further increase in an agent's expected payoff is possible in the Nash equilibrium by only changing its own strategy while the environment remains stationary . In these steady-state conditions – where the opponent's strategy is assumed to be stationary – pRL is proven to maximize the expected reward ( Supporting Text S2 ) . The simulations show that the equilibrium is indeed reached by two pRL agents playing against each other , with a pure ( deterministic ) Nash equilibrium in blackjack and a mixed ( stochastic ) Nash equilibrium in the inspector game . As predicted by the theory [34] , [35] , the strategies oscillated around the mixed Nash equilibrium when both players used the same gradient algorithm based on the others stationarity assumption , i . e . when one network played against another both using pRL ( with a small learning rate ) . Averaging over long enough time windows , i . e . long compared to the oscillation period , yields the Nash equilibrium values . However , when implementing only the employee by a gradient pRL network and the employer by a non-gradient computer algorithm [11] , the two players do not play exactly equally well . In this case no oscillations occurred and both converged to and stayed at the optimal Nash equilibrium . For mathematical clarity we presented the spike-based pRL for an episodic learning scenario . But a biologically plausible implementation of a fully online scheme is also possible: to avoid an explicit separation of stimuli in time , the rectangular window function used to temporally integrate the eligibility trace ( Eq . 8 in Methods ) can be replaced by an exponentially decaying window function to get a low-pass filtered eligibility trace , and concentrations of neuromodulators can be used to encode feedback about the population decision and the global reward signal ( e . g . acetylcholine or dopamine ) [22] . We considered reward delivery immediately after stimulus presentation , but reward could also be substantially delayed when considering a further eligibility trace incorporating the population decision [7] . Moreover , since learning in general speeds up with population size ( up to 1-shot learning for stimulus-response associations [31] ) we expect that the convergence for pRL towards the Nash equilibrium can be much faster than in our example where parameters were fit to reproduce human data . The mixed Nash equilibrium represents a special case of Herrnstein's matching law [40] , according to which the number of times an action is chosen is proportional to the reward accumulated from choosing that action . This is true both for the pure and mixed Nash optimum . In the special case that the current reward only depends on the current action , but not on past actions , reward maximization always implies matching . ( In fact , if one action would yield a higher ( average ) payoff per choice , then this action must be chosen with probability 1 to maximize expected reward , and matching ( ) is trivially satisfied ( since for the non-chosen action ) . If both actions yield the same payoff per choice ( ) , then matching is again trivially satisfied . ) In turn , a reward-based learning rule which only empirically maximizes reward in this case leads to only an approximated matching [42] . Choice probabilities which maximize the expected reward are trivially also fixed points of any learning rule defined by the covariance between reward and neuronal activity . ( In fact , at the reward maximum there is no change in neuronal activity which , in average , would lead to an increase ( and in the opposite direction to a decrease ) of the expected reward , and hence the covariance between activity and reward must vanish . ) The other direction , again , is not true: a covariance-based rule does not necessarily lead to reward maximization or a Nash equilibrium [37] , [38] . Indeed , our simulations of the inspector game with the canonical covariance-based plasticity rules show that these rules do not necessarily lead to the mixed Nash equilibrium , but instead can result in deterministic ( non-Nash ) strategies . Similarly , basic reinforcement rules studied in the context of economics and human decision making [17] are neither compatible with the mixed Nash equilibrium for the inspector game . The performance of spike-based pRL is also superior to TD-learning [2] which is often discussed in the neuro-economical context [1] . With the parameter values for which TD-learners came closest to human data ( although without matching them as closely as pRL ) , the mixed Nash equilibrium in the inspector game was not reached within the long learning times . Instead , TD-learner first adopted a deterministic strategy , transiently switched their behavior , and swapped back to the same deterministic strategy . We attributed this mismatch to a general failing of TD-learning in correctly mapping action values to choice probabilities in probabilistic decision making tasks . TD-learning with the softmax choice policy , in particular , fails when matching of choice probabilities with average payoff is required [40] . Different generalizations have been considered to approach the shortcomings of algorithms in socio-economic games . TD-learning has been extended to not only assign values to its own decisions , but to pairs of own and opponent decisions . This enables the learning of minimax strategies where reward is maximized for the worst of the opponents actions [45] . While for zero-sum games minimax may realize a mixed Nash equilibrium , it results in a deterministic strategy in the inspector game: minimizing the maximal loss implies for the employee to always work ( to prevent being caught while shirking ) , and for the employer to always inspect ( to prevent undetected shirking ) . Another approach is to separately learn its own and the opponents action values and then calculate the Nash equilibrium [46] , but such explicit calculations do not seem to be the typical human behavior in socio-economic interactions . Instead , it is tempting to consider pRL with long eligibility traces which , as it performs policy gradient in POMDPs [7] , should find cooperative strategies with , on average , higher than Nash payoffs for all agents . For the inspector game such a co-operative strategy is that the employer should let the employee sporadically shirk ( say with probability ) without inspection , but with the common agreement that shirking will not prevail ( leading to average payoffs and for the employee and employer , respectively ) . Although under the specific experimental conditions of the inspector game humans did not show cooperation , they often do so in other game-theoretic paradigms , as e . g . in the prisoner's dilemma , and hence deviate from the Nash equilibrium ( for a review see [47] ) . It remains a challenge for future modeling work to capture such cooperative behavior . Likely , this will involve modeling the prediction of other player's reactions in response to ones own actions , as considered in the theory of mind [48] and as being a hallmark of successful socio-economic behavior . Given the difficulties of modeling genuine social behavior , and the difficulties humans effectively have in stacked reflexive reasoning , the assumption of the opponent's stationarity considered here appears as a reasonable approximation for decision making even in complex situations . In view of its success in matching behavioral and theoretical data we may ask how far human decision making is in fact determined by cognitive reasoning , or whether decisions should rather be attributed to automated neuronal processes steered e . g . by pRL ( which can also encompass input from a cognitive module as it is suggested for the production rules in the ACT-R theory , [18] ) . In fact , daily experience tells us that decisions are often more appropriate when we listen to our gut feeling , while we tend to merely add justifications post-hoc . Or put in Schopenhauer's words , “that in spite of all his resolutions and reflections he does not change his conduct” [49] .
Focusing on one neuron we denote by its input , which is a spike pattern made up of spike trains , and by its output spike train . The membrane potential can be written as ( 3 ) The postsynaptic kernel and the reset kernel vanish for . For they are given byFor the resting potential we use ( arbitrary units ) . Further , is used for the membrane time constant and for the synaptic time constant . Action potential generation is controlled by an instantaneous firing rate which increases with the membrane potential . So , at each point in time , the neuron fires with probability where represents an infinitesimal time window ( we use in the simulations ) . Our firing rate function iswith and ( parameter values taken from [29] , see also [19] ) . We consider a population of neurons and an input layer of size for each player that is represented by a neural net . We assume that each population neuron synapses onto a site in the input layer with probability of , leading to many shared input spike trains between the neurons . The population response is read out by the decision making unit based on a spike/no-spike code . We introduce the coding function , with , if neuron does not spike , otherwise . The population activity being read out by the decision making unit is:Note that such a formal summation could be implemented in terms of a neuronal integrator ( forming a ‘line attractor’ ) which continuously integrates excitatory and inhibitory input and keeps the neuronal activity at a constant level in the absence of input [50] . Using this activity readout , the behavioral decision is made probabilistically , with likelihood given by the logistic function ( 4 ) and being the counter probability . The normalization of the activity with ensures that , thus being of same order as the noise in the decision readout . We now describe the terms , modulating synaptic plasticity in Eq . ( 1 ) . The reward feedback encodes the reward prediction error , as observed in experiments [20] , ( 5 ) Here is a running mean estimate of the expected reward , , where we set . The parameter is the positive learning rate which , for notational convenience , we absorb into the reward signal . In all pRL simulations we used the value . Both values and ( rounded ) were chosen to minimize the Mean Squared Error ( MSE ) between the average model and human data ( for the shirk rate and for the employee's reward in the inspector game ) . All other parameter values were taken from [7] . The decision feedback is given by ( 6 ) which is the derivative of , see Eq . ( 4 ) , with respect to ; so decision feedback measures how sensitive the decision is to changes in activity . As shown in [29] , the probability density , , that a neuron actually produces the output spike train in response to the stimulus during a decision period lasting from to satisfies: ( 7 ) The derivative of with respect to the strength of synapse is known as characteristic eligibility in reinforcement learning [51] . For our choice of the firing rate function one obtains for the last term in ( 1 ) ( 8 ) In all the simulations initial values for the synaptic strength were picked from a Gaussian distribution with mean zero and standard deviation equal to , independently for each afferent and each neuron . In the Supporting Text S2 we show that the plasticity rule ( 1 ) composed of the factors ( 5 , 6 , 8 ) and the decision follows the stochastic gradient of the expected reward . For TD-learning we used the SARSA control algorithm [2] which estimates the values of state-action pairs . At each point in time , the value estimates are updated according toHere is similar to a learning rate and has values between and . is the reward immediately obtained after performing action . In the case of blackjack it is defined as zero if the game is not over and the player chooses to draw another card , otherwise it is determined by the payoffs of the considered game . When in state , the next action is chosen using softmax , i . e . according to the probability . In all simulations we used the rounded values and as they minimized the MSE between averaged model and human data ( for the shirk rate and for the employee's reward in the inspector game ) . Note that in both TD-learning and pRL we adapted the same number of free parameters ( TD: and ; pRL: and ) , making it possible to directly compare the quality of the fit . In both Roth-Erev models [17] the probabilistic choice rule is parametrized using propensities . The probability that a specific player ( who's index is omitted ) plays his th pure strategy is . In blackjack we assume an infinite number of card decks . Independently of the history , the drawing probability therefore remains constant , with a probability to draw a card with value being , and the probability to draw any other value from 2 to 11 being . For a strategy determined by the stopping value we calculated analytically the probability distribution of hand values after drawing the last card . The drawing process is iterated for those hand values that are smaller than until there is only probability mass on hand values greater than or equal to . Because the lowest card value on the desk remains always , drawing times in a row yields a lowest possible hand value of . Hence up to cards are drawn in order to obtain a hand value greater or equal to . Let us denote the value of the th card by and its probability distribution by , To obtain the probability distribution we sum up the probabilities of all possible combinations to draw cards that yield hand value , with the condition that the sum of the first drawn cards is smaller than , such that a th card is actually drawn under the stopping strategy . where II is the indicator function which is one if its argument is true and zero else . The product of the is the joint probability that the first card has value , the second and so on . The first indicator function ensures that all drawn cards sum up to , the second that cards are drawn , i . e . the sum of the first cards has to be smaller than the stopping value , because otherwise no further card would be drawn . For instance in the case of and one obtains the distributions in Table 3 . Denoting the hand value of the gambler by and that of the croupier by the payoff of the bank isAveraging of with respect to the joint distribution yields the entry in the average payoff matrix Table 1 for the strategy pair ( ) . For instance for ( ) , . We defined the drawing probabilities in Fig . 2 for a hand value at a certain game number as the frequency with which another card has been drawn upon the last presentations prior to of the corresponding stimulus . The evolution of the average reward in time in Fig . 2C are the low pass filtered reward sequences , where is the reward in the -th game and was used . The initial value was calculated assuming a random choice behavior prior to learning . The initial weights mimicking the prior strategy of instructed humans were obtained by training our network to make a decision with a certain probability . This is possible by adapting pRL to perform regression ( as will be published elsewhere ) . The evolution of the rates in time in Fig . 3E are the low pass filtered decision sequences , e . g . where if the employee shirks in trial , otherwise . We used a value of and assumed again an initial random choice behavior . The rate change in Fig . 3F was determine by binning the obtained time course of the rate into bins of width , calculating the mean of each bin , and the differences between succeeding bins . The result was further low pass filtered once more with an exponential running mean ( ) to reduce the noise . | Socio-economic interactions are captured in a game theoretic framework by multiple agents acting on a pool of goods to maximize their own reward . Neuroeconomics tries to explain the agent's behavior in neuronal terms . Classical models in neuroeconomics use temporal-difference ( TD ) -learning . This algorithm incrementally updates values of state-action pairs , and actions are selected according to a value-based policy . In contrast , policy gradient methods do not introduce values as intermediate steps , but directly derive an action selection policy which maximizes the total expected reward . We consider a decision making network consisting of a population of neurons which , upon presentation of a spatio-temporal spike pattern , encodes binary actions by the population output spike trains and a subsequent majority vote . The action selection policy is parametrized by the strengths of synapses projecting to the population neurons . A gradient learning rule is derived which modifies these synaptic strengths and which depends on four factors , the pre- and postsynaptic activities , the action and the reward . We show that for classical game-theoretical tasks our decision making network endowed with the four-factor learning rule leads to Nash-optimal action selections . It also mimics human decision learning for these same tasks . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"computational",
"neuroscience",
"biology",
"neuroscience"
]
| 2012 | Spike-based Decision Learning of Nash Equilibria in Two-Player Games |
Two-component signal transduction systems , where the phosphorylation state of a regulator protein is modulated by a sensor kinase , are common in bacteria and other microbes . In many of these systems , the sensor kinase is bifunctional catalyzing both , the phosphorylation and the dephosphorylation of the regulator protein in response to input signals . Previous studies have shown that systems with a bifunctional enzyme can adjust the phosphorylation level of the regulator protein independently of the total protein concentrations – a property known as concentration robustness . Here , I argue that two-component systems with a bifunctional enzyme may also exhibit ultrasensitivity if the input signal reciprocally affects multiple activities of the sensor kinase . To this end , I consider the case where an allosteric effector inhibits autophosphorylation and , concomitantly , activates the enzyme's phosphatase activity , as observed experimentally in the PhoQ/PhoP and NRII/NRI systems . A theoretical analysis reveals two operating regimes under steady state conditions depending on the effector affinity: If the affinity is low the system produces a graded response with respect to input signals and exhibits stimulus-dependent concentration robustness – consistent with previous experiments . In contrast , a high-affinity effector may generate ultrasensitivity by a similar mechanism as phosphorylation-dephosphorylation cycles with distinct converter enzymes . The occurrence of ultrasensitivity requires saturation of the sensor kinase's phosphatase activity , but is restricted to low effector concentrations , which suggests that this mode of operation might be employed for the detection and amplification of low abundant input signals . Interestingly , the same mechanism also applies to covalent modification cycles with a bifunctional converter enzyme , which suggests that reciprocal regulation , as a mechanism to generate ultrasensitivity , is not restricted to two-component systems , but may apply more generally to bifunctional enzyme systems .
To rationalize the occurrence of concentration robustness in the EnvZ/OmpR system of E . coli , Batchelor and Goulian proposed a simple mathematical model based on the three activities of the bifunctional EnvZ ( denoted by HK in Fig . 2B ) . Guided by the observation that the total OmpR concentration is much larger than that of EnvZ [23] ( ) they have argued that , in the limit , the steady state concentration of OmpR-P ( denoted by in Fig . 2B ) is determined by a quadratic equation [11] , which can be written in the form ( SI Text S1 ) ( 1 ) Here , denotes the total OmpR concentration , and the parameters and are proportional to the Michaelis-Menten constants associated with the phosphatase ( ) and phosphotransferase ( ) reactions . Note that Eq . ( 1 ) does not depend on the total EnvZ concentration ( ) . Hence , the Batchelor-Goulian model predicts that , in the limit , the concentration of OmpR-P is approximately independent of variations in the total concentration of the sensor kinase , i . e . [OmpR-P] exhibits ( concentration ) robustness with respect to changes in . Interestingly , Eq . ( 1 ) also predicts concentration robustness of with respect to the total concentration of the response regulator ( ) under certain conditions . To see this more explicitly , it is worth mentioning that a structurally similar equation has been analyzed previously in the context of concentration robustness for covalent modification cycles with a bifunctional converter enzyme [24] . This analysis has shown that the shape of the stimulus-response curve , described by Eq . ( 1 ) , depends on the relative magnitude between the two parameters and [18] . To this end , it is useful to consider two limiting cases corresponding to and . In the first case , the physiologically reasonable solution of Eq . ( 1 ) can be approximated by ( SI Text S1 ) ( 2 ) whereas , in the second case , one obtains the approximate solution ( 3 ) In any case , from the expressions in Eqs . ( 2 ) and ( 3 ) it is readily apparent that becomes independent of the total RR concentration if the latter is sufficiently large , i . e . if ( Eq . 2 ) or ( Eq . 3 ) . Hence , if , the parameter determines both , the threshold concentration beyond which becomes approximately constant as well as the value of that constant . In contrast , if , the predicted threshold concentration ( ) is much larger than the asymptotic phosphorylation level of the response regulator ( ) . Also , the approach to the asymptotic level is different for the two regimes: If , increases approximately linearly with up to the threshold ( Eq . 2 ) whereas , in the opposite case , it increases hyperbolically ( Eq . 3 ) . Due to the linear relationship between and in Eq . ( 2 ) the regime has been called ‘signal-transducing’ in Ref . [25] . Together , Eqs . ( 2 ) and ( 3 ) suggest that there exist two different regimes for the occurrence of concentration robustness and , as will be shown below , there is experimental evidence for either case . To test the predictions of their model , Batchelor and Goulian measured changes in the transcriptional activity of OmpR-controlled genes using a two-fluorescent reporter strain , which provided indirect evidence for concentration robustness of OmpR-P . Recently , Gao and Stock directly confirmed the predictions of the Batchelor-Goulian model in the PhoR/PhoB system using a Phos-tag based method allowing for a quantification of the PhoB-P levels as a function of total PhoB amounts [26] . Experiments were performed with the wild-type ( WT ) system as well as with a PhoB mutant ( ) which exhibits reduced interaction strength ( affinity ) with PhoR . Both measurements could be well described by Eq . ( 1 ) with a ratio varying between 0 . 1–0 . 2 ( Fig . 3A , solid lines ) . Overlaying the response curves with the respective values ( dotted lines ) indicates that the PhoR/PhoB system operates in the regime since the threshold concentration ( ) , beyond which PhoB-P becomes constant , is approximately equal to the value of that constant , as expected from Eq . ( 2 ) . The observed shift of the threshold concentration in the mutant strain results from the reduced affinity of which is associated with a larger value for . Since , increasing leads to an increased value of so that the asymptotically constant phosphorylation level of is reached at higher total PhoB concentrations , i . e . for total ( Fig . 3A ) . Concentration robustness has also been observed in the reconstituted NRII/NRI system of E . coli under in vitro conditions [27] . However , in that case the shape of the response curve is quite different ( Fig . 3B ) : The dependence between [NRI-P] and total [NRI] does not appear to be linear below the threshold concentration and the asymptotically constant phosphorylation level ( ) is only reached for very large values of total [NRI] ( ) . Together , this indicates that the NRII/NRI system operates in the regime and , indeed , fitting the measurement data to Eq . ( 3 ) supports this view ( Fig . 3B , solid line ) . Moreover , since in vivo concentrations of NRI are typically much lower than the threshold concentration of [28] it has been argued that , in the NRII/NRI system , concentration robustness will most likely not play a role under physiological conditions [27] .
To understand how ultrasensitivity may arise in TCSs with a bifunctional HK it will be helpful to analyze the consequences of reciprocal regulation in a related , but more simple system first . To this end , the reaction mechanism in Fig . 4A , which describes the reversible phosphorylation of a substrate by a bifunctional enzyme , is considered . The enzyme exhibits both , kinase ( ) and phosphatase ( ) activities , which catalyze the phosphorylation ( ) and dephosphorylation reactions ( ) , respectively . The transition between the two activity states is mediated through binding of an allosteric effector . For simplicity , it is assumed that has no phosphatase activity and , conversely , has no kinase activity so that effector-binding effectively inhibits the enzyme's kinase activity and , concomitantly , activates its phosphatase activity . Note that this system is similar to TCSs with a bifunctional sensor kinase where the autophosphorylation and phosphotransfer reactions are replaced by a covalent modification ( cf . Figs . 2C and 4A ) . Also , the bifunctional converter enzyme is assumed to have just a single catalytic site , which is supposed to mimic the fact that the phosphotransferase and phosphatase activities of the sensor kinase in TCSs are also likely to occur on a single catalytic site [17] . The dynamics of this system is described by the set of ordinary differential equations ( ODEs ) ( 4 ) together with the conservation relations for the total concentrations of substrate ( ) , converter enzyme ( ) and allosteric effector ( ) ( 5 ) ( 6 ) ( 7 ) If the substrate concentration is much larger than that of the converter enzyme ( ) , one can neglect the concentrations of the enzyme-substrate complexes ( since by Eq . 6 ) in the conservation relation for the substrate ( Eq . 5 ) , and the concentration of unmodified substrate can be expressed as ( 8 ) For later comparison , it will be useful to employ the quasi-steady state approximation ( QSSA ) in order to derive an effective equation for . By construction , the QSSA preserves the steady state structure of the underlying system [29] ( which is the main focus here ) although , for a better description of the transient dynamics , application of the total QSSA may be advantageous [30] . To apply the QSSA , it is assumed that , after a short transient period , the enzyme-substrate and the enzyme-effector complexes reach a quasi-steady state , defined by , and , which leads to the algebraic relations ( 9 ) Here , and denote Michaelis-Menten constants associated with the kinase and phosphatase activities , respectively , whereas denotes the dissociation constant for the enzyme-effector complex . Using the QSSA condition , it follows that ( 10 ) where Eqs . ( 4 ) , ( 8 ) and ( 9 ) have been used . In Eq . ( 10 ) , and have to be found as functions of from the conservation relations ( Eqs . 6 and 7 ) ( 11 ) ( 12 ) Intuitively , it is clear that if the effector concentration is sufficiently large ( ) the amount of effector that can be sequestered by the enzyme will be small since . Under this condition the conservation law for the effector ( Eqs . 7 and 12 ) always reduces to independent of whether the binding affinity of the effector is high ( if is small ) or low ( if is large ) . The latter only becomes important when the effector concentration is equal to or smaller than the enzyme concentration ( ) , e . g . under effector-limiting conditions . In the following , it will be shown that the type of effective equation , that is obtained from Eqs . ( 10 ) – ( 12 ) , depends on the ratio which may be regarded as a relative binding affinity for the enzyme-effector complex . The Batchelor-Goulian model is based on the three activities of the sensor kinase shown in Fig . 2B , i . e . it essentially focuses on the signal transduction layer in the general scheme for two-component signaling depicted in Fig . 1 . However , within the context of this model it may become difficult to predict the input-output behavior as a function of the input stimulus , especially if the latter affects multiple enzyme activities as observed in the PhoQ/PhoP/ and NRII/NRI/PII systems ( Fig . 2A ) . Guided by these examples the Batchelor-Goulian model will be extended by incorporating a mechanism that accounts for reciprocal regulation of the sensor kinase's autokinase and phosphatase activities by an allosteric effector . Analysis of this model shows that a low-affinity effector may lead to stimulus-dependent concentration robustness whereas a high-affinity effector may generate ultrasensitivity . In the latter case , the underlying mechanism is essentially the same as for covalent modification cycles ( cf . Fig . 4 ) . To implement reciprocal regulation it is assumed ( cf . Fig . 2C ) that , in the absence of the effector , the free form of the sensor kinase ( ) can undergo autophosphorylation and mediates the phosphotransfer to the response regulator ( step 1 and 2 ) , but does not exhibit phosphatase activity ( step 3 ) . The latter is assumed to be activated through effector-binding ( step 4 ) , so that the phosphatase activity is carried by the ligand-bound form of the sensor kinase . Since cannot undergo autophosphorylation ( and phosphotransfer ) binding of the ligand effectively leads to inhibition of the HK's autokinase activity and , concomitantly , activates its phosphatase activity . The dynamics of the extended model , as shown in Fig . 2C , is described by the five ODEs ( 22 ) ( 23 ) ( 24 ) ( 25 ) ( 26 ) together with the three conservation relations ( 27 ) ( 28 ) ( 29 ) where , and denote the total concentrations of response regulator , histidine kinase and effector , respectively . Measurements in the PhoQ/PhoP and NRII/NRI systems have shown that the ratio between the total concentrations of RR and HK is large ( ) [22] , [28] in which case one can use the simplified conservation relation ( cf . Eq . 8 ) ( 30 ) instead of Eq . ( 27 ) . Similar as in the case of covalent modification cycles ( Eqs . 10–12 ) , the steady state behavior of the system , described by Eqs . ( 22 ) – ( 30 ) , depends on the affinity of the allosteric effector ( ) relative to the total enzyme concentration ( ) . Note that for the derivation of Eqs . ( 22 ) – ( 30 ) it has been assumed that signal-sensing and the reactions describing the catalytic activities of the sensor kinase take place in the same compartment ( the cytoplasm of the cell ) . Hence , this model directly applies to cytosolic TCSs , such as the NRII/NRI system , but not to systems with a transmembrane sensor kinase where signal-sensing typically occurs in a different compartment . For example , in the PhoQ/PhoP system the sensor kinase PhoQ responds to changes of the concentration in the periplasm [20] . However , since effector-binding does not involve mass transfer the conditions for the occurrence of concentration robustness and ultrasensitivity are essentially the same ( up to a factor accounting for the different compartment volumes ) as those which are derived below on the basis of Eqs . ( 22 ) – ( 30 ) ( see Methods ) .
In many two-component systems , the phosphorylation level of the response regulator protein is modified by a bifunctional sensor kinase which , apart from exhibiting autokinase and phosphotransferase activity , also catalyzes the dephosphorylation of the response regulator through a phosphatase activity . In the present study , I have argued that the spectrum of potential input-output behaviors of such bifunctional systems does not only comprise graded responses [8]–[10] and concentration robustness [11] , [12] , but also ultrasensitivity as it is well-known from phosphorylation-dephosphorylation cycles with distinct converter enzymes [31] . To this end , I have proposed and analyzed an extension of the Batchelor-Goulian model [11] which considers the biologically motivated case where the autokinase and phosphatase activities of the sensor kinase are reciprocally regulated by an allosteric effector ( Fig . 2 ) . The analysis of the extended model showed that there exist two operating regimes under steady state conditions depending on the effector affinity: If the affinity is low compared to the total concentration of the sensor kinase ( ) the system produces a graded response to changes in the effector concentration ( Eqs . 33 and 34 ) and exhibits stimulus-dependent concentration robustness , which means that the maximal phosphorylation level of the response regulator does not only depend on kinetic model parameters ( as in the original Batchelor-Goulian model ) , but also on the effector concentration . Consistent with experiments in the PhoQ/PhoP system [22] , the extended model predicts an increase in the maximal phosphorylation level as the effector concentration is lowered ( Eq . 32 ) . However , if the effector affinity is sufficiently high ( ) the steady state equation for the extended model ( Eq . 35 ) becomes structurally identical to that for covalent modification cycles with distinct converter enzymes ( Eq . 19 ) so that ultrasensitivity may arise from the zero-order effect [31] . Apart from enzyme saturation due to the zero-order effect , sequestration of a signaling molecule into an inactive complex represents an alternative mechanism for the generation of ultrasensitivity in signal transduction networks [33]–[35] . Often , sequestration involves a reaction of the form ( 41 ) where , by definition , is sequestered by into the complex . In this sense , regulation of enzyme activities by an allosteric effector may also be regarded as a form of sequestration . In the case of reciprocal regulation shown in Fig . 2C , the enzyme-effector complex ( ) is not catalytically inactive , but rather has a different activity compared to the free form of the enzyme ( ) . Buchler and Louis have shown that the simple mechanism in Eq . ( 41 ) can give rise to ultrasensitivity in the concentrations of and if the stoichiometric binding parameter ( where ) exceeds unity , and the degree of ultrasensitivity increases as [36] . In the present study , the stoichiometric binding parameter ( ) plays a different role for the generation of ultrasensitivity since the condition does not guarantee the occurrence of ultrasensitivity per se , but only the validity of the reduced model , described by the steady state equation in Eq . ( 35 ) . To obtain ultrasensitivity within the reduced model , the ( apparent ) Michaelis-Menten constants for the phosphotransferase and phosphatase activities of the sensor kinase also have to be sufficiently small ( Eq . 37 ) , which distinguishes the mechanism , proposed in the present study , from purely sequestration-based mechanisms . Interestingly , the idea of reciprocal regulation , as a mechanism to generate ultrasensitivity , does not seem to be restricted to two-component systems as the same mechanism may also apply to covalent modification cycles with a bifunctional converter enzyme ( Fig . 4A ) . In both cases , reciprocal regulation may lead to ultrasensitivity if the stoichiometric binding parameters ( in the case of covalent modification cycles or in the case of two-component systems ) are sufficiently large . In this case , almost all free effector molecules are bound to the respective enzyme which leads to a tight partition of enzyme states into those with phosphatase activity and those with kinase activity ( cf . Eqs . 20 and 52 ) . As a consequence , the system behaves as if phosphorylation and dephosphorylation were catalyzed by independent enzyme subpopulations , which rationalizes why the corresponding steady state equations ( Eqs . 18 and 35 ) are structurally identical to that for covalent modification cycles with distinct converter enzymes ( Eq . 19 ) . However , this mechanism only ‘works’ as long as the enzyme is not saturated by the effector , which restricts the occurrence of ultrasensitivity to effector concentrations that are smaller than that of the respective enzyme ( Figs . 4B and 6A ) . To assess the potential relevance of reciprocal regulation for the occurrence of ultrasensitivity under physiological conditions one has to evaluate to what extent the requirements for its occurrence ( substrate excess , a large stoichiometric binding parameter and saturation of the sensor kinase's phosphatase activity ) are satisfied in a particular system in vivo . Based on measurements in the EnvZ/OmpR , PhoQ/PhoP and PhoR/PhoB systems , it seems that the requirement of substrate excess does not represent a limitation for the occurrence of ultrasensitivity as response regulator proteins are typically much more abundant than their respective sensor proteins [22] , [23] , [26] . In contrast , estimation of the stoichiometric binding parameter appears more difficult due to the limited knowledge on the range of input signals for a particular sensor kinase and their affinities relative to the total enzyme concentration . In general , histidine kinases may sense different signals ( such as ions , metabolites , small peptides or auxiliary proteins ) with widely different affinities [3] . Hence , it is conceivable that the same system produces a graded response with respect to a low-affinity effector and an ultrasensitive response with respect to another effector with a high affinity . For example , apart from mediating adaptation to -limiting conditions the PhoQ/PhoP system is also involved in the regulation of bacterial virulence . This transcriptional program is initiated by antimicrobial peptides that seem to bind to the same periplasmic site in the sensor domain of PhoQ as , but with a 100-fold higher affinity [37] , which could potentially shift the stoichiometric binding parameter into a regime where sigmoidal responses become possible . The occurrence of ultrasensitivity also requires saturation of the sensor kinase's phosphatase activity which means that the Michaelis-Menten constant , associated with that activity , has to be smaller than the total concentration of the response regulator . Measurements in the EnvZ/OmpR system have shown that the dissociation constant for the EnvZ-OmpR complex is 5-fold smaller than the total OmpR concentration which indicates that enzyme saturation is , in principle , possible under physiological conditions [23] . However , the occurrence of ultrasensitivity can also be compromised by a sufficiently strong , unregulated phosphatase activity which may arise from a basal phosphatase activity of the sensor kinase ( Fig . 7D ) or from an autophosphatase activity of the response regulator . The latter might explain why the NRII/NRI/PII system exhibits only a weak sensitivity with respect to changes in the effector ( PII ) concentration ( Fig . 8B ) . Alternatively , it has been speculated that the observed weak sensitivity results from a non-saturable phosphatase activity of NRII [27] which is consistent with the prediction that ultrasensitivity requires the phosphatase activity to operate in the zero-order regime ( Eq . 37 ) . On the other hand , it has been shown that single mutations in the dimerization domain of a sensor kinase can substantially affect its interaction strength with cognate and even non-cognate response regulator proteins [26] , [38] , which suggests that binding affinities between sensor kinases and response regulator proteins are highly evolvable . Hence , it is conceivable that one may employ directed evolution or site-directed mutagenesis to ‘adjust’ these binding affinities in a favorable range for ultrasensitivity to occur . In this sense , the results presented here may also guide the design of synthetic regulatory circuits which aim to implement ultrasensitive response behavior at the level of two-component systems [39] .
Under steady state conditions , the right-hand sides of Eqs . ( 22 ) – ( 26 ) are set to zero so that summation of Eqs . ( 22 ) and ( 26 ) readily yields ( 42 ) Similarly , summation of Eqs . ( 23 ) and ( 26 ) leads to the steady state relation ( 43 ) where denotes the dissociation constant for the enzyme-effector complex . From Eqs . ( 24 ) – ( 26 ) together with Eqs . ( 30 ) and ( 43 ) one obtains the expressions ( 44 ) where is defined by ( 45 ) whereas and denote Michaelis-Menten constants associated with the phosphotransferase and phosphatase activities of the sensor kinase , respectively . Using the expressions from Eqs . ( 43 ) and ( 44 ) in Eq . ( 42 ) and in the conservation relations , Eqs . ( 28 ) and ( 29 ) , yields the set of algebraic equations ( 46 ) and ( 47 ) ( 48 ) from which the steady state concentrations , and have to be found . Similar as in the case of a covalent modification cycle with a bifunctional enzyme the type of steady state solution , that is obtained from Eqs . ( 46 ) – ( 48 ) , depends on the affinity of the allosteric effector . If this affinity is low ( ) the concentration of free effector is approximately equal to the total effector concentration ( ) . Replacing by in Eq . ( 46 ) readily yields the quadratic equation in Eq . ( 31 ) with and defined in Eq . ( 32 ) . In contrast , if the affinity of the effector is sufficiently high ( ) the combination of Eqs . ( 47 ) and ( 48 ) yields a quadratic equation similar to that in Eq . ( 15 ) ( 49 ) where and denote the rescaled enzyme concentration and the relative binding affinity , respectively . In the limit , the solution of Eq . ( 49 ) can be approximated by [40] ( 50 ) With this approximation the concentration of free effector becomes ( cf . Eq . 48 ) ( 51 ) Using this expression for in Eq . ( 46 ) yields the equationwhich can be rewritten in the form shown in Eq . ( 35 ) of the main text . Similar to the case of covalent modification cycles it is straightforward to show ( using Eqs . 43–45 , 50 and 51 ) that a high-affinity effector leads to a partition of enzyme states according to ( cf . Eq . 20 ) ( 52 ) so that and may be regarded as apparent phosphatase and kinase concentrations , respectively . For TCSs with a transmembrane sensor kinase autophosphorylation , phosphotransfer and dephosphorylation occur in the cytosol whereas signal-sensing typically takes place in the periplasm ( for gram-negative bacteria ) or directly in the extracellular space ( Fig . 1 ) . Hence , a proper model would have to distinguish at least 3 compartments: The cytosol ( where the response regulator is located ) , the plasma membrane ( to which the sensor kinase is confined ) and the extracellular space ( where the effector is located ) . For gram-negative bacteria one would also have to consider a periplasmic compartment as many sensor kinases seem to respond to signals in the periplasmic rather than directly in the extracellular space [3] . Together , this makes it difficult to propose a generic model for TCSs that are regulated by non-cytosolic effectors which will , therefore , not be attempted here . Instead , to evaluate the impact of compartmentalization on the conditions for the occurrence of ultrasensitivity and concentration robustness it seems reasonable to consider ( as a first approximation ) a simplified model where the reactions describing the catalytic activities of the sensor kinase occur in the cytosol ( similar as assumed in the original Batchelor-Goulian model ) whereas binding of the effector to the regulatory site of the sensor kinase occurs either in the periplasm or in the extracellular space . Because effector-binding does not involve mass transfer between the extracellular space ( or the periplasm ) and the cytoplasm the equations for such a two-compartment model are essentially the same as those for a single compartment ( Eqs . 22–30 ) if the mass-balance equations are written in terms of average molecule numbers ( rather than concentrations ) . The corresponding ODE system then reads ( 53 ) where denotes the average amount of species ( measured in ) . Compared to Eqs . ( 22 ) – ( 26 ) the second-order rate constants , and are now measured in units of , i . e . they are independent of the volume of the compartment in which the corresponding reaction occurs . In contrast , first order rate constants ( , , , , , and ) have the same unit ( ) as before . Mass conservation is now expressed in terms of molecule number conservation for the total amount of response regulator ( ) , sensor kinase ( ) and effector ( ) as ( 54 ) Since the structure of Eqs . ( 53 ) and ( 54 ) is identical to that of Eqs . ( 22 ) – ( 29 ) it is clear that the conditions for the occurrence of concentration robustness and ultrasensitivity are identical in both cases if concentration-based quantities are replaced by their respective molar counterparts . Specifically , ultrasensitivity is predicted to occur if the amount of response regulator is much larger than that of the sensor kinase ( ) and if the affinity of the effector is sufficiently high . The latter condition is now expressed as ( 55 ) where the dissociation constant is measured in . Under these conditions , the steady state amount of phosphorylated response regulator is determined by the analog of Eq . ( 35 ) ( 56 ) where and are defined by the same expressions as in Eq . ( 36 ) . Similar as , the Michaelis-Menten constants and are measured in units of . Conversely , if the effector has a low affinity ( ) the steady state amount of is determined by the analog of Eq . ( 31 ) ( 57 ) where the rescaled Michaelis-Menten constants and are defined by the same expressions as in Eq . ( 31 ) . To analyze the impact of the compartment sizes on the input-output behavior one has to rewrite Eqs . ( 56 ) and ( 57 ) in terms of concentration-based quantities . For this purpose , the concentrations of the response regulator and that of the sensor kinase ( 58 ) are measured with respect to the cytosolic volume , whereas the effector concentration ( 59 ) is measured with respect to the extracellular ( or periplasmic ) volume . In the case of an extracellular effector , one may think of as the effective volume that is accessible to each cell in a population . In general , the effective volume decreases as the number of cells increases , e . g . due to cell growth . However , for the present purpose will be taken as a constant parameter . In addition , it is assumed that the extracellular space is a well-mixed compartment so that effector-diffusion can be neglected . Using the definitions in Eqs . ( 58 ) and ( 59 ) , Eqs . ( 56 ) and ( 57 ) can be written in the form ( 60 ) and ( 61 ) where ( 62 ) denotes the ratio between the cytosolic volume and that of the extracellular ( or periplasmic ) space . Also , in Eqs . ( 60 ) and ( 61 ) the dissociation constant and the Michaelis-Menten constants have been rescaled according to ( 63 ) which gives them the conventional unit . The rescaling is motivated by the fact that , in a concentration-based description of chemical reactions , second-order rate constants have to be proportional to the volume of the compartment in which the corresponding reaction occurs [41] , i . e . , and giving them units of . Similar as Eq . ( 35 ) , Eq . ( 60 ) predicts that ultrasensitivity may occur at low effector concentrations ( ) if the affinity of the effector is sufficiently high ( ) . The latter condition follows from Eq . ( 55 ) using that ( Eq . 63 ) and ( Eq . 58 ) . Hence , depending on the volume ratio the occurrence of ultrasensitivity may be favored ( if ) or suppressed ( if ) compared to a system that is regulated by a cytosolic effector ( for which ) . For example , if regulation occurs via a periplasmic effector may vary between 1 . 5 and 4 corresponding to a periplasmic volume fraction of 20–40% of the total cell volume [42] . In contrast , if regulation occurs via an extracellular effector the volume ratio may be substantially smaller than 1 ( ) ( especially at low cell densities ) which would make the condition less likely to hold and , therefore , suppress the occurrence of ultrasensitivity . Interestingly , Eq . ( 61 ) does not explicitly depend on the volume ratio . Hence , if reciprocal regulation occurs via a low-affinity extracellular effector ( ) the stimulus-response curves predicted by Eq . ( 61 ) are identical with those depicted in Fig . 5 if one replaces and by their extracellular ( or periplasmic ) counterparts and , respectively . The response curves in Fig . 7C and 7D have been generated using the following set of equations ( the corresponding reaction mechanism is shown in Fig . 7A and 7B ) ( 64 ) where , and have to be replaced using the conservation relations | Bacteria often use two-component systems to sense and respond to environmental changes , which involves autophosphorylation of a sensor kinase and phosphotransfer to a cognate response regulator . However , despite conservation of this ‘classical’ scheme there exist substantial variations in the mechanism of phosphotransfer among systems . Also , many sensor kinases exhibit phosphatase activity raising the question whether such a bifunctional architecture enables special regulatory properties in the response behavior to input signals . According to previous studies , classical two-component systems are unlikely to produce sigmoidal response curves ( ultrasensitivity ) if the sensor protein is bifunctional . Here , I argue that this is not necessarily true if the input stimulus ( allosteric effector ) reciprocally affects multiple activities of the sensor kinase , as it seems to be common for bifunctional enzymes . To this end , I propose and analyze an extension of the experimentally well-supported Batchelor-Goulian model which shows that ultrasensitivity requires a high-affinity effector and saturation of the phosphatase activity . The underlying mechanism involves sequestration of the effector by the sensor kinase which restricts the occurrence of ultrasensitivity to sufficiently low effector concentrations . Hence , this operating regime might be useful to sense effector limitations or to amplify weak input signals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"computer",
"and",
"information",
"sciences",
"systems",
"science",
"mathematics",
"network",
"analysis",
"regulatory",
"networks",
"biology",
"and",
"life",
"sciences",
"signaling",
"networks",
"computational",
"biology",
"nonlinear",
"dynamics",
"physical",
"sciences"
]
| 2014 | Reciprocal Regulation as a Source of Ultrasensitivity in Two-Component Systems with a Bifunctional Sensor Kinase |
Although acetylated α-tubulin is known to be a marker of stable microtubules in neurons , precise factors that regulate α-tubulin acetylation are , to date , largely unknown . Therefore , a genetic screen was employed in the nematode Caenorhabditis elegans that identified the Elongator complex as a possible regulator of α-tubulin acetylation . Detailed characterization of mutant animals revealed that the acetyltransferase activity of the Elongator is indeed required for correct acetylation of microtubules and for neuronal development . Moreover , the velocity of vesicles on microtubules was affected by mutations in Elongator . Elongator mutants also displayed defects in neurotransmitter levels . Furthermore , acetylation of α-tubulin was shown to act as a novel signal for the fine-tuning of microtubules dynamics by modulating α-tubulin turnover , which in turn affected neuronal shape . Given that mutations in the acetyltransferase subunit of the Elongator ( Elp3 ) and in a scaffold subunit ( Elp1 ) have previously been linked to human neurodegenerative diseases , namely Amyotrophic Lateral Sclerosis and Familial Dysautonomia respectively highlights the importance of this work and offers new insights to understand their etiology .
Microtubules ( MTs ) are polymers of α/β-tubulin heterodimers that associate head-to-tail and laterally to form hollow tubes . MTs are usually highly dynamic and undergo rapid turnover by exchange of subunits where the ends of MTs undergo transitions between growth and shrinkage . It has been postulated that this “dynamic instability” provides a space-probing mechanism that creates contact between MTs and target organelles [1] . However , the cytoplasmic network also contains a subpopulation of stable MTs [2] , [3] , [4] that have an undefined cellular function , possibly required for cellular morphogenesis [5] . A distinguishing and evolutionarily-conserved feature of stable MTs is that they acquire posttranslational modifications ( PTMs ) in a time-dependent manner [6] . An example of a PTM is the acetylation of α-tubulin on a lysine residue at position 40 . HDAC6 and SIRT2 have been shown to act as microtubule de-acetylases [7] , [8] , [9] . Elimination of acetylation has no obvious phenotypic consequences in Chlamydomonas or Tetrahymena , thus leading to the notion that α-tubulin acetylation is per se not required for cell survival . In Caenorhabditis elegans the expression of a non-acetylatable α-tubulin rescues the touch insensitivity phenotypes of neurons lacking MEC-12 , the only C . elegans α-tubulin that contains lysine at position 40 [10] , [11] , [12] . However , by using other systems , acetylation of α-tubulin has been proposed to control transport mechanisms such as e . g . the dynein/dynactin transport of aggresomes [13] , [14] or the selective transport of the Kinesin-1 cargo JIP1 [15] . Furthermore , MT acetylation may also be important for cell motility since overexpression of HDAC6 leads to decreased acetylation and increased cell motility , the opposite is true when HDAC6 is chemically inhibited [7] , [16] . In vertebrates , a family of GTPases ( Rac , Rho and Cdc42 ) triggers localized changes in actin and MT dynamics resulting in changes in cell motility [17] . Moreover , downstream effectors of these GTPases include the MT plus-end tracking proteins ( +TIPs ) that “capture” and “stabilize” the ends of MTs stimulating the detyrosination and acetylation of MTs in fibroblasts [18] , [19] , [20] . These data suggest that GTPases are involved in MT dynamics and PTMs . In C . elegans the Rac GTPase ced-10 and the Rac-like GTPase mig-2 have overlapping functions in axon guidance and cell migration [21] , but their role in MT function has not yet been studied . The Elongator , a component of the hyperphosphorylated holoenzyme RNA polymerase II ( RNAPII ) , was originally identified in yeast [22] . Significantly , one subunit of Elongator , Elp3 , harbors motifs found in the GNAT family of histone acetyltransferases ( HATs ) [23] . Both yeast and human Elongator have HAT activity in vitro , primarily directed toward histone H3 [24] , [25] , [26] , and yeast elp3 mutation results in decreased histone H3 acetylation levels in chromatin in vivo [24] , [27] . The Elongator is associated with nascent RNA that emanates from elongating RNAPII in yeast [28] and human cells [29] , [30] , and thus is classified as a key component of transcript elongation . However a substantial portion of Elongator is cytoplasmic [25] , [26] , [31] and may function as a scaffold involved in exocytosis [32] , tRNA modification [33] , activation of JNK [31] or actin dynamics [34] . The precise cytoplasmic function of the Elongator remains however controversial [35] . Recently , it was reported that the Elongator can acetylate α-tubulin in vitro [36] . This work uncovers that ( i ) elongator is a regulator of α-tubulin acetylation in vivo; ( ii ) elongator is important for MT function in correct loading and velocity of vesicles in vivo and ( iii ) acetylation has a novel function in fine-tuning intrinsic dynamics of MTs by modulating α-tubulin turnover . Altogether , this concept adds an additional layer of understanding explaining how acetylation by the Elongator can affect MT and neuronal function .
Using various “loss of function ( lf ) ” and “gain of function ( gf ) ” alleles of the C . elegans Rac GTPases rac-2 , mig-2 and ced-10 confirmed that lf and gf alleles of mig-2 hamper movement as revealed by reduction in the frequency of body bends in liquid ( Figure 1A , [21] , [37] ) . VD/DD motoneurons are located ventrally and send dorsal projections called “commissures” . Using unc-25::gfp as a marker for these motoneurons [21] , it was possible to quantify fully developed commissures in L1 larvae . Whilst wt and mig-2 ( lf ) each displayed an invariable number of 6 commissures ( with no differences in neuron morphology ) , commissures were significantly reduced in mig-2 ( gf ) ( Figure 1B and 1C ) . This supports previous reports [21] , [37] that mig-2 is required for correct body movement and that mig-2 ( gf ) exerts uncoordinated body movement ( unc ) and reduces commissures in VD/DD neurons . Rac GTPases may affect α-tubulin acetylation [19] , a notion that was assessed by western blot analyses and whole mount immunofluorescence in neurons with the monoclonal antibody 6-11B-1 that specifically recognizes MEC-12 , a neuronal α-tubulin that is acetylated on a single lysine residue at position 40 and that is expressed also in all motoneurons ( Figure S1A , S1B , S1C , S1D , S1E , S1F , S1G ) . Western blot analysis uncovered that acetylation of MEC-12/α-tubulin is absent in eggs but can be reliably detected in all other stages ( Figure 2B ) . As a control , we observed that acetylation of MEC-12 was absent in all stages within a mec-12 ( e1607 ) background , an allele previously described to be a putative null mutant ( data not shown , [10] ) . Whilst wt and mig-2 ( lf ) were indistinguishable from each other ( data not shown ) , there was a marked increase of MEC-12/α-tubulin acetylation in mig-2 ( gf ) , which was most pronounced in eggs and L1 stages ( Figure 2B asterisks , and Figure 2C; Figure S5A for details and quantification ) . Although the signals were strongest in touch neurons ( Figure 2C ) , acetylation was also shown to be present in other neurons ( e . g . motoneurons , [10] ) . In mec-3 ( e1338 ) , where the touch neurons are not specified [38] , the developmental acetylation pattern was analogous to wt , although slightly reduced due to the missing contribution of the touch neurons ( Figure S1H and S1J ) . This finding confirms that acetylation of MEC-12/α-tubulin is not limited to touch neurons . If the increase in acetylation of MEC-12/α-tubulin in mig-2 ( gf ) is central to the observed body movement defects and the neuronal phenotypes , a reduction of acetylation of MEC-12/α-tubulin in a mig-2 ( gf ) background should suppress the phenotypes . This was investigated by overexpressing HDAC-6 , an α-tubulin deacetylase [7] . Overall acetylation of MEC-12/α-tubulin was reduced in eggs and L1 stages of wt and mig-2 ( gf ) ( Figure 2B ) . Whilst the overexpression of HDAC-6 did not cause detectable phenotypic changes in a wt background , a suppression of the unc phenotype ( as scored by a body bends per minute assay , which simply sums the number of “left to right” and “right to left” movements of worms put into water [see also Methods] , Figure 1D ) and an amelioration of the neuronal morphology ( Figure 1E ) was apparent in the mig-2 ( gf ) background ( notably , the suppression of the morphological defects was not complete thus suggesting that mig-2 may also be involved in other pathways ) . This finding was shown to be specific to the regulation of acetylation of MEC-12/α-tubulin , as overexpression of HDAC-6 in a mec-12/α-tubulin mutant background abolished the suppression of the movement defects in mig-2 ( gf ) . In addition , a synthetic movement phenotype was observed in the mec-12/α-tubulin mutant strain overexpressing HDAC-6 ( Figure 1D ) . Hyperacetylation of MEC-12/α-tubulin in mig-2 ( gf ) can therefore be linked to the observed movement and neuronal phenotypes . Based on the observation that a reduction of MEC-12/α-tubulin acetylation in mig-2 ( gf ) partially ( but nevertheless clearly ) suppresses the movement defects , a whole genome RNAi suppression screen was performed to possibly identify regulators of α-tubulin acetylation ( data not shown ) . We found a series of suppressors , including elpc-1 , which upon RNAi distinctly improved the movement and neuronal phenotype of mig-2 ( gf ) ( Figure 2D and 2E ) in an allele unspecific manner ( since it suppressed the gf alleles: gm38 and gm103 , data not shown ) . ELPC-1 , the homologue of human IKAP/hELP1 , is predicted to be a subunit of the elongator complex , of which ( based on sequence homology ) 4 members can be identified in C . elegans: namely elpc-1 – 4 ( Figure S2A ) . The core complex is made up of ELPC-1 – 3 and ELPC-4 is a member of the accessory complex ( Figure 2A ) . Western blot analyses of the elongator mutants confirmed that the overall acetylation was reduced especially in L1 stage ( elpc-1 ) and L1–L4 stage ( elpc-3 ) when tested in a wt background ( Figure 2B blue asterisks , and Figure 2C ) and in eggs and L1 in the mig-2 ( gf ) background ( Figure 2B blue asterisks , and Figure 2C ) [for quantification of immunohistochemical data and Western blots see Figure S5] . Residual acetylation by a yet-to-be-defined pathway can be seen primarily in later stages of development . RNAi knock-downs of individual members of the elongator complex demonstrated that all four were able to suppress the movement phenotypes of mig-2 ( gf ) ( Figure 2D ) . Worms harboring a chromosomal deletion of elpc-1 and elpc-3 or a point mutation in elpc-3 produced by tilling ( Figure S2 and Figure S3 ) showed no gross neuronal or movement phenotype by themselves however , all three were able to suppress the mig-2 ( gf ) phenotypes ( Figure 2D and 2E ) . Interestingly , the double mutant elpc-1; elpc-3 did not enhance the suppression ( Figure 2D and 2E ) providing supporting evidence that elpc-1and elpc-3 are not involved in parallel pathways . A palette of translational rescue constructs ( Figure S3 ) revealed that elpc-1 and elpc-3 are expressed ubiquitously . The subcellular localization of ELPC-1 is mainly in the cytoplasm , whereas ELPC-3 is predominantly nuclear but also cytoplasmatic ( data not shown ) . This finding is analogous to the observations in human fibroblasts [25] . To test whether the core complex subunits also associate in C . elegans , co-immunoprecipitation assays ( CoIPs ) were performed with a specific polyclonal antibody directed against endogenous ELPC-1 . In lysates derived from elpc-3 deletion mutants expressing ELPC-3::GFP ( Figure S4 ) or from wt animals expressing GFP alone , ELPC-3::GFP associated specifically with endogenous ELPC-1 , while GFP alone did not ( Figure 2F ) . These results were further confirmed by co-precipitation of endogenous ELPC-1 or ELPC-3::GFP ( the reciprocal CoIPs ) using antibodies against GFP ( Figure 2G ) . We therefore conclude that the core elongator is conserved in worms . Besides the previously identified function in histone acetylation , this work also suggests that the elongator may be involved in the regulation of α-tubulin acetylation . Suppression of the mig-2 ( gf ) phenotype by the elongator was analyzed in detail by re-introducing rescue constructs as transgenes into the elongator; mig-2 ( gf ) double mutant background ( constructs described in Figure S3 ) . The translational gfp fusion construct of elpc-1 was able to rescue the elpc-1 effect in the mig-2 ( gf ) background ( Figure 3A ) . The overexpression of elpc-1 in a wt background had no effect ( Figure 3A ) . A similar observation was obtained with two minigene constructs , where the elpc-1 cDNA was placed under the control of two different neuronal promoters , namely unc-119 and rab-3 . This denotes that the suppression of mig-2 ( gf ) by elpc-1 is due to its function in neurons . Interestingly , a truncated version of elpc-1 ( which resembles the ELP1 mutation present in patients suffering from Familial Dysautonomia ( FD ) [39] , was not able to rescue ( Figure 3A ) even though its localization in the cytoplasm was unaffected by the truncation ( data not shown ) . Finally , the Prab-3::elpc-3 minigene in the respective elpc-3 ( lf ) ; mig-2 ( gf ) background resulted in a complete rescue , demonstrating that elpc-3 is required in neurons . elongator mutants should behave like strains overexpressing HDAC-6 , at least if acetylation is critical and therefore elongator; mec-12/α-tubulin double mutants should be equivalent to HDAC-6-overexpressing mec-12/α-tubulin strains . Indeed , two of the four mec-12/α-tubulin alleles tested ( u63 and u76 , Table S1 ) showed synthetic movement defects with elpc-1 ( lf ) ( Figure 3B and 3C ) . Only the alleles with altered structure of mec-12/α-tubulin showed a synthetic phenotype with elpc-1 ( lf ) . [This is due to the fact that mec-12/α-tubulin is accumulated in elpc-1 ( lf ) therefore compromising neuronal function ( the detailed analysis and the reasons of this synthetic phenotype are presented later in Figure 7 and Figure 8 ) ] . Moreover , all four mec-12/α-tubulin alleles completely abolished the suppression in the elpc-1 ( lf ) ; mig-2 ( gf ) double mutant ( Figure 3B and 3C ) thus highlighting that neurons require MEC-12/α-tubulin for elpc-1 to correctly exert its function . The suppression of mig-2 ( gf ) seems to require a precise equilibrium between MEC-12/α-tubulin and its acetylation . Interfering with this equilibrium by introducing point mutations into MEC-12/α-tubulin in fact abolishes the suppression . Moreover , a genetic interaction was observed in the elpc-3 ( lf ) ; mec-12/α-tubulin double mutant . Furthermore , the suppression of mig-2 ( gf ) by elpc-3 required wt mec-12/α-tubulin ( Figure 3D ) . Taken together , these findings reveal that the elongator is likely a regulator of neuronal function mediated by MEC-12/α-tubulin . The movement phenotypes analyzed here are independent from touch neuron function where mec-12/α-tubulin is required [10] . This is confirmed by the fact that elpc-1 can suppress mig-2 ( gf ) phenotypes in a mec-3 ( e1338 ) background , where the touch neurons are not specified ( Figure S1I ) . Moreover , elpc-1 ( lf ) and elpc-3 ( lf ) are normal in touch sensitivity ( Figure S4A ) and it has been shown that acetylation of mec-12/α-tubulin is not critical for touch sensitivity [10] . Besides neuronal function , mig-2 is also involved in vulva development and mig-2 ( gf ) alleles are egg laying defective ( egl ) [37] . In isolation , elongator alleles did not display notable egg laying defects and the elongator; mig-2 ( gf ) strains could not suppress the egl phenotype ( data not shown ) . Furthermore , it is known that mig-2 is required for correct distal tip migration and phagocytosis of apoptotic cell corpses in the gonad [21] , [40] . Various combinations of elongator , mig-2 and mec-12/α-tubulin mutants were tested , but no phenotypes were observed ( Table S2 and Table S3 ) . This suggests that elongator and mec-12/α-tubulin are not required for vulval development , distal tip migration or cell corpse phagocytosis . Endogenous ELPC-1 and α-tubulin co-immunoprecipitated in a wt and elpc-3 deleted background ( Figure 3E ) . This suggests that the ELPC-1/α-tubulin association is independent of ELPC-3 . The association between ELPC-1 and α-tubulin was further analyzed by expressing full length and FD truncated GFP tagged versions of ELPC-1 in elpc-1 ( lf ) mutants followed by the CoIP of α-tubulin . A similar result with α-tubulin was observed with both variants of ELPC-1::GFP ( Figure 3F ) . This is not surprising since FD-ELPC-1::GFP localized like the full length protein . ELPC-3::GFP was also shown to co-immunoprecipitated with α-tubulin ( Figure 3F ) , while GFP alone did not ( data not shown ) . Immunoprecipitation with 6-11B-1 , the antibody directed against acetylated α-tubulin , revealed that ELPC-1::GFP did not associate with the acetylated form of α-tubulin ( data not shown ) . This suggests that the affinity of ELPC-1 to α-tubulin is lost after acetylation . The observed genetic interaction between elongator and mec-12/α-tubulin , the physical association of the core Elongator and α-tubulin and the fact that MEC-12/α-tubulin is less acetylated in elongator mutants suggest that elongator plays a role in α-tubulin acetylation . To further analyze the role of acetylation of mec-12/α-tubulin in neurons of mig-2 ( gf ) animals we expressed gfp::mec-12/α-tubulin mutated at lysine 40 ( K40Q ) . Overexpressing a non acetylable gfp::mec-12/α-tubulin in mig-2 ( gf ) should out-compete the overacetylated endogenous mec-12/α-tubulin and therefore reduce the overall acetylation . This should suppress the mig-2 ( gf ) phenotypes . The results shown in Figure 4A show how expression of the gfp::mec-12/α-tubulin K40Q transgene suppresses the mig-2 ( gf ) phenotypes in a dose dependent manner . High doses are deleterious , as was also seen with an integrated version of mec-12/α-tubulin K40Q used in Figure 4B and Figure 7A and 7C ( data not shown ) . elongator mutants should behave like strains overexpressing HDAC-6 , at least if acetylation is critical and therefore elongator; mec-12/α-tubulin double mutants should be equivalent to HDAC-6-overexpressing mec-12/α-tubulin strains . Indeed , two of the four mec-12/α-tubulin alleles tested ( u63 and u76 , Table S1 ) showed synthetic movement defects with elpc-1 ( lf ) ( Figure 3B and 3C ) . Only the alleles with altered structure of mec-12/α-tubulin showed a synthetic phenotype with elpc-1 ( lf ) . [This is due to the fact that mec-12/α-tubulin is accumulated in elpc-1 ( lf ) therefore compromising neuronal function ( the detailed analysis and the reasons of this synthetic phenotype are presented later in Figure 7 and Figure 8 ) ] . The dose of gfp::mec-12/α-tubulin K40Q required for suppression of mig-2 ( gf ) in a mig-2 ( gf ) ;mec-12 ( e1607 ) double background appears to be higher than in a single mig-2 ( gf ) mutant . We hypothesize that the difference observed in the two backgrounds is due to negative effects of MEC-12/α-tubulin overexpression . In the mig-2 ( gm38 ) background , the total levels of functional MEC-12/α-tubulin protein are higher , because of the concomitant presence of the endogenous mec-12 gene . For this reason the mig-2 ( gm38 ) worms reach a threshold of toxicity at lower mec-12/a-tubulin array expression levels . The general reduction of movement seen in the left panel in mig-2 ( gm38 ) is probably due to negative effects of the array in this background . Moreover , the expression of gfp::mec-12/α-tubulin K40Q also suppressed the defects in commissure formation ( Figure 4B ) . This indicates a possible role of mec-12/α-tubulin acetylation during neuronal development . We conclude that acetylation on MEC-12 K40 is an important ( but probably not the sole ) aspect of the elongator phenotype . To study the cytoplasmic function of the Elongator , neuronally expressed elpc-3 were tagged with the well characterized nuclear localization signal [NLS] ( to localize elpc-3 into the nucleus ) or with a nuclear export signal [NES] ( in order to direct the expression mainly in the cytoplasm ) . No well-established NES has been characterized in worms yet . NESs are leucine-rich sequences with an evolutionary conserved consensus LX1–3 LX2–3 LXL [41] . For the purpose of this study the well known NES of HIV Rev protein was chosen . If elpc-3 is required in the cytoplasm to exert its function with mec-12/α-tubulin then only the NES version , but not the NLS version , should be able to rescue . The correct nuclear and/or cytoplasmic localization was assayed by microscopy since both construct are translational gfp fusions and expressed in predicted subcellular compartment ( data not shown ) . Figure 4C shows the body bends assays confirming the hypothesis that NES::ELP-3 was able to rescue the suppression in the mig-2 ( gf ) ; elp-3 ( lf ) background , whereas NLS::ELP-3 was not . To further confirm the cytoplasmic requirement of the Elongator in our system we carried out body bends assays of worms treated with Leptomycin A and B . Leptomycin B has previously been shown to be an important tool in the study of nuclear-cytoplamic transport in C . elegans [42] . The molecular composition of Leptomycin A is very similar to Leptomycin B and both were used to substantiate the results . We hypothesized that treating mig-2 ( gf ) animals with Leptomycin would sequester the Elongator to the nucleus and therefore reduce the acetylation of mec-12/α-tubulin , which in turn should suppress the body bends defects of mig-2 ( gf ) . Worms were raised in the presence of the drugs , so ensuring that possible developmental effects were not excluded when the locomotion tests were performed ( at the young adult stage ) . The correct nuclear localization of elpc-3 upon drug treatment was assayed by microscopy using a transgenic strain bearing a translational rescuing elpc-3::gfp fusion which correctly localized to the nucleus in worms treated with Leptomycin ( data not shown ) . The movement data obtained not only confirmed the hypothesis ( Figure 4D ) , but also show that the suppression of mig-2 ( gf ) is dependent on mec-12/α-tubulin , since the suppression was abolished in a mig-2 ( gf ) ; mec-12/α-tubulin double mutant . Rescue experiments were performed using neuronally expressed elpc-3::gfp in elpc-3 ( lf ) ; mig-2 ( gf ) double mutants . As seen in Figure 4D , the elpc-3::gfp construct is functional . Furthermore , this experiment confirmed the cytoplasmic requirement of the Elongator , since the rescue was progressively lost by increasing the concentration of Leptomycin A ( Figure 4D ) . The effect of Leptomycin A was persistently higher than Leptomycin B , which was toxic at high concentrations ( data not shown ) . It has been shown that mig-2 is important for cell migration and axon pathfinding in the CAN neuron ( Lundquist et al and references therein [21] ) . Since the body sizes of elongator mutants and wt worms are identical , the position of the CAN cell body was measured to assay cell migration and posterior process outgrowth in L1 larvae ( Figure 3G ) . In elpc-1 and elpc-3 mutants the CAN cell migrated less and the posterior process was shorter than in wt . In the elpc-1; elpc-3 double mutant the phenotypes were not enhanced , indicating that elpc-1 and elpc-3 are not involved in parallel pathways ( Figure 3G ) . No misguidance or branching phenotypes were observed . Notably , elongator mutants were able to ameliorate the CAN cell migration defect , but not the posterior process outgrowth of mig-2 ( gf ) ( Figure 3G ) . These results suggest that the elongator is required for correct neuronal migration and axonal extension . Acetylation of MTs has been proposed to be important for vesicle transport [15] . In C . elegans a unique tool to measure vesicle velocities exists: namely IDA-1::GFP which marks the dense core vesicles in some neurons [43] . The velocity of dense core vesicles ( DCVs ) along MT tracks was measured in the ALA lateral process of living worms using a fast time-lapse method [43] ( Figure 5A ) . If the acetylation of MT positively regulates DCV velocity , then a reduced function of the elongator should slow down the vesicles . The mean velocity of vesicles measured in wt animals reflected the published data for wt animals ( 2 , 25 µm/sec . vs 2 , 10 µm/sec . [43] ) . Indeed , reducing the function of the elongator diminished the average velocity of DCVs by about 35% of wt young adults ( Figure 5B ) . In contrast , the mean velocity of mig-2 ( gf ) mutant was increased by about 13% . In addition , the phenotype of mig-2 ( gf ) was suppressed in the mig-2 ( gf ) ; elpc-1 ( RNAi ) double mutant . Analogous to this , the increased velocity of mig-2 ( gf ) was suppressed by mec-12 ( u76 ) , even though mec-12 ( u76 ) on its own had no velocity phenotype ( Figure 5B ) . Moreover mec-12 ( e1607 ) in which mec-12 is not acetylated showed a marked reduction of DCV velocity ( Figure 5B ) . These experiments confirm that a pathway with mig-2 , elongator and mec-12 affects vesicle transport in vivo . It is known that transport of DCV relies , in part , on the same mechanisms as the transport of clear core vesicles [43] . Clear core vesicles transport neurotransmitters are important for movement ( see below ) , however , tools to measure the transport of clear core vesicles directly are not available yet . Since MT are also acetylated in motoneurons , acetylation may play a role in regulating transport of clear core vesicles and therefore also movement . In C . elegans movement is controlled by the major neurotransmitter acetylcholine ( ACh ) . Aldicarb , a potent ACh esterase inhibitor that potentiates ACh response , can be used to test for changes in ACh concentration at neuromuscular junctions ( NMJs ) . In a standard test , which utilizes different concentrations of aldicarb , the sensitivity of adult elongator mutants and wt were identical ( [44]; and data not shown ) . However , a differential sensitivity towards aldicarb was observed when exposure time was reduced from 4 hours to 1 hour . Compared to wt elpc-1 , elpc-3 and the elongator double mutant less sensitive to aldicarb ( Figure 6A–6C ) . The double mutant was indistinguishable from single mutants ( Figure 6C ) . All elongator mutants suppressed the aldicarb phenotype of mig-2 ( gf ) mutants ( Figure 6D–6F ) . The sensitivity of mig-2 ( gf ) was not as pronounced as in goa-1 ( n363 ) [a classical hypersensitive mutant [45] that displays a strong phenotype ( Figure 6G ) ] , this might at least in part be due to developmental defects . The resistance to aldicarb of elpc-1 mutants was comparable to the syntaxin mutant unc-64 ( e264 ) ( Figure 6G ) . elpc-1 ( lf ) could be rescued by re-expressing elpc-1 in neurons ( Figure 6G ) . This proves that pre-synaptic elpc-1 is required to ensure correct ACh concentrations at NMJs . The reason for the reduction of ACh at NMJs might in part be due to reduction of velocity of vesicles . On the other hand the hypersensitive phenotype of mig-2 ( gm38 ) might be due to the overacetylation . Although this is a possible explanation it has to be considered that developmental defects in mig-2 ( gm38 ) might further influence the phenotype ( e . g . by wrong NMJ wiring ) . Since no method exists to date to directly measure the ACh flow towards the NMJs other functions than acetylation may be additionally responsible for the mig-2 ( gf ) phenotype in these tests . Using levamisole , an agonist of the nicotinic ACh receptor [44] , it was possible to assess neurotransmission and function of the postsynaptic apparatus in elongator mutants . The response of adult elongator mutants and mig-2 ( gf ) animals was statistically equal to wt but clearly different to unc-29 ( e1072 ) , a mutant with a defective nicotinic ACh receptor ( Figure 6H ) . These data underline the notion that the elongator and the mig-2 ( gf ) mutants regulate ACh at NMJs due to presynaptic defects . Interestingly the degree of sensitivity to aldicarb correlated with the level of acetylation of the MTs . Whether changes of ACh levels at NMJs is a result of the direct or indirect regulation of acetylation of microtubules and changes of vesicle behavior remains to be analyzed in other systems . To verify that the change of ACh at NMJs is not due to a defect in ACh synthesis , a chromatographic method was devised capable of measuring ACh concentrations directly in worm extracts . The amount of ACh per adult worm was determined to approximate 3 fmol , a concentration statistically invariable in the elongator mutants ( Figure 6I ) . Therefore the phenotypes are unlikely a result of differences in presynaptic ACh production . Even though elongator and mec-12/a-tubulin were shown to interact genetically ( Figure 3 ) , DNA microarray and RT-PCR of elpc-1 ( ng10 ) and wt revealed no differential regulation of mec-12/α-tubulin transcripts ( data not shown ) . The levels of MEC-12/α-tubulin were analysed via a functional GFP-tagged version ( to date a specific antibody to endogenous MEC-12 has not been raised ) . Surprisingly , the amount of GFP::MEC-12 protein was found to be inversely proportional to the degree of acetylation ( the cell body of the touch neuron ALM was chosen to measure the GFP signal in L1 of all genetic backgrounds ) . In detail , compared to wt , the GFP signal of elongator and mig-2 ( gf ) mutants was about 20% higher and 20% lower , respectively ( Figure 7A ) . The reduced level of GFP::MEC-12 signal observed in mig-2 ( gf ) mutants could be suppressed by crossing in elongator mutants ( Figure 7A ) . The double mutant elpc-1; elpc-3 did not enhance the effect ( Figure 6A ) , suggesting that elpc-1 and elpc-3 act together in this function . We wanted to investigate whether this effect is transcriptional or posttranslational . The first evidence that this regulation is posttranslational was provided by the fact that GFP::MEC-12 cDNA driven by a heterologous promoter with a heterologous 3′ UTR did not change the relative levels of GFP::MEC-12 . This analysis was performed using the CAN-neuron specific promoter ceh-10 ( to rule out possible effects of differences in mec-12 expression in mechanosensory and non-mechanosensory neurons ) . As before , the signal from this GFP::MEC-12 minigene was increased in the elpc-1 and mig-2 ( gf ) ; elpc-1 background , but diminished in the mig-2 ( gf ) background ( Figure 7B ) . In a previous experiment , a point mutation that changed lysine 40 to glutamine ( Q ) in MEC-12/α-tubulin was still able to rescue a touch sensitivity phenotype , therefore MEC-12 ( K40Q ) was considered to be structurally functional [10] . Follow up experiments presented here were used to investigate whether lysine 40 has an impact on the regulation of GFP::MEC-12 levels . Given that the regulation was abolished in GFP::MEC-12 ( K40Q ) transgenic animals ( Figure 7C ) , provides strong evidence that the acetylation of lysine 40 is important for the posttranslational fine-tuning of MEC-12/α-tubulin levels presented here . These observations also explain the unexpectedly high levels of acetylation in L4 and young adult worms . This might be due to the elevated amount of MEC-12/α-tubulin levels leading to a misinterpretation of the Western data . In fact we observe an accumulation of MEC-12/α-tubulin in adults ( see Figure 8 ) . To decipher whether the increase of MEC-12/α-tubulin alters dynamics of MTs in vivo , the movement phenotype was scored in worms grown in the presence of sublethal concentrations of the MT depolymerizing drug nocodazole or the MT stabilizing drug taxol [46] . This assay was designed to uncover possible effects of the drug during development , e . g . commissure formation defects that persist from L1 to adulthood ( since the worms developed through all stages in the presence of the drug ) . In wt , movement was impaired by nocodazole in a linear dose dependent manner . elpc-1 and to a lesser extent elpc-3 mutants were both more resistant to nocodazole treatment with no additive effect seen in the double mutant elpc-1; elpc-3 ( Figure 7D ) . Compared to mig-2 ( gf ) , the suppression of mig-2 ( gf ) by single or double elongator mutants resulted in an increased resistance to nocodazole ( Figure 7D ) . To gain an insight into whether the observed nocodazole resistance in elongator mutants is neuron specific or due to a broader effect , elpc-1 was re-introduced under the control of a neuron specific promoter into elpc-1 mutant worms . The resistance to nocodazole was abolished and essentially wt , suggesting that this phenotype of elpc-1 is neuron specific ( Figure 7D ) . If stabilization of MTs is fundamental for the suppression of mig-2 ( gf ) by elongator mutants ( as suggested by the nocodazole experiment ) then MT stabilization by taxol should have a similar effect . In turn , elongator mutants may be expected to have an increased sensitivity towards taxol . Indeed , exposure to taxol suppressed the movement phenotype in mig-2 ( gf ) , an effect that was dependent on mec-12/α-tubulin . In contrast , elongator mutants by themselves displayed an increased sensitivity to high concentrations of taxol ( Figure 7E ) . In conclusion , these experiments show that elongator fine-tunes the amount of MEC-12/α-tubulin protein , which in turn may regulate the sensitivity of MTs to depolymerizing and stabilizating drugs revealing a change in their dynamic properties . Resistance to nocodazole in elpc-1 ( lf ) may arise due to MTs being stabilized by MT associated proteins . However this is unlikely , as the comparative expression analysis of wt and elpc-1 mutants by DNA microarray technology yielded no MT associated proteins ( data not shown ) . Alternatively , MTs number or length may be increased , a notion that was addressed by transmission electron microscopy ( TEM ) . An ideal target for investigation proved to be the ALM neuron because firstly , it has an unusually high number of special MTs ( Figure 8A ) and secondly , MEC-12/α-tubulin is highly expressed [10] , [47] ) . In randomly selected TEM sections , the overall number of MTs was indistinguishable in wt and mig-2 ( gf ) but significantly increased in elpc-1 ( lf ) ( Figure 8A and 8B ) . This was confirmed in serial sections ( in 3 . 2 µm increments ) of single wt and elpc-1 ( lf ) mutant worms ( Figure 8C ) . The identified variation in MT number along the axon was fed into a mathematical model ( see Experimental Procedures for details ) which suggested that , the difference identified in elpc-1 ( lf ) may be also the result of MTs being longer rather than solely based on an increase of MT number ( Figure S5B ) . This is , as said above , further supported by the fact that the DNA microarray data could not pinpoint MT nucleation factors ( e . g . γ-tubulin ) in elpc-1 ( lf ) that are instrumental for the increase in MT numbers . In summary , this provides at least circumstantial evidence that an increase in MTs number is unlikely ( Figure 8A–8C ) . If increasing the amount of MEC-12/α-tubulin causes longer MTs and subsequently suppresses mig-2 ( gf ) movement defects , any alternative means that increases MEC-12/α-tubulin levels should be able to replicate these observations . Treatment with lactacystin , a potent and specific inhibitor of the proteasome used in C . elegans [48] , imposed no movement effect on wt animals , but was able to partially suppress the movement phenotype of mig-2 ( gf ) at 15 µM ( Figure 8D ) . As mentioned before , the treatment was continuous during development of the worms , thereby ensuring that defects during early larval stages impose effects on locomotion at young adult stage . Moreover , the suppression was dependent on MEC-12/α-tubulin ( Figure 8D ) , thus ( re ) confirming the taxol data ( Figure 7E ) . Interestingly , compared to wt , mec-12/α-tubulin and elongator mutants were all sensitive to high concentrations of lactacystin ( Figure 8D ) . In both cases high levels of lactacystin are predicted to impair MT function , where accumulation of a mutated MEC-12/α-tubulin and “overstabilized” MTs in elongator mutants alter MT function; findings that are analogous to the taxol experiments ( Figure 7E ) . Ubiquitin mediated proteolysis is the key player in the control of cytoplasmic protein turnover . To investigate whether impaired proteasomal proteolysis of α-tubulin might be linked to the phenotypes observed in elongator mutants , immunoprecipitated α-tubulin from different genetic backgrounds was probed with an anti ubiquitin antibody ( Figure 8E ) . The ubiquitination level was higher in elpc-3 mutants compared to worms with an elpc-1 background . Interestingly , this observation parallels the strong effects on acetylation levels in elpc-3 mutants ( Figure 2B ) . These differences in post-translational modification are possibly due to the fact that ELPC-3 is the catalytic subunit of the elongator complex . This experiment demonstrated that α-tubulin is strongly modified by endogenous ubiquitin in elongator mutants . We conclude that elongator is a critical factor for α-tubulin turnover .
The results presented here demonstrate how a genetic animal model can be used to study acetylation of MTs . In detail , it was possible to corroborate that mig-2 regulates α-tubulin acetylation and ascertain that the elongator is a new regulator of α-tubulin acetylation . Recently , Creppe and colleagues showed that elongator is involved in this process in mouse cortical neurons [36] . Their data confirm the hypothesis , presented here , namely that elongator activity regulates α-tubulin function . In our work , regulation of acetylation by the elongator was shown to be strong in early stages of development and acetylation of α-tubulin lower , but nevertheless still present , in late stages of elongator mutants . The fact that elongator also regulates other processes and that some acetylation remains in the absence of elongator suggests the presence of some degree of redundancy . It should be noted that during embryonic development and early larval stages numerous neurons undergo dramatic morphological change and in consequence , a tight control of acetylation by elongator has a deep impact on neuronal structure during development . This notion is discussed below . Gain of function alleles of the Rac-like GTPase mig-2 display defects in neuronal shape due to the hyperacetylation of α-tubulin . The defects are partially suppressed ( although not fully ) when acetylation is reduced either by overexpression of HDAC-6 , the α-tubulin deacetylase , or by knocking out the elongator . This implies that mig-2 modulates other pathways as well . For example mig-2 is also involved in actin dynamics to regulate neuronal shape and axon pathfinding [49] . But if acetylation of MTs is regarded as a distinct fact , how can acetylation of MTs affect cell shape ? It has been shown in fibroblasts and neurons that Rac GTPases regulate MT dynamics by different mechanisms . One of those is the recruitment of +TIPs such as APC and EB1 to promote stabilization via Rac effector mDia [17] . The same +TIPs are proposed to act in the linkers between actin cortex and the protruding MTs in the growth cone of neurons [50] . Although acetylation of MTs seems to be secondary to stabilization [6] , it has not been excluded that RacGTPases may actively regulate acetylation . If α-tubulin is hyperacetylated , as in the Rac gain of function mutants , the cellular consequences are twofold: firstly , neurons fail to project their axons correctly and secondly neuronal migration is , at least in some cases , downregulated . The discovery that acetylation affects α-tubulin levels by regulating its turnover is important , since the dynamics of MT may also be altered in a system where α-tubulin is “hyperacetylated” . This may be the result of MTs being too short or prone to depolymerization . Whilst a regulation in number of MT cross-sections was not seen in the mig-2 ( gf ) mutant , it was possible to suppress its phenotypes by either raising the stability of the MTs with taxol or by mimicking an α-tubulin accumulation by blocking the proteasome . This supports the hypothesis that the regulation of α-tubulin levels affects the dynamics of MTs . In contrast , if acetylation is reduced , as seen in the elpc-1 mutants , then the length of MTs was probably increased . This evidence was supported by the observed resistance to the depolymerizing drug nocodazole and the increased sensitivity to taxol and proteasome inhibitor . In addition , a significant change in neuron shape was seen in elongator mutants . Altogether , the results advocate that the formation of stable MTs requires the fine-tuning of α-tubulin levels . Moreover , since Rac GTPases are upstream of these effects , it is possible that they contribute to the coordination of multiple events that lead to changes in cell shape , namely: a ) actin cytoskeleton dynamics , b ) stabilization of MT ends and c ) fine-tuning of MT dynamics via regulation of α-tubulin turnover . A further aspect is that axons of mig-2 ( gf ) mutants do not only elongate less , but also show misguidance phenotypes [21] . Since acetylation of α-tubulin affects vesicle loading and transport , it is conceivable that extracellular axon guidance cues might be deregulated by transport problems due to incorrect acetylation of MTs [6] . Recently it has been shown that mig-2 and vab-8 ( kinesin-like motor protein ) are required to regulate the correct subcellular localization of UNC-40 ( a homolog of the netrin receptor ) which specifies cell polarity of neurons [51] . In addition , the authors showed that UNC-40 was mislocalized in a mig-2 gain of function mutant . Of course MTs were hyperacetylated in that background , as shown by the results presented here . This leads to the assumption that UNC-40 was in fact mislocalized due to the improper loading and/or velocity of VAB-8 dependent vesicles . A more direct involvement of elongator complex in the regulation of intracellular trafficking that is critical for cell shape remodeling during migration and terminal branching of mouse cortical neurons ( corticogenesis ) has recently been shown [36] . In summary , this adds further evidence that acetylation may affect neuronal shapes by regulating a ) the dynamics of MTs and their propensity to be stabilized via the turnover of α-tubulin and b ) the loading and transport of vesicles ( and their cargo ) . The elongator mutants were shown to be resistant to the MT depolymerizing drug nocodazole . Based on this alone , one may conclude that MTs are more stable in elongator mutants , however further experiments showed that the resulting increase of α-tubulin levels also altered the dynamics of MTs . This can be confused with the notion that the “stability” of MTs requires an involvement of active stabilizing factors ( such as +TIPs ) which in the end change the half life of MTs [6] , [50] . Although not the subject of this study , the possible link between elongator and +TIPs is an interesting point that should be investigated in the future . Others have shown that changes in acetylation levels ( mostly achieved by blocking the α-tubulin deacetylase HDAC6 ) modify the stability of MTs [8] , [52] . Indeed , the experiments presented here confirm this point of view , and in addition allow a re-interpretation , namely that their findings are likely a result of the α-tubulin pool being regulated and not due to the direct regulation of MT stability . Deregulation of acetylation not only affects vesicle dynamics but also α-tubulin turnover ( summarized in model Figure 8F ) . Indeed , this is important for cellular function as revealed by the movement defect of the elongator; α-tubulin double mutant . It is noteworthy that the defects can only be seen when crossing structural mutants of α-tubulin and not regulative or null mutants ( Figure 3B–3D , Table S1 ) . These defects may thus be the consequence of the accumulation of damaged α-tubulin , which in turn alters MT and neuronal function . In contrast , precise regulation of the levels of functional α-tubulin is required in the suppression experiments , since all mutations in α-tubulin abolish suppression . This provides further evidence that acetylation is important for the regulation of α-tubulin levels . Besides this work , a significant and growing amount of studies identifies acetylation of various proteins as a key regulator for their turnover . A very interesting finding is that molecules that participate in protein deacetylation , for example HDAC6 , can directly interact with CDC48 , a factor required for quality control and polyubiquitination of substrates . This reveals a link between control of acetylation states , active deacetylation and degradation [53] , [54] . Recently it has been shown that elp3 was linked to neurodegenerative diseases and more specifically to Amyotrophic Lateral Sclerosis ( ALS ) in three families . ALS is the most common adult onset human motor neuron disease typically resulting in death from respiratory muscle weakness within three years [55] . Furthermore , mutations that cause tissue-specific exon skipping thereby truncating human Elp1 result in the autosomal recessive Familial Dysautonomia ( FD ) , one of the most frequent hereditary neuropathies [39] , [56] . Affected individuals are born with a reduced number of neurons within the autonomic and sensory nervous system . However , penetrance of the FD mutation is typically incomplete and a low level of full-length IKAP protein can prevail in brain tissues from FD patients [57] , [58] . DNA microarray analysis on human cells revealed that the depletion of Elongator ( via RNAi of ELP1 ) in fibroblasts modulates the expression of genes , several notably linked to motility and migration . This led to the conclusion that impaired cell movement within the nervous system might be involved in the neuropathology of FD patients [59] . Recently , a mouse knock out system showed that complete ablation of ELP1 leads to embryonic lethality due to defects in neurolation and blood vessel development [60] . As the worm knock-out of elpc-1 is viable , it is arguably at present , the most suitable system to study and model the ELP1 function in neurons . Indeed , it may at first seem surprising that the defects in the worm nervous system are more similar to the human disease , while the mouse model has such a severe phenotype . However , several differences within the systems may offer plausible explanations . In humans , ELP1 is only downregulated , rather than completely deleted as in the mouse ELP1 experiment . Moreover , whilst only one α-tubulin is acetylable in the nematode , namely MEC-12/α-tubulin ( expressed exclusively in neurons , which are not critical for survival as in mouse ) , all vertebrate α-tubulins possess the acetylatable lysine 40 . Therefore regulation of vertebrate turnover of α-tubulin may have a greater impact at the cellular and organismal level and contribute to the severe phenotype observed in the knock out mouse . The mouse phenotypes may all be a result of defective cytoskeletal dynamics , such as reduced polarization and/or vesicular transport due to the downregulation of α-tubulin acetylation . The results presented here offer exquisite support for this line of argumentation , but clearly need further investigations into the transcriptional function of the elongator in vertebrates . A further aspect is the degeneration of neurons present in FD patients as well as in ALS patients . In the worm model α-tubulin turnover is regulated through acetylation . Degenerative aspects may arise from deregulation of transcription as both ELP1 ( involved in FD ) and Elp3 ( involved in ALS ) are part of the same transcriptional elongator complex . Nevertheless , accumulation of “hypoacetylated” targets may be an additional explanation , since acetylation is important in the regulation of protein turnover . In addition , the reduced transport along MTs might lead to an accumulation of proteins destined for degradation . Indeed , the results show that transport is hampered in elongator mutants ( Figure 4 ) , that the elongator regulates α-tubulin acetylation and that downregulation of acetylation leads to accumulation of α-tubulin which in turn alters MT dynamics . Accumulation of α-tubulin , as well as the resulting changes in MT dynamics , may constitute an additional stress leading to degeneration and cell death . The HAT activity of the Elongator , which is directed toward histone H3 , has been proven in various systems to be crucial for transcription and affects genes important for cell movement in human fibroblasts [59] . Furthermore , it has been speculated that defects in this cellular function may underlie FD . In normal cells ( fibroblasts and neurons ) cell migration is always the concerted interaction between transcription and cytoskeletal dynamics . In this study α-tubulin was identified as an additional genetic target of the elongator . By acetylating histone H3 and regulating α-tubulin acetylation , it is conceivable that the elongator links and “synchronizes” the two functions to modulate the migration event . Rac GTPases may be involved in the response to outer signals and by doing so control the function of the elongator . To date , nothing is known about the regulation of elongator activation . This study highlights that acetylation in mig-2 ( gf ) mutants is increased . However whether this is due to the activation of pathways leading to acetylation regulated by the elongator , to the downregulation of HDAC-6 function or an indirect consequence of upregulation of stabilizing +TIPs still awaits to be unraveled . This study uncovers that the elongator is linked to the regulation of α-tubulin acetylation . In addition the results presented here indicate that elongator is not only important for vesicle transport but also for α-tubulin turnover . This in turn seems to affect MT dynamics . These observations are instrumental for the introduction of a novel point of view , ultimately explaining how neuronal function is perturbed in neuropathies of Amyotrophic Lateral Sclerosis and Familial Dysautonomia . These new insights offer intriguing cues that targeting microtubules may lead to changes in the design of future therapies .
Worms were handled according to standard procedures [61] . Strains: N2 ( wt ) , mig-2 ( mu28 ) X , mig-2 ( gm38 ) X , mig-2 ( gm103 ) X , rac-2 ( ok326 ) IV , ced-10 ( n1993 ) IV , mec-12 ( u76 ) III , mec-12 ( u63 ) III , mec-12 ( e1605 ) III , mec-12 ( e1607 ) III , unc-29 ( e1072 ) I , goa-1 ( n363 ) I , unc-64 ( e246 ) III , unc-104 ( e1265 ) II , mec-3 ( e1338 ) IV . elpc-1 ( ng10 ) I and elpc-3 ( ng15 ) V were generated by ourselves: elpc-1 ( ng10 ) I by TMP/UV mutagenesis and elpc-3 ( ng15 ) V by tilling . elpc-3 ( tm3120 ) V was kindly provided by S . Mitani . GABA-ergic motoneurons were visualized by crossing the strain juIs76[Punc-25::gfp] , analogously the CAN neuron was marked using lqIs4[Pceh-10::gfp] ( gifts from E . A . Lundquist ) . Dense core vesicles were visualized using the strain BL5752 bearing the double insertion: inIs181[ida-1::gfp] and inIs182[ida-1::gfp] . Transgenics made by ourselves: ( plasmid GU327 ) : ngEx2[elpc-1::gfp] ( 50ng co-injected with 75ng of ttx-3::gfp ) , ( plasmid GU329 ) : ngEx8[Ex[Punc-119::elpc-1::gfp] ( 50ng co-injected with 50ng of ttx-3::gfp ) , ( plasmid GU330 ) : ngEx43[Prab-3::elpc-1::gfp] ( 20ng co-injected with 80ng of ttx-3::rfp and 20ng of lin-15 ) , ( plasmid GU446 ) : ngEx94[Prab-3::elpc-1-FD::gfp] ( 20ng co-injected with 80ng of ttx-3::rfp and 20ng of lin-15 ) , ( plasmid GU320 ) : ngEx19[elpc-3::gfp] ( 20ng co-injected with 80ng of ttx-3::gfp ) and ( plasmid GU448 ) : ngEx33[Prab-3::elpc-3] ( 20ng co-injected with 80ng of ttx-3::rfp ) ( see Figure S3 ) ; ( construct GU326 ) : ngIs14[hdac-6::gfp]I ( 20ng co-injected with 80ng of ttx-3::gfp ) ( wormbase . org: “HDAC-6” = F41H10 . 6c ) , ( construct GU312 ) : ngIs9[Pmec-12::gfp::mec-12]III ( 5ng co-injected with 80ng of ttx-3::rfp and 20ng of lin-15 ) , ( construct GU331 ) : ngIs10[Pceh-10::gfp::mec-12]X ( 20ng co-injected with 50ng of ttx-3::rpf and 50ng of lin-15 ) and ( construct GU314 ) : ngIs11[Pmec-12::gfp::mec-12 ( K40Q ) ]IV ( 5ng co-injected with 80ng of ttx-3::rfp and 20ng lin-15 ) ( see Figure 6A–6C ) . The construct GU314 was also used for the differential expression of mec-12 in Figure 4A: ngEx98 showed no detectable GFP signal ( control strain for no expression of mec-12::gfp ) , ngEx99 and ngEx100 showed weak , ngEx101 medium and ngEx102 strong expression of mec-12::gfp ( 3ng co-injected with 80ng of ttx-3::rfp and 20ng lin-15 ) . Constructs for nuclear or cytoplasmic expression of elpc-3: ( construct GU449 ) : ngEx103 [Prab-3::NLS::elpc-3::gfp] ( containing the NLS sequence from the Fire Lab vector kit ) ( 20 ng co-injected with 80 ng of ttx-3::rfp and 20 ng of lin-15 ) , ( construct GU450 ) : ngEx104 [Prab-3::NES::elpc-3::gfp] ( 20 ng co-injected with 80 ng of ttx-3::rfp and 20 ng of lin-15 ) . Constructs for ectopic expression of mec-12 , elpc-1 and elpc-3 were designed using the promoter regions of ceh-10 for CAN neuron expression [21] , unc-119 and rab-3 for panneuronal expression ( Figure S4 ) . Plasmids GU327 , GU329 and GU330 contain PCR amplified elpc-1 cDNA cloned into pPD95 . 75 using the PstI and XmaI sites . Prab-3::elpc-1-FD::gfp ( GU446 ) was obtained deleting the 3′ end of Prab-3::elpc-1::gfp ( GU330 ) using the convenient HpaI site ( in the cDNA ) and the MscI site ( in the vector ) and blunt ligated , in frame , with gfp ( Figure S4 ) . GU320 contains the whole elpc-3 operon amplified by PCR and cloned into the XmaI site of pPD95 . 75 . GU448 contains the elpc-3 cDNA under the control of the rab-3 promoter cloned into the NotI XmaI site of pPD95 . 75 vector . The NLS and NES versions of elpc-3 were obtained by PCR amplification of the respective sequences and cloned into vector GU448 . The coding sequence for the NES was: CTT CCA CCA CTC GAG AGG CTT ACG CTT . GU326 was obtained by PCR amplification of F41H10 . 6c ( long isoform of HDAC-6 ) and cloning into pPD95 . 75 using the SphI and XbaI sites . GU312 contains mec-12 genomic locus ( including the putative promoter ) cloned into PstI and KpnI sites of pPD117 . 01 , followed by insertion of gfp into BglII site in-frame with mec-12 . Pmec-12::gfp::mec-12 ( K40Q ) ( GU314 ) was obtained by standard point mutation method using Pmec-12::gfp::mec-12 as template and subsequently sequenced . GU331 contains ceh-10 promoter inserted via SalI and BamHI sites into pPD117 . 01 and subsequently mec-12 cDNA was added with XbaI site into the compatible NheI site . Primer sequences used for PCR amplifications are available upon request . All expression constructs were cloned into the vector pPD95 . 75 except for mec-12 and mec-12 ( K40Q ) cloned in pPD117 . 01 ( kind gift from A . Fire ) . The elpc-1 ( ng10 ) deletion allele was performed as described [62] . elpc-3 ( ng15 ) was isolated by tilling as described [63] . Integration of constructs GU326 , GU312 , GU331 and GU314 was obtained by UV-irradiation ( 30 , 000J ) of L4 stage worms . Fertile young adults were singled the next day and subsequently starved for 2 weeks to lose the non-integrated extrachromosomal array . Starved worms were transferred onto fresh growth plates and scored for stable inheritance of the respective ttx-3::rfp marker . Integrants were outcrossed twice then mapped by crossing with strain DA438 containing markers for each chromosome ( kindly provided by CGC and constructed by L . Avery ) . Worms grown in liquid at 20°C to the appropriate stage were collected in buffer A ( 20 mM Tris pH 7 . 4 , 200 mM NaCl , 10% ( v/v ) glycerol , 1% Triton ) and stored at −80°C . Protein extracts were prepared adding 1 mM PMSF , 1× complete protease inhibitor ( Roche ) and 300 nM TSA ( Sigma T8552 ) . Acid-washed glass beads ( Sigma G8772 ) and FastPrep-24 sample preparation system ( MP Biomedicals ) were used to homogenize worms ( 4 . 0 m/s for 45 seconds at RT ) . After centrifugation at 4°C , the protein concentration of the supernatants was determined by Bradford assay ( Biorad ) and 15 µg total protein were resuspended in Laemmli buffer . Proteins separated on 10% SDS-PAGE gels were detected by immunoblotting using ECL . Worm extracts for immunoprecipitation were prepared in HNNG buffer ( 15 mM Hepes pH 7 . 5 , 250 mM NaCl , 1% NP- 40 , 5% glycerol , 1 mM PMSF , 10 mM sodium butyrate , -Sigma B5887- , 300 mM TSA and 1× protease inhibitors cocktail ) . Lysates ( 1–2 mg/ml ) were incubated overnight with 1 µg of antibody at 4°C . Immunocomplexes were collected with protein A plus sepharose beads ( Amersham Biosc ) , protein G plus sepharose beads ( Zymed ) , anti-α-tubulin- or anti-ELPC-1-conjugated agarose beads , sequentially washed at 4°C with HNNG buffer and finally resuspended in Laemmli buffer . anti-α-tubulin and anti-ELPC-1 crosslinked resins have been produced by Cogentech ( Consortium for Genomic Technologies , Milan , Italy ) using standard procedures . Antibodies used for western blots and/or immunoprecipitations: anti-α tubulin ( Sigma T5168 ) , anti-acetylated-α-tubulin ( Sigma , T6793 ) , anti-actin ( MP biomedicals , #69100 ) , anti-GFP ( Torrey Pines Biolabs Inc , TP401 ) and anti-Ubi ( Upstate , # 05-944 ) . Rabbit polyclonal anti-ELPC-1 antibody was produced by Cogentech . anti-rabbit or anti-mouse IgG HRP-conjugated antibodies were both from Cell Signaling Tech ( # 7074 and # 7076 , respectively ) . Western blots were performed at least twice to confirm the results . TEM was performed using standard procedures . Briefly , worms were washed in M9 and anesthetized in 8% ethanol in M9 for 5 min . They were placed in a fixative ( 2 . 5% glutaraldehyde , 1% paraformaldehyde in 0 . 1M sucrose , and 10 mM PBS , pH 7 . 4 ) , cut open with a needle at the anterior and posterior ends , and fixed for 2 h . Worms were embedded in 2% agarose , cut into small blocks , and washed three times in PBS . Subsequently , pieces were fixed with a second solution ( 1% osmium tetroxide , 1 . 5% potassium ferrocyanide in PBS ) for 2 h and washed three times in water . Worms were stained with 1% uranyl acetate for 1 h . Samples were dehydrated in ethanol ( 10 min in 50% ethanol , 10 min in 70% ethanol , 10 min in 90% ethanol , and 10 min in 100% ethanol ) and acetone ( 10 min ) . Blocks with worms were embedded in Epon resin ( Fluka , Buchs , Switzerland ) : first in Epon-acetone ( 1∶1 ) for 1–2 h and then in pure resin for 2–4 h . Samples polymerized for 24–48 h at 60°C and in 60-nm sections were prepared with Ultracut E . Sections were stained in uranyl acetate for 60 min and then 2 min in Millonig's lead acetate stain . Pictures were taken on Philips Morgagni 80 KV microscope ( Eindhoven , The Netherlands ) [64] . GFP signal was quantified by taking pictures of ALM or CAN neuronal body of anesthetized L1 at 63× magnification at fixed exposure time ( 2 sec . ) The total signals were measured using ImageJ software . Whole mount staining of acetylated MEC-12/α-tubulin using the specific monoclonal antibody ( see below ) was performed according to previous protocols [10] . Cy3 conjugated secondary antibodies from Jackson Lab . were used . Primary antibody dilution: 1/500 . Secondary antibody dilution: 1/200 . Analysis of dense core vesicles was performed as described [43] . Immunolocalizations were repeated 3× per genotype . In Figure 7C we show the two sets of measurements for the number of intersections of microtubules on eight independent sections of ALM in a wild type and an elpc-1 ( lf ) mutant . In wt the average number of intersections and its standard deviation are μ1 = 67 and σ1 = 9; for elpc-1 ( lf ) μ2 = 85 and σ2 = 5 . The means are statistically different: μ1 differs from μ2 ( one-sided t test , P = 2·10−4; 11–Inf is 95% confidence interval for μ2−μ1 ) . These measurements confirm that the two individuals of different genotype show similar MT intersection levels than the respective ones in Figure 7B . Based on these data , homogeneity of variances cannot be discarded ( one-sided F test , P = 0 . 1 ) . Interpretation model: Assuming that microtubules have a fixed length h , parallel to the axis of the axon and distributed uniformly along an axon of a given length L leads to following calculations: the probability for a microtubule to cross any section is p = h/L and follows the Poisson distribution ( h/L equals the ways a “stick” of length h can be placed in a longer “tube” of length L ) . Performing N trials of a Poisson process with probability p , the expected value and the variance of the number of successes for a random variable X are E ( X ) = Np and VAR ( X ) = Np ( 1−p ) , respectively . Thus , if we have m microtubules , the expected value for the number of intersections is E ( X ) = mh/L , while the variance is VAR ( X ) = mh/L ( 1−h/L ) . Since the elpc-1 ( lf ) mutant has , on average , more microtubules intersections than wild type , using E ( X ) = mh/L we obtain the condition m2h2>m1h1 ( 1 = wt; 2 = elpc-1 ( lf ) ) . In other words , elpc-1 ( lf ) have more microtubules , or the microtubules are longer , or both . It seems unlikely that the main change is the number of microtubules , because the assumption that microtubule length is unchanged , i . e . h2 = h1 and m2>m1 , requires VAR2 ( X ) >VAR1 ( X ) , i . e . σ22>σ21 . The later is unsupported by the measurements . Body bends per minute: Worms grown at 20°C to L4/young adult stage were placed into water . Body bends were counted for 1 minute . Blind scoring was not applied to these tests , because they were highly reproducible and not subjective . The number of body bends per minute for wt worms in our tests reproduced published levels [65] . Drug tests: Standard NGM plates were supplemented with 0 . 1–0 . 5 µM nocodazole , 1–5 µM taxol [46] , 0–75 µM leptomycin B [42] or 0–40 µM lactacystin [48] . DMSO [for nocodazole and taxol] or water [for leptomycin and lactacystin] alone were used as a control . Worms were grown in the presence of the respective drug and body bend assays were carried out on the F1 . The Aldicarb tests were performed as described [44] . Concentration used: 0–40 µM . Modifications to the original protocol: worms were exposed only 1h ( instead of 4h ) and the test was performed in M9 buffer ( this modified procedure increased the sensitivity ) . Levamisole test was performed as described [44] . Aldicarb and levamisole tests were repeated 5 times using each time 12 animals calculating the average and error . The sensitivity of wt worms when exposed 4h to aldicarb and in the levamisole test reproduced the published data ( not shown , [44] ) . Touch sensitivity: young adult worms were touched 10 times with an eyelash , alternating between the anterior and posterior part of the body . Between touches worms were given 1 min to recover . Positive response by movement away form the stimulation was scored for 10 independent worms per strain . Performed as described [21] , [40] . Young adult worms were flash frozen in water ( 1000 worms per pellet ) . Extraction was performed in 0 . 2 M perchloric acid ( PCA ) solution with 100 mM EDTA . Ethylhomocholine ( 0 . 1 µM ) served as internal standard . To homogenize the worms the FastPrep®-24 sample preparation system with beads was used as described above . After centrifugation , the pH of the supernatant was neutralized using 1M KHCO3 and purified using 0 . 45 mm pore size filters . Final extracts were stored at −80°C and analyzed on an electrochemical detector HTEC-500 ( Eicom Co . Kyoto , Japan ) as described [66] . If not otherwise stated , statistical significance was performed by two-tailed unpaired Student's t-test . P value>0 , 05 was scored as “ns” ( not significant ) . | We were able to demonstrate how a screen , that utilized the nematode model organism Caenorhabditis elegans , yielded the novel discovery that the Elongator protein complex is critical for neuronal development and microtubule acetylation . Other scientists have previously shown that the Elongator is important for transcription , but also hypothesized that a hitherto unknown function within the cytoplasm prevails . Regulation of microtubule acetylation is indeed important for cellular function . Our results provide the first tantalizing insights of how worms display neuronal phenotypes that may be linked to changes in degrees of acetylation of microtubules . We also point out that the Elongator itself has a defined acetyltransferase function that could well be directly responsible for the acetylation of α-tubulin . In conclusion , we discuss how the regulation of microtubule acetylation might impact on the understanding of human neurodegenerative diseases , namely Familial Dysautonomia and Amyotrophic Lateral Sclerosis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
]
| [
"genetics",
"and",
"genomics/animal",
"genetics"
]
| 2010 | The Caenorhabditis elegans Elongator Complex Regulates Neuronal α-tubulin Acetylation |
Mutations in sorting nexin 10 ( Snx10 ) have recently been found to account for roughly 4% of all human malignant osteopetrosis , some of them fatal . To study the disease pathogenesis , we investigated the expression of Snx10 and created mouse models in which Snx10 was knocked down globally or knocked out in osteoclasts . Endocytosis is severely defective in Snx10-deficent osteoclasts , as is extracellular acidification , ruffled border formation , and bone resorption . We also discovered that Snx10 is highly expressed in stomach epithelium , with mutations leading to high stomach pH and low calcium solubilization . Global Snx10-deficiency in mice results in a combined phenotype: osteopetrosis ( due to osteoclast defect ) and rickets ( due to high stomach pH and low calcium availability , resulting in impaired bone mineralization ) . Osteopetrorickets , the paradoxical association of insufficient mineralization in the context of a positive total body calcium balance , is thought to occur due to the inability of the osteoclasts to maintain normal calcium–phosphorus homeostasis . However , osteoclast-specific Snx10 knockout had no effect on calcium balance , and therefore led to severe osteopetrosis without rickets . Moreover , supplementation with calcium gluconate rescued mice from the rachitic phenotype and dramatically extended life span in global Snx10-deficient mice , suggesting that this may be a life-saving component of the clinical approach to Snx10-dependent human osteopetrosis that has previously gone unrecognized . We conclude that tissue-specific effects of Snx10 mutation need to be considered in clinical approaches to this disease entity . Reliance solely on hematopoietic stem cell transplantation can leave hypocalcemia uncorrected with sometimes fatal consequences . These studies established an essential role for Snx10 in bone homeostasis and underscore the importance of gastric acidification in calcium uptake .
The function of the bone-resorbing osteoclast is highly dependent on vesicular trafficking pathways [1] . Endocytosis and intracellular trafficking of the endocytosed material are required for many osteoclast functions , including creation of the ruffled border; secretion of ions and proteases to digest bone; to engulf the digested material; to move it across the cell by transcytosis; and to secrete the products of digestion [2 , 3] . Disruption ( genetic or pharmacological ) of osteoclastic vesicle transport abolishes resorptive activity [1] . For example , osteoclasts from human patients and from rats deficient in Plekhm1 , a protein with a critical function in vesicular transport in osteoclasts , develop osteopetrosis and have osteoclasts with secretory defects and which lack ruffled borders [4] . Members of the sorting nexin ( Snx ) family of proteins are known to mediate endosomal sorting , endocytosis , recycling of membrane proteins , and trafficking between various endosomes and Golgi apparatus [5] . The Snx family consists of a diverse group of cytoplasmic and membrane-associated proteins that are unified by a common phospholipid-binding motif , the PX domain . They participate in protein sorting and membrane trafficking by means of protein-protein complexes and protein-lipid interactions . [5] . Overexpression of one Snx family member , Snx10 , induced giant vacuoles in mammalian cells [6] . During investigations of genes differentially expressed during RANKL-induced osteoclast differentiation , we found Snx10 to be very strongly upregulated both in vitro and in vivo [7] . Immunohistochemical analysis of mouse embryo sections showed expression in long bone , calvariae , and developing teeth . Snx10 was expressed in cells that also express tartrate-resistant acid phosphatase ( TRAP ) , demonstrating osteoclastic localization [7] . Snx10 silencing inhibited formation of resorption pits on hydroxyapatite and also TRAP secretion [7] . Taken together , these results indicate that Snx10 is expressed in osteoclasts and is required for osteoclast activity in vitro . In 2012 , SNX10 mutations were discovered in patients with infantile autosomal recessive osteopetrosis . One was a point mutation that caused a single amino acid change in a highly conserved residue , R51Q , [8] and one introduced a premature stop codon [9] . Osteoclasts from these patients showed reduced resorptive capacity and altered endosomal pathways [8] . In 2013 , nine novel mutations in SNX10 were then described in 14 autosomal recessive osteopetrosis ( ARO ) patients , and together , SNX10 mutations are now known to accounting for about 4% of known ARO cases , roughly the same proportion as mutations in the RANK-RANKL pathway or in OSTM1 [10] . Most patients with SNX10 mutations benefited from hematopoietic stem cell transplants ( HSCT ) . However , some patients exhibited symptoms consistent with osteopetrorickets and did not experience improvement after HSCT , which suggests that bone may not be the only site of expression [10] . Several molecules in the acid producing system are expressed in both bone and stomach cell types , including Atp6i , Clc-7 and Ae2 . ATP6i [11] and Ae2 [12–14] loss-of-function mutants develop osteopetrorickets due to simultaneous acidification defects leading to osteoclast dysfunction and impaired calcium absorption . Osteopetrorickets has been described in both humans and in mice and it is often fatal in infants . It manifests as the seemingly paradoxical combination of dense , sclerotic bones , but with defective mineralization of hypertrophic cartilage and of osteoid , often in the presence of elevated parathyroid hormone , alkaline phosphatase , and decreased 1 , 25-dihydroxyvitamin D levels [15 , 16] . While rickets was once considered a rare complication of osteopetrosis , it has been suggested that osteopetrorickets is more common than previously thought [17–19] . Given its critical role in osteoclast function [7] , it seemed likely that Snx10 may also regulate acid production in the stomach . To study the role of Snx10 in bone homeostasis in vivo we generated strains of mice that carry either a global knockdown of Snx10 or an osteoclast-specific knockout . We report here that Snx10 is highly expressed , not only in osteoclasts , but also in gastric epithelium , leading to osteoclastic and gastric acidification defects , similar to the Tcirg1-deficient oc/oc mice [11] . We also show that osteoclast-specific knockout of Snx10 led to severe osteopetrosis , but without rickets , and that dietary calcium supplementation rescued mice with a global knockdown of Snx10 from the rachitic phenotype and prevented juvenile lethality , indicating that this simple remedy should always be considered clinically .
In order to study the role of Snx10 in osteoclast function and bone homeostasis in vivo we generated mice from KOMP gene-trapped ES cells ( PG00216_Z_2_C06 , S1A Fig . ) . This targeting construct places a selection cassette and several recombination sites within intron 3 , flanks exons 4 and 5 with loxP sites , and disrupts transcription . PCR genotyping ( S1B Fig . ) and mRNA levels ( S1C Fig . ) validated correct insertion and revealed an 86% reduction in Snx10 levels in bones from the resulting , homozygous mice compared to wildtype littermates ( relative expression = 0 . 14 vs . 1 . 03 , n = 4 per group , p<0 . 05 , S1C Fig . ) . We have designated these mice Snx10Neo-f/Neo-f ( see EXPERIMENTAL PROCEDURES ) and we hereafter refer to them as “Snx10 KD” for Snx10 knockdown to reflect the global insufficiency of Snx10 expression . Snx10 KD mice die between 3 and 4 weeks post-partum . By 14 days of age Snx10 KD mice exhibited severe growth retardation compared to WT or heterozygous controls ( Fig . 1A ) , with failed tooth eruption ( compare Fig . 1B and 1C ) . The overall skeletal development was impaired , with higher radio-density in the 3-week-old Snx10 KD mice ( Fig . 1E , 1G and 1I ) compared with the wild-type mice ( Fig . 1D , 1F and 1H ) . Further skeletal examination of 3 week-old mice by micro-CT ( Fig . 2 ) confirmed that the long bones of Snx10 KD mice were filled with unresorbed trabecular bone and lacked marrow spaces , consistent with a severely osteopetrotic phenotype . Micro-CT analysis of the femur , skull , and mandible performed in Snx10 KD mice ( n = 4 ) and WT littermates ( n = 3 ) confirmed the osteopetrotic phenotype . In fact , Snx10 KD mice had significantly higher BV/TV ( 0 . 31 +/− 0 . 052 vs 0 . 070 +/− 0 . 004 , P = 0 . 0006 ) , higher trabecular number ( 15 . 17 +/− 2 . 92 1/mm vs 2 . 29 +/− 0 . 004 1/mm , P = 0 . 0007 ) and significantly reduced trabecular spacing ( 0 . 061 +/− 0 . 009 mm vs 0 . 391 +/− 0 . 042mm , P = 1 . 97535E-05 ) than WT . Trabecular thickness values , on the other hand , were not different ( 0 . 032 +/− 0 . 002 mm for the Snx10 KD , 0 . 031 +/− 0 . 004 mm for the WT , p = 0 . 793 ) . The abundance of trabecular bone suggests that the observed phenotype is mainly the result of an osteoclast defect . In addition , all examined Snx10 KD bones exhibited a thinned or absent cortex , producing a dramatic “moth-eaten” appearance ( Fig . 2A , 2B , and 2C ) . Finally , analysis both by radiograph and by micro-CT of transverse sections of long bones ( femur , tibia , humerus ) revealed an inner ring of cortex-like ( denser ) bone within the trabecular bone ( shown in the femur , Fig . 2E and 2G ) . This is reminiscent of the classic “bone-in-bone” or "bone within bone" appearance , a typical radiologic finding in osteopetrosis , usually reserved for the vertebral column [15 , 17 , 20–22] . Radiographs also revealed metaphyseal fraying and cupping ( Fig . 1I ) , indicative of rickets superimposed on osteopetrosis , or “osteopetrorickets . ” To confirm this , we compared Snx10 KD mice with Tcirg1 ( Atp6i ) KO mice [23] , which display both hypocalcemia and osteopetrorickets [11] . Radiograph and micro-CT analysis ( S2 Fig . ) demonstrated that that the Tcirg1 KO mice were strikingly similar to the Snx10 KD mice by radiograph and micro-CT , with lack of cortical bone , metaphyseal cupping and fraying , and non-mineralized condyles and articulations , all characteristic of rickets . Moreover , biochemical investigations showed hypocalcaemia and high PTH together with low levels of serum 25-hydroxy vitamin D ( see below ) . Based on our own findings and reports on mice and humans with TCIRG1 mutations , we conclude that the Snx10 KD mice have a phenotype of osteopetrosis with super-imposed rickets: osteopetrorickets [17 , 24] . Analysis of bone mineral density by DXA demonstrated that BMD is reduced ∼14% in Snx10 KD mice compared to WT mice ( 0 . 0337 +/− 0 . 002 g/cm2 vs . 0 . 0393 +/− 0 . 003 g/cm2 , respectively . n = 6 per group , P = 0 . 02 ) . Furthermore , BMC is reduced 43% ( 0 . 114 +/− 0 . 043 g vs . 0 . 199 +/− 0 . 034 g , Snx10 KD and WT respectively , P = 0 . 006 ) . The greater reduction in BMC is likely a reflection of both reduced BMD and the small size of the Snx10 KD mice These observations were further confirmed by histological analysis of femora from 3-week-old mice . Low magnification images ( Fig . 3A and 3B ) , demonstrated: 1 ) a marrow cavity almost completely filled with unresorbed , poorly mineralized cartilage , and 2 ) thinning of cortical bone ( black arrowheads ) in the Snx10 KD mice . High magnification images of von Kossa/van Gieson stained undecalcified longitudinal sections of femora from WT and Snx10 KD mice confirmed the presence of numerous trabeculae within the bone marrow space of Snx10 KD bones ( Fig . 3D ) . Interestingly , the van Gieson counterstain also revealed the presence of non-mineralized osteoid on the surface of the trabeculae ( Fig . 3D ) . Histomorphometry performed on longitudinal femoral sections confirmed a significant increase in growth plate thickness ( GpTh , ) in Snx10 KD mice compared to WT mice ( 0 . 286 +/− 0 . 062 mm and 0 . 175 +/− 0 . 010 mm , respectively , P = 0 . 04 , n = 3 per group , S2 Table ) . We also detected a significant increase in Osteoid volume per Bone volume ( OV/BV , % ) in Snx10 KD mice ( 26 . 97 +/− 8 . 19% and 2 . 88 +/− 1 . 63% , respectively , P = 0 . 01 ) and in Bone volume / Tissue volume ( BV/TV , % ) ( 20 . 69 +/− 2 . 15% for the WT and 28 . 11 +/− 0 . 98% for the Snx10 KD , P = 0 . 02 , S2 Table ) . Histomorphometry performed on sections from lumbar vertebral bodies ( S4 Table ) and from skull base/floor bones ( S5 Table ) show similar results . Put together , these results confirm a combined phenotype in Snx10 KD mice characterized by both a bone mineralization defect ( rickets ) and osteopetrosis due to defective osteoclast resorption . We performed mechanical testing of tibial diaphyses ( three-point-bending ) to obtain the following parameters: maximal load , stiffness , energy to failure , and maximal displacement , which together provide an accurate picture of the biomechanical status of bone [25] . We found that the tibiae of Snx10 KD mice have significantly lower maximal load ( P = 0 . 0025 ) , stiffness ( P = 0 . 001 ) , energy to failure ( P = 0 . 027 ) , and greater maximal displacement ( P = 0 . 05 ) compared to WT ( n = 6 , S1 Table ) . Therefore , compared to WT , the long bones of Snx10 KD mice are weaker , less mineralized , and easier to deform in response to mechanical stress . It should be noted that all the mentioned values ( maximal load , stiffness , energy to failure and displacement ) are dependent on bone size . Therefore the values obtained for bones from Snx10 global deficient mice ( Snx10 KD ) reflect the fact that they have different material properties and that they are smaller . These results are consistent with rickets , which leads to softening and weakening of the bones ( osteomalacia ) [24] , whereas osteopetrosis alone results in hard , brittle bones due to the persistence of woven bone [22 , 26] . Taken together , these results show that Snx10 is essential for bone homeostasis and to maintain normal biomechanical properties of long bones . An osteopetrotic phenotype can be the result of absent osteoclasts ( e . g . , RANKL knockout ) , which indicates a defect in osteoclast differentiation . Osteopetrosis can also result from impaired osteoclast function ( e . g . , c-Src knockout [27] ) . Histological analysis showed that bones from Snx10 KD mice have TRAP-positive osteoclasts ( S2A Fig . ) , which suggests that Snx10 deficiency affects osteoclast function . Due to the lack of marrow in Snx10 KD mice , we used splenocytes as a source of osteoclast precursors from WT and Snx10 KD mice for ex vivo differentiation experiments . Snx10 KD splenocytes gave rise to multinucleated osteoclasts , confirming that Snx10 deficiency does not inhibit osteoclast formation ( S2C Fig . ) . In pit formation assays we found that , although the Snx10-deficient cells gave rise to TRAP-positive multinucleated osteoclasts , the total area resorbed by Snx10 KD osteoclasts was reduced by 94% ( S2D Fig . , top panel ) ( 3 . 99 +/− 1 . 41 mm2 for the WT , 0 . 26 +/−0 . 15 mm2 for the Snx10 KD , n = 6 , P = 0 . 04 ) , in agreement with our knockdown results [7] . Finally , infection of Snx10 KD splenocytes with a retrovirus expressing Snx10 reintroduced Snx10 expression ( S2D Fig . , bottom panel ) and corrected the pit formation defect ( 2 . 25 +/− 0 . 62 mm2 n = 6 , p = 0 . 008 ) , confirming that Snx10 deficiency is responsible for the resorption defect ( S2D Fig . , top and middle panels ) . To determine the effect of Snx10 deficiency on bone turnover , we assessed levels of serum markers for bone formation ( osteocalcin ) and resorption ( collagen I C-telopeptide; CTX ) . We found no significant differences in osteocalcin ( 18 . 35 +/− 1 . 81 ng/ml in the WT and 15 . 09 +/− 0 . 18 ng/ml in the Snx10 KD , n = 6 , P = 0 . 127 ) , suggesting that the bone phenotype observed in Snx10 KD is not primarily a bone formation defect . In contrast , serum CTX levels were significantly elevated in Snx10 KD mice compared to WT mice ( 88 . 56 +/− 5 . 67 ng/ml in the WT and 220 . 09 +/− 16 . 77 ng/ml in Snx10 KD , n = 6 , P = 0 . 0002 ) . This seemingly contradictory finding is consistent with reports in other osteopetrosis models [28–30] , and may be due to the very large bone area resulting in a much larger total number of osteoclasts in the Snx10 KD mice ( see Discussion ) . Morphological examination of stomachs ( n = 6 ) showed that Snx10 KD mice have an abnormal digestive tract , with stomachs prone to hemorrhagic necrosis ( Fig . 4A ) , suggesting that Snx10 deficiency may result in a functionally impaired digestive system . Since Snx10 is required for extracellular acidification in osteoclasts , we hypothesized that it might also be required for gastric acid production and subsequent normal intestinal calcium absorption . To test this , we first examined Snx10 expression by qPCR analysis and found it to be expressed in the stomach and bone , but not in heart , muscle or intestine ( Fig . 4B , left panel ) . In the stomach , expression was restricted to the oxyntic mucosa of the body/corpus ( Fig . 4B , right panel ) , the region producing gastric acid . Expression of Snx10 mRNA in the stomach of Snx10 KD mice is reduced by 78% compared to WT . This is a similar reduction to that in bone ( S1C Fig . ) . Immunofluorescence of stomach sections confirmed that Snx10 is normally present predominately in the intracellular secretory granules of zymogenic chief ( ZC ) cells ( Fig . 4C , WT ) . Immunofluorescence shows that Snx10 antibody labels ZC cells at the base of the gastric units only in WT mice ( S4A Fig . , bottom ) . ZC in WT are large cells containing apical cytoplasm filled with secretory granules which show intense labeling for the marker , GIS , whereas in Snx10 KD mice , ZCs are smaller and have sparser , more punctate granules with weak GIS staining ( S4A Fig . ) . This resembles other mouse models with defects in secretory granule maintenance/formation , such as the Mist1-/- mouse [31 , 32] . GSII labels mucous neck cells , which are interspersed between parietal cells and define the mid/neck region of each gastric unit . Their abundance and location in Snx10 KD mice are not distinguishable from WT . H&E staining ( S4B Fig . ) suggests a small size and high nuclear:cytoplasmic ratio of parietal cells in Snx10 KD mice . The ZCs , lacking their abundant apical granules , are also smaller , and there are increased parietal cells in the base ( ZC ) zone ( S4B Fig . , bottom , black arrowheads ) . Stomach sections were also stained for the parietal cell marker VEGF-B [33] . We confirmed that , compared to their WT counterparts , Snx10-deficient parietal cells are smaller and have an increased nuclear to cytoplasmic ratio ( Fig . 4D ) . Also , VEGF-B is focused in a ring-like cytoplasmic pattern in Snx10-deficient parietal cells , whereas it is broadly distributed baso-laterally in WT cells , suggesting a defect in cellular protein trafficking . To study possible functional consequences of Snx10 deficiency in the stomach , we measured gastric pH in WT and Snx10 KD stomachs . We detected a significant increase in stomach pH of Snx10 KD mice compared to WT mice ( Fig . 4E , 4 . 62 +/− 0 . 43 vs . 2 . 50 +/−0 . 32 , respectively , n = 5 per group , P<0 . 001 ) . Taken together , these findings demonstrate that Snx10 is expressed in the stomach and its expression is required for stomach acidification . Because impaired stomach acidification reduces calcium absorption by the intestine , we investigated whether Snx10 was required for maintaining calcium homeostasis , which could explain the rachitic aspects of the phenotype . Assessment of serum calcium confirmed that Snx10 KD mice are severely hypocalcemic compared to their WT littermates ( KO 6 . 40 +/− 2 . 46 mg/dl vs WT 10 . 21 +/− 1 . 81 mg/dl , n = 6 per group , P = 0 . 01 , Fig . 4F ) . They also had higher serum parathyroid hormone PTH ( 0 . 748 +/− 0 . 127 ng/ml , Fig . 4G ) compared to the WT ( 0 . 142 +/− 0 . 023 ng/m , n = 3 per group , P = 0 . 001 ) and lower 1 , 25-dihydroxyvitamin D levels ( 0 . 071 +/− 0 . 014 nmol , Fig . 4H ) , compared to WT ( 0 . 129 +/− 0 . 017 nmol , n = 3 per group , P = 0 . 01 ) ; both typical manifestations of osteopetrorickets [17–19] . Taken together , these results suggest that the rachitic features of Snx10 KD are due to hypocalcemia caused by the acidification defect in the stomach . To confirm this , we performed calcium supplementation to rescue mice from the Snx10 KD rachitic phenotype . The diet included 2% calcium gluconate and 2 , 000 IU kg-1 vitamin D as described [11] . Due to the tissue damage seen in the stomachs of Snx10 KD mice we sought to improve calcium uptake by also giving sub-cutaneous injections of 2% calcium gluconate diluted in saline ( 50μl per day ) . Calcium supplementation was initiated at 14 days after birth and continued for 10 days . In a parallel survival experiment , mice ( n = 5 ) were treated at 14 days after birth for 16 weeks . Calcium supplementation of Snx10 KD mice prevented the premature death seen at 3–4 weeks in the untreated mice . All mice receiving supplementation survived for the duration of the survival experiment . Treatment restored normocalcemia ( 10 . 52 +/− 0 . 43 mg/dl for treatment group vs . 10 . 70 +/− 0 . 46 mg/dl for the WT group , n = 6 , P = 0 . 661 ) ( Fig . 5A , left ) , bone mineral density ( n = 6 , P = 0 . 006 ) and bone mineral content ( n = 6 , P = 0 . 001 ) . Radiographic analysis of femora demonstrated mineralization of the condyles and patellae in the calcium supplementation group ( Fig . 5B , Snx10 KD+Ca panel ) in contrast to the untreated Snx10 KD mice ( Fig . 1I ) . Also , mutant mice receiving calcium supplementation lacked frayed or cupped metaphyseal plates ( compare 1I and 5B ) , consistent with rescue from rickets . These results were further confirmed by histomorphometry . In fact , calcium supplementation of Snx10 KD mice restored growth plate thickness ( GpTh , ) and osteoid volume per bone volume ( OV/BV , % ) to values undistinguishable form WT ( Fig . 5A , center and top right and S6 Table ) . Bone volume / tissue volume ( BV/TV , % ) , on the other hand , remained significantly higher in the treated group compared to the WT ( 64 . 21 +/− 6 . 31% and 28 . 35 +/− 8 . 68% , respectively , P = 0 . 04 , n = 3 per group , Fig . 5A , bottom right and S6 Table ) . We next analyzed femora by micro-CT ( Fig . 5C ) and confirmed restoration of cortical bone in the calcium supplementation group . Taken together , these results demonstrate that both the mortality and the mineralization defect observed in Snx10 KD mice were caused by hypocalcaemia and were prevented by calcium supplementation . To determine the role of osteoclastic expression of Snx10 on bone homeostasis in vivo , we used the recombination sites in the gene trap ( S1A Fig . ) . Snx10 KD mice were crossed with flippase-expressing mice to delete the selection cassette in exon 3 , leaving a floxed allele that deleted exons 4 and 5 of Snx10 specifically in osteoclasts when crossed with mice expressing Cre under the control of the cathepsin K promoter ( generously provided by Dr . Shigeaki Kato , Institute of Molecular and Cellular Biosciences , The University of Tokyo ) . Exons 4 and 5 encode the PX domain , which is essential for phospholipid binding . We designated this allele Snx10OC- . Snx10OC-/OC- homozygous animals are hereafter referred to as “Snx10 OC KO” , for osteoclast knockout . Snx10 OC KO mice showed a 95% reduction in Snx10 expression in bone ( Fig . 6A ) ; however expression in the stomach was not affected ( Fig . 6B ) . Snx10 OC KO mice were viable and had survival times that did not differ from their WT littermates . However , by 3 weeks of age homozygotes exhibited mild growth retardation ( Fig . 6C ) with failure of tooth eruption ( Fig . 6E and 6G versus 6D and 6F ) . Radiograph analysis showed higher radio-density but no metaphyseal plate widening , fraying or cupping of the tibia and femur , compared with the WT ( Fig . 6H-M ) . DXA analysis of 9 week-old bones revealed a significant 30% increase in Snx10 OC KO mice compared to WT ( n = 8 , 0 . 074 g/cm2 vs . 0 . 057 g/cm2 , P = 0 . 0001 ) . Micro-CT analysis demonstrated the presence of cortical bone in the femur , and the skull surface was also mineralized . Consequently , there was no "moth-eaten" appearance ( Fig . 7A , top and bottom ) . Similar to the Snx10 KD , the long bones of the Snx10 OC KO mice have marrow cavities filled with unresorbed bone ( Fig . 7A , center and 7C ) . Therefore , osteoclast-specific Snx10 deficiency resulted in osteopetrosis . Gastric pH was normal ( 3 . 06 +/− 0 . 3 ) and so was serum calcium ( 11 . 00 +/− 1 . 29 mg/dl for the Snx10 OC KO mice vs . 11 . 89 +/− 1 . 33 mg/dl for the WT , Fig . 7B , left panel ) and PTH ( 0 . 176 +/− 0 . 012 ng/ml for the Snx10 OC KO mice vs . 0 . 142 +/−0 . 023 ng/ml for the WT , Fig . 7B , right panel ) . Hematoxylin/eosin and Von Kossa/van Gieson staining of undecalcified femur sections from Snx10 OC KO mice confirmed the presence of mineralized trabeculae within the bone marrow space ( Fig . 7C , top center panel ) . However , unlike the Snx10 KD mice , the trabeculae were thoroughly mineralized and not covered by thick layers of unmineralized osteoid ( Fig . 7C , FEMUR center panel ) . The growth plate thickness ( GpTh , mm ) was significantly larger than the WT ( 0 . 160 +/− 0 . 038 mm and 0 . 092+/− 0 . 002 mm , P = 0 . 03 , n = 3 per group , S3 Table ) and so was the bone volume / tissue volume ( BV/TV , % ) . The osteoid volume per bone volume ( OV/BV , % ) , on the other hand , did not vary significantly between WT and Snx10 OC KO mice confirming that Snx10 OC KO mice are osteopetrotic but not rachitic ( S3 Fig . ) . Histomorphometry performed on sections from lumbar vertebral bodies ( Fig . 7C , VERTEBRA panel and S4 Table ) and from skull base/floor bones ( Fig . 7C , SKULL panel and S5 Table ) show similar results . We cultured osteoclast precursors from WT and Snx10 KD mice on dentine slices with M-CSF and RANKL to induce osteoclast differentiation and incubated them with the pH indicator dye , acridine orange . Confocal microscopy revealed the presence of orange fluorescence ( i . e . low pH ) in WT osteoclasts ( Fig . 8A ) . Orange label was either absent or very faint in Snx10 KD osteoclasts , indicating decreased acidification capacity ( Fig . 8B ) . The endocytic vesicles that contribute to the formation and maintenance of the ruffled border originate in the basolateral membrane of the osteoclast [1] . To study the effect of Snx10 deficiency on endosomal trafficking/vesicle formation and resorption , we performed dextran internalization assays and assessed resorption capacity on HA coated plates . WT cells internalized dextran normally ( green punctuate pattern , Fig . 8C ) . Snx10 KD cells , on the other hand , did not ( Fig . 8D ) . Ultrastructural examination of bone sections ( we examined 8 bone sections—4 sections for WT and 4 sections for Snx10 KD—and observed 3 osteoclasts per section ) by TEM demonstrated a well-developed ruffled border in WT osteoclasts ( Fig . 8E , black arrowheads ) whereas a rudimentary ruffled border was observed in Snx10-deficent osteoclasts ( Fig . 8F ) . Thus , Snx10 deficiency prevented endocytosis , severely impaired ruffled border formation , and blocked resorption of HA .
In this study we report that global Snx10-deficient mice ( Snx10 KD ) die at 3 weeks post-partum and exhibit severe osteopetrosis with a superimposed mineralization defect . Snx10 KD osteoclasts fail to endocytose dextran , form a ruffled border , or acidify the resorption lacuna . Despite these defects , CTX levels are greater in the Snx10 KD deficient mice . This is a seemingly paradoxical result , given that Snx10 deficient-osteoclasts have a severe bone resorption deficiency . However , similar results have been reported for severely osteopetrotic mice with deletions in the osteoclastic anion exchanger 2 ( Ae2 ) [28] , and with loss-of-function mutations in the chloride channel 7 ( clc-7 ) [34] and in cathepsin K [30] . This phenomenon may reflect the fact that in osteoclast-rich osteopetrosis , as in these examples , there is a very large increase in bone mass and bone surface area , resulting in a corresponding increase in the total number of osteoclasts . One study of clc-7 KO mice [34] demonstrated that CTX per individual osteoclast was reduced by 50% compared to the WT , but because there also was an increase in the total number of osteoclasts , the total serum CTX was increased . We also observed a mineralization defect in the Snx10 KD mice , consistent with rickets , suggesting that Snx10 is also required for calcium homeostasis . Pursuing this , we found Snx10 to be expressed in gastric zymogenic cells . Similar to osteoclasts , Snx10 KD zymogenic cells exhibit a defect in secretory vesicle formation/maintenance , leading to hypochlorhydria and hypocalcemia . The Snx10 KD mice thus exhibit a phenotype that is a combination of osteopetrosis ( due to impaired osteoclast resorption ) and rickets ( impaired mineralization due to poor calcium absorption ) . This is borne out by the correction of the rachitic phenotype when Snx10 is knocked out specifically in osteoclasts , with normalization of stomach pH , circulating calcium , and bone mineralization . Our results are in line with those of Schinke et al . [11] who analyzed the combined acid defect phenotype by using mutations in different genes ( Cckbr for the stomach and Src for the osteoclasts ) . Similarly , the Tcirg1-deficient oc/oc mice have an osteopetrorickets phenotype due to impaired acidification in both stomach and bone [11] . These observations unify the complex phenotype seen in Snx10 KD mice and osteopetrorickets patients as being due to the simultaneous inhibition of osteoclast function and gastric acid production . Shinke et al [11] reported that 21 patients with diagnosed osteopetrosis had not been genotyped and 10 of these patients had histological evidence of osteopetrorickets that had not been diagnosed . Because calcium supplementation rescued mice from the rachitic phenotype and dramatically extended life-span in Snx10 KD mice , our findings suggest that this may be a critical component of the clinical approach to Snx10-dependent human osteopetrorickets . Because calcium gluconate is taken up at higher pH than calcium carbonte , it is the form of choice . There is a complex system of cross-regulation that exists between bone and most other organs [35] . Our studies provide further evidence that there are instances of shared molecular machinery for the production of acid in the stomach and in osteoclasts . TreeFam , a database of phylogenetic trees from animal genomes that can be used to infer the evolutionary history of genes , shows that Snx10 appears first in Osteichthyes , the bony fishes ( http://www . treefam . org/family/TF332117 ) . The acid-secreting stomach also evolved in bony fishes . This raises the intriguing prospect that the osteoclastic and gastric acid producing systems evolved simultaneously during vertebrate evolution , together with emergence of a mineralized skeleton , utilizing some of the same genes . Snx10 appears to mediate both bone resorption and stomach acidification by regulating vesicular trafficking . Our data and several other reports support a concordance between the secretory trafficking mechanisms in OCs and zymogenic cells . For example , GNPTAB [36–38] , the enzyme required for adding mannose-6-phosphate to lysosomal hydrolases , when knocked out in mice , causes defective OC and zymogenic cell trafficking . Gnptab-null mice also have low BMD [38 , 39] . In the stomach , trafficking of digestive enzymes like PGC and acid secretion ( and , thus , calcium homeostasis ) are linked . Loss of parietal cell acid secretion and/or damage to parietal cells causes loss of mature zymogenic cells [31 , 40] . It is well-known that the fate and function of parietal and zymogenic cells are profoundly intertwined [41] . Further studies may reveal whether there is Snx10 expression in parietal cells , which may be too low to be seen in the assays used here or if their function depends indirectly on Snx10 expression in zymogenic cells . Gastric acidification is essential for calcium absorption , and this regulatory pathway may be disrupted in a variety of clinical conditions associated with hypochlorhydria ranging from osteopetrorickets to antacid therapy . Proton-pump inhibitors , which block gastric acid production and increase gastric pH , have recently been associated with hip fracture risk [42 , 43] . It has been speculated that hypochlorhydria results in impaired calcium absorption thereby increasing fracture risk . Similarly , endosteal bone resorption has been reported following gastric bypass surgery in rodents and humans [44 , 45] . The marked hyperparathyoirism of the Snx10 KD mice was corrected in the Snx10 OC KO model , consistent with the hypocalcemia in the former which was prevented by normal gastric acdification in the latter . Thus , our findings underscore the relationship between gastric acidification and calcium absorption and its impact on bone health in the general population ( for an authoritative review on the topic see [46] ) , and provide novel insights into the mechanisms governing the regulation of bone accrual via the gastrointestinal tract .
All animal procedures were approved by IACUC of The Forsyth Institute . We obtained the Snx10 targeting vector , PG00216_Z_2_C06 , from the European Conditional Mouse Mutagenesis Program ( EUCOMM ) . This vector is a “knockout first” gene trap ( see S1A Fig . ) which inserts a flippase site-flanked Neo selection cassette with an IRES and LacZ reporter into intron 3 and inserts LoxP sites flanking exons 4 and 5 . Exons 4 and 5 contain the PX domain required for phospholipid interactions , including all three phospholipid contact amino acid sequences . This allele is designated Snx10tm1a ( EUCOMM ) Hmgu . Hereafter in this report , we refer to the resulting targeted allele as Snx10Neo-f . Mice homozygous for this allele are severely deficient in Snx10 globally ( see Results ) , so we hereafter refer to Snx10Neo-f/Neo-f mice as “Snx10 KD , ” for Snx10 knockdown . The gene-trap vector was electroporated into V6 . 5 ES cells . Neomycin resistant clones were picked , expanded and screened for correct insertion by Long Range Genomic PCR using the following vector-specific primers and gene-specific primers: 5' Integration Gene Specific Forward ( GF3 ) . 5'-GCTTATGGTCGACTCATCGGAGAATC-3' LacZ Reverse ( LAR7 ) , 5'-GGTGTGGGAAAGGGTTCGAAGTTCCTAT-3' Amplicon size = 5 , 170 bp 3' Integration LacZ Forward ( LAF ) , 5'-GAGATGGCGCAACGCAATTAATG-3' Gene Specific Reverse ( GR4 ) , 5'-CACAGAAGTAATGTACGCTAATGGCAACG-3' Amplicon size = 5 , 680 bp The resulting clones were injected into host blastocysts to generate mouse chimeras . Two male chimeras were bred with C57BL/6J females . Germ line transmission was confirmed by PCR , using primers flanking the third loxP site ( LoxP3 Forward: 5’-ATAACTAACCCAGGCAAACA-3’ and LoxP3 Reverse: 5’-TTGTCAAGTGCGTGTGTCGT-3’; S1A Fig . , red arrows ) . Snx10Neo-f/+ offspring were bred to homozygosity to generate animals for experiments at the expected Mendelian ratio of approximately 25% . PCR genotyping using the preceding primer pair produced bands of 213 bp for the WT ( +/+ ) , 273 bp for the Snx10Neo-f/Neo-f , and both bands for the heterozygotes ( S1B Fig . ) . For qPCR amplification of cDNA we used the following primers ( S1A Fig . , black arrows ) : SNX10 ( Forward ) : 5’-GAACAATCGCCAGCATGTCGAC-3’ and SNX10 ( Reverse ) : 5’-ATGTCC TCGGAGTTCAGATGGC-3’ . Osteoclast-specific Snx10-deficient mice were generated in two steps . First , the Neo cassette was removed by crossing Snx10Neo-f/+ females with males homozygous for Rosa26-driven FLP recombinase ( strain B6 . 129S4-Gt ( ROSA ) 26Sortm1 ( FLP1 ) Dym/RainJ ) [47] , resulting in an allele with exons 4 and 5 flanked by loxP sites ( Snx10f ) . Snx10f/f animals were viable and fertile . Then , heterozygous males were crossed with females homozygous for Cre recombinase driven by the cathepsin K promoter ( Ctsk-Cre ) [48] . The resulting mice carry an osteoclast-specific Snx10 allele with exons 4 and 5 deleted , and therefore a null allele , which we designate Snx10OC- . Snx10OC-/OC- homozygous animals are hereafter referred to as “Snx10 OC KO” , for osteoclast knockout . Radiographs were obtained with a Faxitron cabinet radiograph system ( Model 43855A , Hewlett Packard , McMinnville , Oregon ) with Kodak high-speed holographic film at 40 kV . For micro-CT , mice were sacrificed by CO2 asphyxiation . Left femora , tibiae , mandibles and calvariae were then dissected and fixed in 4% paraformaldehyde ( PFA , pH = 7 . 4 ) at 4°C for 18 hours , followed by 70% ethanol , then stored at 4°C until scanned . Samples were scanned ( μCT 40 , Scanco Medical ) for 3D reconstruction and the following parameters were estimated: BV/TV ( bone volume per tissue volume ) , Tb . N ( trabecular number ) , Tb . Th . ( trabecular thickness ) , Tb . Sp ( trabecular spacing ) , and bone mineral density ( BMD ) . Micro-CT slices , each with a slice thickness reconstructed to 10 μm ( 100 slices/mm ) , were obtained delivering a 3-dimensional representation of approximately 3 mm of anatomy . The imaging was done at 55kV , 145μA . We used splenocytes as the source of osteoclast precursors because long bones in osteopetrotic mice lack marrow cavities . Spleens were resected from 4 week-old mutant and WT mice and mashed through a cell strainer ( BD Falcon , Mesh Size: 40μm ) To induce osteoclast differentiation , splenocytes were plated in a 24-well plates ( 1 x 106 cells/well ) in alpha-MEM/10% FBS supplemented with M-CSF ( 25 ng/ml ) and RANKL ( 50 ng/ml ) ( both from Peprotech ) . Cells were cultured for 7 days with changes of medium and cytokines every other day . Spleen cells were cultured overnight in alpha-DMEM/10% FBS supplemented with 50ng/ml M-CSF in 24-well plates ( 1 x 106 cells/well ) . The next day , the medium was removed and the cells were incubated overnight with lentiviral particle-containing medium in the presence of polybrene ( 6ug/ml ) . The following day , the culture medium was replaced with fresh medium containing M-CSF and RANKL . The cells were cultured under osteoclast differentiation conditions for 7 days , with changes of medium and cytokines every other day . Spleen cells were plated ( 1 x 106 cells/well ) on 24-well Osteo Assay Surface plates ( Corning ) in alpha-MEM supplemented with 50ng/ml M-CSF . The following day cells were infected with Snx10 lentiviral particles and 24 hours later the medium was replaced with fresh medium containing M-CSF and RANKL . The cells were cultured under osteoclast differentiation conditions for 7 days as above . Cells were then removed by incubation with 10% bleach solution . The wells were washed with PBS , air-dried and photographed under a light microscope . The area of resorption pits was quantified using Image J ( National Institutes of Health , Bethesda , MD ) . Osteoclast precursors were cultured on bovine dentine slices in alpha-MEM supplemented with M-CSF and RANKL to induce osteoclast differentiation . Differentiated osteoclasts were incubated with 200 μg/ml aldehyde fixable Alexa 488-dextran ( Life Technologies , Grand Island , NY , USA ) overnight at 37°C . Cultures were fixed in 4% PFA for 15 minutes , counterstained with DAPI ( 4 , 6-diamidino-2-phenylindole ) to visualize nuclei , and mounted on glass cover slips for fluorescence microscopy . Digital images ( obtained using fluorescence microscope FSX100 , Olympus ) were processed with Image J to obtain an interactive 3D surface plot , where the height of the surface plot reflects image brightness , corresponding to the amount of internalized dextran . Multinucleated cells were counted and each cell was assessed for dextran endocytosis . Osteoclasts cultured on dentine slices as above were incubated with 5ug/ml acridine orange ( Sigma ) for 15 minutes at 37°C , rinsed with PBS and chased for an additional 15 minutes . Images were obtained using a Leica SP5X Laser Scanning Confocal Microscope with a 490-nm excitation filter and a 525-nm emission filter . Digital images were processed with Image J . Multinucleated cells were counted and assessed for the presence of extracellular acidification . Experiments were performed in triplicate and the results were expressed as the proportion of acidification-positive osteoclasts in the total number of osteoclasts ( ±sd ) . Differences were analyzed using Student’s t-test and considered significant if P<0 . 05 . After culturing cells for 7 days in the presence of RANKL and M-CSF , cultures were washed with PBS , fixed first in 4% PFA for 5 minutes , then briefly in ethanol/acetone , 50%/50% , and air dried for 2 minutes . Cells were then incubated in tartrate resistant acid phosphatase ( TRAP ) staining solution ( Napthol AS-MX phosphate and Fast Red Violet LB Salt ) at 37°C until the color developed ( 10 minutes to 1 hour ) . The wells were then washed with PBS , air-dried and photographed under a light microscope . Mouse femora were fixed in 4% PFA overnight at 4°C and stored in 50% alcohol at room temperature . For paraffin embedding , the specimens were washed with 5 , 10 , and 15% glycerol in PBS , each for 15 min and decalcified with 10% EDTA in 0 . 1M TRIS for 2 weeks . Specimens were then embedded in low melting temperature paraffin and sectioned at 5-μm intervals . Decalcified paraffin sections were used for H&E , IHC and TRAP staining . For plastic embedding , the specimens were dehydrated in a graded series of alcohols then infiltrated and embedded in a methyl and butyl methacrylate resin . Each femur was bisected longitudinally with a precision saw and cross-sections ( transverse ) were cut at the mid-diaphysis ( plane 1 ) and longitudinal planes in the proximal-to-distal orientation ( plane 2 ) . Plane 1 and 2 segments were sectioned using a Leica RM2165 microtome and collected onto coated slides , press-mounted and dried in an oven set at 50°C for several hours . Slides were deplasticized in xylenes , rehydrated in a graded alcohol series into deionized water , stained with Von Kossa-Van Gieson’s stains [11] , and covered with a cover slip . Micro-ground sections were prepared from planes 1 and 2 sawed segments , mounted onto plastic slides , then ground and polished to a thickness of <80 microns using the EXAKT CS400 system ( EXACT Technologies , Oklahoma City , OK ) . The micro-ground slides were etched in a solution of 50% acetone + 50% dehydrated ethanol , rinsed in deionized water , then stained with Von Kossa-Van Gieson’s stain , and covered with a cover slip . Freshly obtained stomach samples were fixed overnight in 4% formalin at 4°C , rinsed in 70% ethanol , arranged in 2% agar in a tissue cassette , and paraffin embedded . 5μm sections were cut , paraffin was removed , and the sections were rehydrated . Antigen retrieval was done for 20 minutes in boiling Trilogy solution ( Cell Marque , Rocklin , CA ) , blocked for 1 hour in 1% bovine serum albumin , 0 . 3% Triton X-100 in PBS and incubated with primary antibodies overnight . Primary antibodies used were Snx10 ( Santa Cruz SC-104657 , 1:200 ) , VEGF-B ( Santa Cruz SC-1876 , 1:250 ) , and GIF ( a gift from Dr . David Alpers , Washington University School of Medicine , 1:2000 ) . Secondary antibodies , GSII lectin and Hoechst 33358 labeling were as described [32] Using morphometric software ( i . e . , Olympus MicroSuite Biological Suite or cellSens ) , five high magnification images of methylmethacrylate sections were acquired from the sub-articular region ( i . e . , immediately subjacent to the growth plate ) of the proximal femur ( 40x ) , the bones of the skull base/floor ( i . e . , basisphenoid ) ( 40x ) and lumbar vertebral bodies ( 20x ) . Calibrated images of the femora , calvariae , and vertebrae were then morphometrically assessed to determine: osteoid volume/bone volume ( OV/BV , % ) and bone volume/tissue volume ( BV/TV , % ) , and , in addition , the growth plate thickness ( GpTh , mm ) of distal femur , all as described [49] . Bone and osteoid volume were calculated based on the summation of all five sampling fields . Ratios ( i . e . , OV/BV and BV/TV ) are presented as the average of all five fields . Since our morphometry was done in 2D we measured areas and then calculated volumes ( i . e . , by multiplying perimeter area values 4/π to getsurface/volume ) . Data were consolidated and presented as mean ± sd Mice were perfused with 4% ( w/v ) PFA and 1% ( v/v ) glutaraldehyde in 0 . 1 M phosphate buffer ( pH 7 . 4 ) . Femora were isolated , defleshed and immediately fixed in 4% ( w/v ) PFA and 1% ( v/v ) glutaraldehyde in 0 . 1 M phosphate buffer ( pH 7 . 4 ) at 4°C for 1 h prior to cutting the femur into approximately 1 mm slices , which continued to be fixed for additional 24 h and then rinsed in 0 . 1M sodium phosphate buffer . Bone slices were decalcified in 10% ( w/v ) EDTA in PBS for 3 days , post fixed in 1% ( v/v ) OsO4 and embedded in Epon-812 resin ( Tousimis , Rockville , MD ) . Ultrathin sections ( n = 4 per group ) were then made and stained with uranyl acetate and lead citrate , and examined with a FEI Tecnai G2 Spirit Transmission Electron Microscope . Gastric pH was measured as described [50] . Briefly , mice were fasted for 2 hours prior to the test and anesthetized . The stomach was then exposed after abdominal midline incision , followed by ligation of the pylorus and esophagus . Saline solution was then injected into the stomach and the fluids collected . The pH of the collected fluid was measured using a Star A321 pH Meter ( Thermo Scientific Orion ) equipped with a PerpHecT ROSS pH electrode with micro tip ( Thermo Scientific Orion ) . For analysis of calcium homeostasis , blood was obtained by cardiac puncture and collected in heparinized Eppendorf tubes . Serum was isolated by centrifugation at 3000 g for 10 min and stored at −80°C . Serum calcium was measured by the Clinical & Epidemiologic Research Laboratory at Children’s Hospital ( Boston , MA ) , a nationally certified laboratory , using a colorimetric assay on a Hitachi 917 analyzer using Roche reagents ( Roche Diagnostics , Indianapolis , IN ) . The lowest detection limit of this assay is 0 . 2 mg/dL and the day-to-day imprecision values at concentrations of 8 . 38 , 9 . 12 and 13 . 13 mg/dL are 1 . 5% , 1 . 6% , and 0 . 8% , respectively . PTH was determined using the Mouse PTH 1–84 ELISA Kit ( Cat . # 60-2305 , Immutopics , Inc . , San Clemente , CA ) . 1 , 25- dihydroxyvitamin D was determined using the Mouse Vitamin D , VD ELISA Kit ( CSB-E07912m , Cusabio Life Science ) . Quantification of fragments of type I collagen in serum was done using the RatLap EIA kit from IDS ( Immunodiagnostic Systems , AC-06F1 ) following the manufacturer's instructions . Serum osteocalcin was measured using the Mouse Osteocalcin ELISA Kit from Immutopics ( 60-1305 ) , following the manufacturer's instructions . | We found that Snx10 , a molecule expressed in osteoclasts , was also expressed in the stomach . Studies in tissue specific or global knock-down mice showed that Snx10 deficiency resulted in a phenotype that was a consequence of deficiencies in both osteoclasts and gastric zymogenic cells . Our studies add to a growing list of genes , including atp6i ( Tcirg1 ) , whose expression is required both in bone and stomach to maintain normal gastric acidification and calcium absorption . This work provides additional insight into the mechanisms governing the regulation of bone accrual by the gastrointestinal tract . Because osteopetrorickets has not been described clinically in Snx10-related osteopetrosis , these findings highlight the importance of considering impaired acidification in both stomach and bone in osteopetrotic patients with mutations in SNX10 and other genes with similar patterns of expression and activities . Because defects in gastric differentiation and/or gastric acidification may cause or contribute to hypocalcemia , bone insufficiency , and early death , our results suggest that dietary calcium supplementation could be a life-saving intervention in these patients . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Osteopetrorickets due to Snx10 Deficiency in Mice Results from Both Failed Osteoclast Activity and Loss of Gastric Acid-Dependent Calcium Absorption |
IRF8 ( Interferon Regulatory Factor 8 ) plays an important role in defenses against intracellular pathogens , including several aspects of myeloid cells function . It is required for ontogeny and maturation of macrophages and dendritic cells , for activation of anti-microbial defenses , and for production of the Th1-polarizing cytokine interleukin-12 ( IL-12 ) in response to interferon gamma ( IFNγ ) and protection against infection with Mycobacterium tuberculosis . The transcriptional programs and cellular pathways that are regulated by IRF8 in response to IFNγ and that are important for defenses against M . tuberculosis are poorly understood . These were investigated by transcript profiling and chromatin immunoprecipitation on microarrays ( ChIP-chip ) . Studies in primary macrophages identified 368 genes that are regulated by IRF8 in response to IFNγ/CpG and that behave as stably segregating expression signatures ( eQTLs ) in F2 mice fixed for a wild-type or mutant allele at IRF8 . A total of 319 IRF8 binding sites were identified on promoters genome-wide ( ChIP-chip ) in macrophages treated with IFNγ/CpG , defining a functional G/AGAAnTGAAA motif . An analysis of the genes bearing a functional IRF8 binding site , and showing regulation by IFNγ/CpG in macrophages and/or in M . tuberculosis-infected lungs , revealed a striking enrichment for the pathways of antigen processing and presentation , including multiple structural and enzymatic components of the Class I and Class II MHC ( major histocompatibility complex ) antigen presentation machinery . Also significantly enriched as IRF8 targets are the group of endomembrane- and phagosome-associated small GTPases of the IRG ( immunity-related GTPases ) and GBP ( guanylate binding proteins ) families . These results identify IRF8 as a key regulator of early response pathways in myeloid cells , including phagosome maturation , antigen processing , and antigen presentation by myeloid cells .
The defense mechanisms of mononuclear phagocytes that are circumvented by successful intracellular pathogens are poorly understood [1] . Genes and proteins in these pathways may represent valuable targets for therapeutic interventions in the corresponding diseases . Such host defense mechanisms can manifest themselves as genetic determinants of innate resistance or susceptibility to infections in human populations [2] , [3] , and in corresponding animal models of experimental infections [4] , [5] . Forward genetic studies of naturally occurring or experimentally induced mutations in mice may identify such genes and proteins [5] , [6] , which relevance to the corresponding human infection can be established in parallel studies of human populations from areas of endemic disease [2] , [5] , [6] . In inbred mouse strains , susceptibility to infection with several intracellular pathogens including Mycobacterium , Salmonella and Leishmania , is determined in part by the natural resistance-associated macrophage protein 1 ( Nramp1 ) gene ( Slc11a1 ) . In resistant mice , Slc11a1 functions as an efflux pump for Fe2+ and Mn2+ ions at the membrane of microbe-containing phagosomes formed in macrophages , thereby restricting microbial access to these essential nutrients [7] . In humans , polymorphic variants at or near SLC11A1 have been associated with differential susceptibility to mycobacterial infections including tuberculosis , leprosy , and Buruli ulcer [6] . In addition , monocytes derived from individuals bearing SLC11A1 alleles associated with tuberculosis susceptibility in field studies , display reduced functional activity of the SLC11A1 protein [8] . A search for genetic modifiers of the protective effect of Slc11a1 identified the BXH2 mouse strain as highly susceptible to Mycobacterium bovis ( BCG; bacillus Calmette-Guérin ) infection despite presence of resistance-associated Slc11a1 alleles ( Slc11a1Gly169 ) [9] . By positional cloning , we determined that susceptibility to infection in BXH2 is caused by a mutation ( R294C ) in the interferon regulatory factor 8 ( IRF8 ) ( Ensembl:ENSMUSG00000041515 ) [9] . IRF8 is one of 9 members of the Interferon Regulatory Factor ( IRF ) family . IRF8 has a DNA binding domain ( DBD; 120 a . a ) of the helix-turn-helix type that binds to ISRE ( Interferon Stimulated Response Elements ) sites present in the proximal promoters of type II IFN-regulated genes . IRF8 also has an IRF association domain ( IAD ) that serves as a recruitment module for other transcription factors . IRF8 is expressed primarily in macrophages and dendritic cells , but is also detected in T and B lymphocytes [10]; and upon stimulation with interferon gamma ( IFNγ ) , lipopolysaccharide ( LPS ) , pathogen-associated molecular patterns ( PAMPs ) and other microbial stimuli , IRF8 binds to ISREs in association with other members of the IRF ( e . g . IRF1 ) , or ETS ( e . g . PU . 1 , TEL ) families , or other hetero-dimerization partners , to activate or repress gene expression in these cells [11] . The IRF8-PU . 1 heterodimer leads to the activation of genes containing ETS-IRF composite element ( EICE , GGAAnnGAAA ) , the ETS-IRF response element ( EIRE , GGAAAnnGAAA ) or to the IRF-ETS composite sequence ( IECS , GAAAnn ( n ) GGAA ) [12] . IRF8 plays an important role in several physiological aspects of myeloid cells development and function . IRF8 drives differentiation of myeloid progenitors towards mononuclear phagocytes , while positively regulating apoptosis of the granulocytic lineage [11] , [13] . Macrophages from IRF8-deficient mice remain immature , including altered expression of intrinsic macrophage anti-microbial defenses [11] , and are susceptible to ex vivo infection with M . bovis [14] , Salmonella typhimurium [14] , and Legionella pneumophila [15] . IRF8-deficient mice also show a profound defect in dendritic cells ( DCs ) , as they lack both CD11c+CD8α+ DCs and pDCs [16] . In addition , the small number of CD11c+CD8α+ and CD8α− DCs present in these mice remain immature and fail to up-regulate co-stimulatory molecules and to produce key cytokines in response to microbial products [16]–[18] . In addition , IRF8 is required for Th1 polarization of early immune response [11] . This cooperation between antigen presenting cells ( APCs ) and T/NK cells , involves IFNγ binding to its receptor ( IFNγR ) which causes STAT1 ( signal transducers and activators of transcription ) activation . STAT1 trans-activates IRF8 expression leading to IL-12p40 production by dendritic cells , and engagement of the interleukin 12 receptor ( IL12R ) on Th1 cells further amplifies IFNγ production [11] . IRF8 binds to the promoter regions , and is required for activation of IL-12p40 [19] , [20] , IL-12p35 and IL-18 genes in DCs in response to IFNγ [11] , [19] , [20] . IRF8−/− mice do not produce IL-12p40 , lack Th1 polarization ( absence of antigen specific CD4+ , IFNγ producing T cells ) , and are susceptible to in vivo infection with intracellular pathogens [19] , [21]–[24] . We have shown that the IRF8R294C isoform of BXH2 behaves as a partial loss-of-function which is associated with impaired IL-12p40 production by BXH2 splenocytes , and loss of trans-activation of a IL-12p40 reporter construct in vitro . The IRF8R294C mutation results in increased M . bovis ( BCG ) multiplication both early and late during infection , with uncontrolled replication linked to inability to form granulomas in infected liver and spleen . The IRF8R294C mutation also causes susceptibility to S . typhimurium to a level comparable to that seen for mice lacking functional Nramp1 or Tlr4 ( Toll-like receptor 4 ) , and impairs innate and adaptive immune defenses against the blood-stage malarial parasite Plasmodium chabaudi AS [25] . BXH2 mice are also extremely susceptible to aerosol infection with Mycobacterium tuberculosis , showing uncontrolled intracellular pathogen replication in lung macrophages , impaired granuloma formation , rapid dissemination of the infection to distant sites , and rapid necrosis of infected tissues , and early death . There was complete absence of IL-12p40 induction , severely reduced IFNγ production , and impaired T cell priming in the lungs of infected BXH2 , highlighting the critical role of IRF8 in this response [26] . These studies have identified IRF8 as a key regulator of host defenses against Mycobacteria . In this study , we have used transcript profiling with microarrays and chromatin immunoprecipitation ( ChIP ) hybridization on genomic DNA arrays ( ChIP-chip ) in macrophages from normal and IRF8-deficient mice , to systematically identify genes transcriptionally regulated by IRF8 a ) during ontogeny and maturation of macrophages , and b ) in response of these cells to combined exposure to IFNγ and Tlr9 ( Toll-like receptor 9 ) ligand ( CpG ) , and c ) during pulmonary tuberculosis in vivo . In these studies , we incorporated an experimental strategy based on the co-segregation of IRF8-dependent differential gene expression in macrophages from [BALB/c×BXH2] F2 animals selected for homozygosity for either wild-type ( wt; IRF8R294 ) or mutant ( IRF8C294 ) IRF8 alleles . These studies have identified a critical role for IRF8 in regulating expression of genes and associated cellular pathways responsible for early interaction with pathogens , phagosome maturation , antigen processing and antigen presentation to CD4+ and CD8+ T cells .
To identify transcriptional targets of IRF8 that play a role in a ) macrophage maturation , and b ) in activation in response to IFNγ and microbial products , we used transcript profiling to compare RNA expression in macrophages bearing either a wild-type ( wt ) or a mutant allele at IRF8 ( R294C ) . For this , we used bone marrow-derived macrophages ( BMDMs ) from individual [BALB/c×BXH2] F2 mice of mixed genetic background but that were identified as homozygote for either wt ( IRF8R294 ) or mutant ( IRF8C294 ) IRF8 alleles . This strategy [27] is based on the observation that complex gene expression profiles ( eQTLs ) caused by a null mutation at a specific gene show extremely robust segregation in F2 animals [28] , congenic strains [29] or recombinant congenic lines [30] derived from parental strains bearing wt and mutant alleles at the gene of interest . Gene expression profiles detected in common in macrophages from F2 animals of either wt or mutant IRF8 genotypes but that show mixed genetic background ( C57BL/6J , C3H/HeJ , BALB/cJ ) , can distinguish true IRF8-dependent effects from irrelevant ones caused by differences in genetic background of the two parental mouse strains . This strategy is well suited to study eQTLs caused by absence versus presence of a transcription factor such as IRF8 . In this approach , individual F2 mice ( 6 samples per experimental group ) are used both as biological and technical replicates , to increase the stringency of the analysis . BMDMs from wt and IRF8C294 F2 mice were stimulated or not with IFNγ/CpG , and RNA was isolated and used for transcript profiling . IRF8 plays a critical role in maturation of monocytes , macrophages and dendritic cells , and mice bearing mutations at IRF8 have defects in these cell types [11] . To identify IRF8 transcriptional targets that may play a role in maturation of the myeloid lineage , we compared transcript profiles in resting BMDMs from wt and IRF8 mutant F2 mice ( Figure 1A ) . A pairwise analysis ( t test p value<0 . 05; fold change ≥1 . 5X ) identified a total of 454 genes differentially expressed in an IRF8-dependent fashion in these cells at basal level ( Table S1 ) . Of these 454 genes , 219 were more highly expressed in wt cells , while 235 were more highly expressed in mutant BMDMs ( Table S1 ) . Hierarchical clustering of the 454 genes according to expression pattern similarities in the 12 independent microarrays readily separated the 6 individual wt mice from the 6 individual mutant mice ( data not shown ) , illustrating the robustness of the approach . A gene ontology ( GO ) report on these 219 and 235 genes separately , revealed that 39 ( 17 . 8% ) and 45 ( 19 . 1% ) of them were associated with ‘response to stimulus’ , representing the most abundant group ( data not shown ) . Additional enriched gene clusters included GO-terms such as immune system development , immune system process , response to stress , intracellular signaling cascade , immune response , defense response , transcription , and others . Genes most positively regulated by IRF8 in resting cells included genes involved in a ) antigen processing and presentation ( CD74 , H2-AbI , H2-Eb1 , H2-Ea ) , b ) cytokines and chemokines production and signaling ( Cxcl14 , Cxcl16 , Socs2 , Ciapin1 , the C1q complex , Il17ra ) , c ) growth regulation ( Csf3r ) , d ) tissue remodeling ( Timp1 , Vcam1 ) , and e ) rapid response to microbial insults ( Mx1 , Ifitm1 , Tnfaip3 , Ly86 ) [see Table S1 for annotation] . Together these genes may correspond to direct IRF8 targets or may represent markers of maturation differentially expressed in response to the block caused by loss of IRF8 function in BMDMs . To systematically identify IRF8 targets that are important for IFNγ-induced macrophage activation , we compared gene expression profiles obtained in BMDMs from wt and IRF8 mutant F2 mice following exposure to IFNγ/CpG ( stimulated versus control ) . A first pairwise analysis ( t test p value<0 . 05; fold change ≥1 . 5X ) identified a total of 2501 ( 1247 induced; 1254 repressed ) and 1904 ( 828 induced; 1076 repressed ) genes significantly regulated by IFNγ/CpG in wt and IRF8 mutant mice , respectively ( Figure 1A ) . A subset of these genes ( 76 genes in wt only , 40 genes in IRF8 mutant only , and 138 genes in both groups ) were also regulated by IRF8 at the basal level , in the absence of IFNγ/CpG stimulation ( Figure 1B and Table S2 ) . Secondly , and to take into account possible IRF8-dependent expression differences at basal level , we carried out a two-way ( 2×2 interaction ) Anova analysis [29] . In this analysis , expression levels before and after IFNγ/CpG treatment are calculated and expressed as ratios , and a statistical analysis is conducted to identify genes which ratio of expression are affected by IRF8 [wt , stimulated versus unstimulated; compared to IRF8 mutant , stimulated versus unstimulated] with a t-test p value<0 . 05 and a fold change ≥1 . 5X . This comparison identified 368 genes that were significantly regulated by IRF8 in response to IFNγ/CpG ( Figure 2A and Table S3 ) . Hierarchical clustering according to expression pattern similarities not only distinguished the control from treated groups ( IFNγ/CpG ) , but also separated wt from IRF8 mutant BMDMs and this for both conditions ( dendrogram in Figure 2A ) . A subset of 80 genes showed particularly robust IRF8 dependence in expression in response to IFNγ/CpG , while showing no significant IRF8-dependent effects at basal levels ( indicated in bold in Table S3 ) . This list contained many genes known to play a key role in several aspects of macrophage function , including a ) cytokine-cytokine receptor interaction ( Ccl8 , Ccr3 , Il13ra1 ) , b ) antigen presentation ( H2-DMb2 , Ciita ) , c ) tissue remodeling ( Angptl4 , Col18a1 , Mmp13 ) , d ) detoxification ( Cyp27a1 , Cyp4f18 , Cyp51 , Por , Ephx1 ) , e ) cell surface receptors ( Igh-6 , Tfrc ) and adhesion molecules ( Siglec1 ) , and Irf4 , a member of the IRF family know to functionally interact with IRF8 to regulate gene expression [11] . A subset of 8 transcripts ( Ephx1 , Cyp27a1 , Ciita , Il10ra , Ms4a7 , C1qb , Angptl4 , and Slc40a1 ) strongly induced by IFNγ/CpG in an IRF8-dependent manner , were selected for further validation by quantitative PCR ( qPCR ) . For all the genes tested , we observed an excellent correlation between the level and degree of differential expression initially detected by transcript profiling and results from qPCR analysis ( Figure 2B and 2C ) . Transcript profiling analyses revealed that IRF8 intervenes in a complex transcriptional network . To identify which genes in this network are direct IRF8 transcriptional targets of ( as opposed to secondary targets ) , we hybridized IRF8-bound chromatin obtained by immunoprecipitation ( ChIP ) from cultured macrophages treated with IFNγ/CpG to Agilent promoter tiling arrays ( ChIP-chip ) . Following normalization and statistical analysis , we identified 319 IRF8 binding events corresponding to 333 different genes ( Table S4 ) . These binding sites were selected for a ) a minimum of 2-fold enrichment over control ChIP carried out using non-immune serum , and b ) a p value≤0 . 001 . In this list , we validated IRF8 recruitment to Ifnβ promoter by ChIP-qPCR ( data not shown ) . Moreover , this list contains several published IRF8 binding sites ( Tlr4 , Oas2 , Cybb , Ifitm3 , Etv3 , Lyz and Tlr9 ) and shows a significant 43% overlap with the recently published IRF8 ChIP-chip study performed on chromatin from human monocytes [31] . To analyse closely the IRF8 binding sequence , we determined the chromosomal position for the center of each binding peak , and extracted 500 bp of peak flanking sequence using the mouse mm8 genome assembly . These sequences were queried for de novo motif discovery with different algorithms ( MEME , MDscan and AlignAce ) , and all produced the same IRF8 DNA binding motif ( G/AGAAnTGAAA ) as the top matrix ( Figure 3A and Figure S1A ) [32]–[34] . This highly significant motif ( MEME E-value = 8 . 3−385 ) is in agreement with the known Transfac database IRF8 binding motif ( ICSBP_M00699; Figure S1B ) , although there is no requirement for the 3′ CTG bases that are more characteristic of the ISGF3 ( Stat1/Stat2/IRF9 ) ISRE binding site ( Figure 3A ) . This de novo binding site is closer to a standard IRF site which is characterized by a 2 nucleotide spaced tandem repeat of GAAA , with an important difference , the first base is mostly occupied by a guanine . A comparable GGAAnnGAAA motif was previously described as ETS-IRF composite element ( EICE ) [12] . However , the motif identified in the present study gives importance to a T placed in the sixth position . This de novo derived motif was found at least once within a 1000 bp segment of 87% ( 277 out of 319 ) of the IRF8 binding sites identified by ChIP-chip . A comparison of the de novo defined IRF8 site with known ISRE and IRF1 binding motif show they all cluster at the peak of enrichment , with the highest number for the de novo IRF8 site ( Figure 3B ) . Other predominant motifs were identified in our dataset by de novo analysis . Their similarity to known Transfac v11 . 3 database was assessed using the STAMP web-tool [35] . With the MatInspector motif search tool , we measured the fold enrichment of each de novo and known motifs occurrence in our dataset compared to similar sets of random sequences ( Table S5 ) . As expected , all the matrices from IRF family were enriched; ETS family motifs were also enriched because the GGAA sequence which forms part of their binding site ( GAGGAA ) is imbedded within the IRF8 binding site . We detected a strong association ( ∼50% of sites ) between the de novo generated IRF8 motif and binding sites for PU . 1 , the major ETS factor in macrophages , with co-localization of the two sites at the binding peak ( Figure 3C ) . In addition , we noted an enrichment of AP-1 sites: of the 277 IRF8 motif containing peaks , 73 ( 26% ) also contain an AP-1 predicted site with a tendency of these sites to be centered at the peak of enrichment , although not as clearly as for PU . 1 ( Figure 3D ) . A Gene Ontology ( GO ) analysis with the DAVID ( database for annotation , visualization and integrated discovery ) web-tool for genes exhibiting an IRF8 binding peak detected by ChIP-chip revealed a strong enrichment for the “immune response” category ( 29 genes , p value = 5 . 3e10−9 ) ( Table S6 ) [36] , [37] . This list includes several genes encoding proteins involved in recognition , processing and presentation of antigens by antigen-presenting cells ( APCs ) . Indeed , it includes members of the Toll-like receptors ( TLR ) family that play a crucial role in recognition of pathogen-associated molecular signatures , including Tlr4 ( interaction with LPS from Gram-negative bacteria ) , Tlr9 ( unmethylated CpG containing DNA ) and Tlr13 ( vesicular stomatitis virus ) ( Figure 4A ) [38] . KEGG ( Kyoto encyclopedia of genes and genomes ) pathway enrichment analysis identifies several genes that play a key role in antigen processing and presentation in dendritic cells and macrophages , including Class I and Class II MHC ( major histocompatibility complex ) molecules , as well as proteases , membrane transporters and structural proteins involved in generation , transport and loading of antigenic peptides onto Class I or Class II molecules ( Figure 4B and Table S6 ) . Finally , we also note an enrichment of IRF8 binding peaks in the GO term nucleotide binding . Strikingly , many of the genes contained in that list include members of the IFN-inducible GTPase superfamily , including the Gbp ( guanylate binding proteins ) , Mx , and p47 ( Irg; immunity-related GTPases ) families which are involved in early innate immune response to intracellular infection in many cell types ( Figure 4B , 4C and Table S6 ) [39]–[41] . To identify direct functional targets of IRF8 in macrophages , we overlapped the list of IRF8 binding peaks ( ChIP-chip ) with the list of genes differentially regulated by exposure to IFNγ/CpG in macrophages from F2 mice bearing wt alleles at IRF8 . This intersection included 145 direct IRF8 targets controlled by 111 IRF8 binding sites ( Table S7 ) . We also examined the overlap between IRF8 binding sites detected by ChIP-chip and the genes which expression in macrophages in regulated by IRF8 in response to IFNγ/CpG ( from 2×2 Anova analysis , Figure 2A and Table S3 ) . We found 21 genes that are regulated in this fashion and that harbour an IRF8 binding site in their vicinity . These genes represent transcriptional targets of IRF8 which expression is regulated by IFNγ/CpG in macrophages in an IRF8-dependent fashion . The vast majority of these 21 genes were included in the intersection detected between IRF8 binding sites and genes regulated by IFNγ/CpG in wt F2 macrophages ( Table S7 ) . We also investigated the relevance of IRF8 targets discovered by ChIP-chip , to host defenses against infections in vivo . IRF8 and IFNγ are required for protection against pulmonary infection with M . tuberculosis [26] , and mice bearing mutations in either gene are hyper-susceptible to pulmonary tuberculosis [26] , [42] . IRF8 is required for development of the dendritic cell lineages , IL-12 production by these cells ( Th1 polarization of immune response ) , recruitment of T cells to the site of infection , macrophage activation and containment of infection by activated macrophages in granulomas [43] . To identify IRF8 targets that may play an important role in host defenses against pulmonary tuberculosis , we investigated which of the 319 IRF8 binding sites and associated genes are significantly regulated in the lungs of C57BL6/J ( B6 ) mice 30 days following aerosol infection with M . tuberculosis ( pairwise analysis of day 30 versus day 0 transcript profiles ) [29] . An intersection of 213 IRF8 binding sites corresponding to 359 associated transcription units was detected in this analysis . Therefore , ∼2/3 of the identified IRF8 targets were found to be modulated during M . tuberculosis infection in vivo . In addition , there was considerable overlap between the list of IRF8 targets which expression was regulated by a ) IFNγ/CpG stimulation in wt F2 macrophages and b ) following M . tuberculosis infection in the lungs in vivo ( Table S7 ) . Gene ontology and KEGG pathway analysis of these two lists once again identified “immune response” and “antigen processing and presentation” as the key functional annotation ( Table S6 ) . This overlap included a strong focus on genes playing a role in antigen presentation by Class I and Class II MHC molecules ( CD74 , H2-D1 , H2-DMa , H2-DMb1/2 , H2-Ea , H2-Eb1 , H2-Q8 , Ltb , Tapbp1 ) , cytokines , chemokines and their receptors ( Ccl6 , Cxcl9 , IL6ra , Csfr3 , Fcgrt , Tlr9 ) , anti-viral and anti-bacterial GTPases ( Gbp2 , 3 , 5 , 6 , Gma1 , Rgl2 ) and other early response genes ( Ifitm1 ) , as well as a numbers of proteolytic enzymes ( erap1 , lysosyme , endopeptidase ) ( Figure 4C , Figure 5 , and Figure 6 ) .
In this study , we have used transcript profiling and chromatin immunoprecipitation on microarrays ( ChIP-chip ) to investigate the role of IRF8 in macrophage function , activation by IFNγ/CpG , and response to M . tuberculosis infection in vivo . For transcript profiling experiments we compared RNA expression profiles from BMDMs obtained from mice that bear either a wt ( R294 ) or a severely hypomorphic IRF8 allele ( C294 ) derived from the mutant BXH2 mouse strain . In addition , biological and technical RNA replicates were from independent [BALB/c×BXH2] F2 mice of mixed genetic background but genotyped for the two IRF8 alleles . This was done to increase the stringency of the analysis , and to distinguish true IRF8-dependent effects on gene expression from irrelevant ones resulting from differences in genetic background of the wt ( C57BL/6J ) and mutant animals ( BXH2; mixed C57BL/6J , C3H/HeJ ) . This approach has been shown to be well suited to map genome-wide eQTLs that segregate as a result of presence or absence of a specific transcription factor [27]–[30] . These experiments produced several lists of genes which levels of expression , under different conditions , is influenced by IRF8 . The first list was obtained by comparing BMDMs from wt and IRF8C294 mutant mice , and corresponds to genes which basal level of expression in macrophages is influenced by IRF8 ( n = 454 ) . However , because IRF8 plays an important role in maturation of the myeloid lineage [11] , [16]–[18] , this list may also include genes not directly regulated by IRF8 , but rather modulated during macrophages maturation . The second list was obtained by comparing BMDMs from wt and IRF8C294 mutant mice treated with IFNγ/CpG ( IRF8 genes regulated during macrophage activation ) . Two sub-lists were generated , one obtained by pairwise comparison , and the other generated by a 2×2 interaction ( Anova ) analysis which takes into account IRF8-dependent differences in basal level of expression in absence of IFNγ/CpG stimulation ( n = 368 ) . Using ChIP-chip experiments with chromatin prepared from IFNγ/CpG activated macrophages and immunoprecipitated with anti-IRF8 antibodies , we identified a total of 319 IRF8 binding events ( minimum of 2 fold enrichment over control ChIP and p value≤0 . 001 ) on a promoter tiling array . From this information , we further extracted two overlaps and associated critical gene lists . The first one contains 145 genes and corresponds to IRF8 targets ( bound by IRF8 ) that are regulated by exposure to IFNγ/CpG in wt F2 macrophages in vitro . The second one contains 359 genes and corresponds to IRF8 targets ( bound by IRF8 ) that are regulated during pulmonary infection with M . tuberculosis in vivo . The above-mentioned lists were generated by comparing wt cells to those from BXH2 that bear the severely hypomorphic Irf8C294 allele; nevertheless , small amounts of residual activity may remain in Irf8C294 and gene lists obtained with this mutant may differ somewhat from those obtained by comparing wt cell to cells bearing a null Irf8−/− allele . These two gene lists are the most biologically relevant with respect to the role of IRF8 in macrophage function and defenses against M . tuberculosis in vivo . A striking feature of these lists is the preponderance of IRF8 targets associated with antigen recognition , processing and presentation by antigen presenting cells ( APCs ) . APCs include dendritic cells , macrophages and B lymphocytes . These cells capture either soluble or particulate antigen by scanning different areas of the body including epithelial surfaces , degrade this antigen and present to T-lymphocytes to activate immune responses . Although virtually all cells can present processed peptide antigens to T cells in association with Class I MHC molecules , so-called “professional APCs” present a wide range of antigens to T cells in association with Class II MHC molecules . Many of the genes coding for proteins involved in antigen recognition , antigen degradation , translocation to a suitable secretory compartment and Class I and Class II molecules are encoded by genes within the major histocompatibility locus ( MHC ) on mouse Chr . 17 and human Chr . 6 . Presentation of cytosolic peptides via association with Class I MHC molecules occurs in all cells , and is critical for protection against viral infection . It involves proteasome-mediated degradation of viral or other proteins into short peptides , which are then translocated into the endoplasmic reticulum ( ER ) via the tapasin-associated ( Tapap in the ER lumen ) ABC transporter heterodimer TAP1/TAP2 . Such peptides entering the ER are further trimmed by ER-specific aminopeptidases to fit on the Class I MHC binding site formed by the Class I α chain in association with β2 microglobulin . Peptide-bound Class I complexes are then released from tapasin-chaperone complexes to be delivered to the cell surface , where they can interact with cytolytic CD8+ T cells leading to destruction of the infected cells . In addition , Class I MHC antigen presentation is up-regulated by INFα , β , and γ as well as by LT and TNFα . As shown in Figure 5 and Figure 6 , we have determined that several genes involved in Class I MHC antigen presentation harbour validated binding sites for IRF8 , and/or are regulated in macrophages upon exposure to IFNγ/CpG and/or in the lungs of M . tuberculosis-infected mice in vivo . These include the PA28 subunit of the proteasome , the TAP1/TAP2 transport system and associated tapasin ( Tapap ) , ER aminopeptidase ( Erap1 ) , as well as Class I MHC a chain , β2 microglobulin and associated ER chaperone Calnexin . On the other hand , antigen presentation via the Class II MHC pathway is carried out by specialized APCs . It involves antigen capture via the endosomal or phagosomal routes through initial interaction with specific cell surface receptors of the TLR ( Toll-like receptor ) , C3R ( complement-3 receptor ) , FcR ( Fc receptor ) and Ig ( immunoglobulin ) families . These antigens are digested by members of the cathepsin family of Cys/Asp proteases in acidic endosomes , lysosomes and phagolysosomes . Class II MHC α and β chains stabilized by chaperones are associated with non-polymorphic Ii protein which prevents antigen binding at the α/β interface . Delivery of this complex to antigen-containing acidified endosomes/lysosomes causes proteolytic degradation of Ii ( Cd74 ) , leaving only the CLIP ( Class II-associated invariant chain peptide ) portion of Ii in the antigen binding site . This CLIP peptide is then removed by the HLA-DM ( major histocompatibility complex , class II , DM ) protein ( H2M in mice ) , freeing up the antigen binding site . Antigen-bound Class II complexes are delivered to the cell surface where they can interact with CD4+ helper T cells to induce production of effector T cells , activation of macrophages to microbicidal function , and antibody production depending on the type of APC involved . The process of Class II MHC antigen presentation can itself be stimulated by secretory products of APCs ( e . g IL12 ) or T cells ( IFNγ ) . As shown in Figure 4 , Figure 5 , and Figure 6 , our analysis shows that several genes of the Class II antigen presentation pathway harbour IRF8 binding sites and/or are regulated in macrophages upon exposure to IFNγ/CpG and/or during pulmonary tuberculosis . These include Tlr4/9/13 , members of the Cathepsin family of proteases , Class II MHC molecules , Ii , and HLA-DM . Together , these results establish a critical role for IRF8 in regulation of the key programmes of antigen presentation in APC . These results are in good agreement with the reported profound defect of IRF8 mutant mice in the DCs compartment , as they lack both CD11c+CD8α+ DCs and pDCs , and the small number of CD11c+CD8α+ and CD8α− DCs present in these mice remain immature and fail to up-regulate co-stimulatory molecules and produce key cytokines in response to microbial products in vitro and in vivo during pulmonary tuberculosis [18] , [26] . Macrophage maturation including expression of cytocidal function is also impaired in IRF8-deficient mice , and their macrophages are susceptible to infection with intracellular pathogens in vitro [14] , [15] . Many of the genes implicated in Class I and Class II MHC antigen presentation are located within the boundaries of the MHC locus . In this locus , we note a striking over-representation of binding sites for IRF8 and the number of genes which expression is regulated by IFNγ/CpG in macrophages . In the case of genes bound by IRF8 and regulated by IFNγ/CpG in macrophages ( 199 probes , corresponding to 145 genes genome wide ) , ∼10% of them map to the MHC region on Chr . 17 , with 11 binding sites mapping near 16 regulated genes . This concentration of direct IRF8 targets genes corresponds to a 14 fold enrichment over genome-wide representation of these 145 genes . Likewise , ∼10% of the genes regulated by M . tuberculosis infection in vivo and that contain a IRF8 binding site in their vicinity map to the MHC locus ( 9 fold enrichment ) , with 12 IRF8 binding sites mapping near 27 regulated genes ( Figure 5 ) . The regulatory role of IRF8 in the MHC locus also seems to include additional genes playing a central role in amplification of early immune response , such as TNFα , LT , and components of the complement pathway . IRF8-dependent transcription of MHC-linked genes may involve direct cis-acting effects of IRF8 binding to de novo motifs identified in our study , or may additionally involve amplification through activation of other transcription factors . For example , Ciita ( class II , major histocompatibility complex , transactivator ) is a non-DNA binding co-activator that binds to the so-called “MHCII enhanceosome” multiprotein complex and that serves as a master control factor for MHCII gene expression [44] . Ciita expression is up-regulated by IFNγ in a STAT1-dependent fashion; through the presence of GAS ( Gamma interferon activation site ) element in the proximal PIV responsive promoter in the Ciita gene [45] . Moreover , we observed that Ciita expression is tightly regulated by IFNγ/CpG in an IRF8-dependent fashion in macrophages from F2 mice ( 2×2 interaction Anova analysis; Table S3 ) , and is also regulated in pulmonary tuberculosis . The PIV promoter region of Ciita also contains a de novo IRF8 binding motif , and a cluster of weak binding sites were experimentally detected in this region by ChIP-chip ( data not shown ) . Therefore , it is possible that IRF8-mediated control of MHC gene expression involves amplification by other transcriptional regulators such as Ciita . Furthermore , our study points at an important role of IRF8 in transcriptional activation of several families of IFN-inducible intracellular GTPases of the p47 ( IRG ) , p65 ( GBPs ) and Dynamin ( Mx ) families which are known to be essential for protection against intracellular bacterial , parasitic and viral infections ( Figure 4C ) [40] . The p47 family of immunity-related GTPase ( IRGs ) contains 18–23 members in mice , with 6 having been characterized in some details [40] , [41] . They are expressed at low levels in different cell types , but mainly myeloid cells , and show dramatic up-regulation upon exposure to IFNγ . Studies in mutant mice have shown that deficiency in Lrg47/Irgm1 causes susceptibility to infection with M . tuberculosis [46] , [47] , while absence of Igtp/Irgm3 and Iigp1/Irga6 causes intracellular replication of Toxoplasma gondii [48] , [49] . In macrophages , the Irgm1 protein is rapidly recruited to the membrane of bacteria-containing phagosomes , where it is believed to facilitate delivery of lysosomal cargo for the destruction of intracellular pathogens , a process that is critically dependent on phosphatidylinositol 3 , 4 bisphosphate ( PI3 , 4P2 ) and PI3 , 4 , 5P3 [50] . As expected , we observed increased expression of several IRGs in macrophages in response to IFNγ and in vivo in M . tuberculosis infected lungs ( Figure 4C ) , but we also detected at least one IRF8 binding site near Irgm1 , with two weaker sites near Irg-47 and Irgm3 . Unfortunately , several of the IRGs promoter regions were not present on the arrays we used , and more experimentation will be required to determine if IRF8 binding sites are present at or near other IRG genes . The family of p65 GBP contains 11 members in mice that map to two gene clusters on chromosomes 3 ( Gbp1 , 2 , 3 , 5 , 6 , 13 ) and 5 ( Gbp4 , 8 , 9 , 10 , 11 , 12 ) [39] , [41] . Gbp mRNAs are induced by IFNγ in macrophages in vitro , and in spleen , liver and lungs of mice infected with intracellular pathogens Listeria monocytogenes and T . gondii [39] . Subcellular localization studies have shown that several Gbp family members are quickly recruited to the membrane of microbe-containing phagosomes formed in infected fibroblasts [39] . Like Irg proteins , Gbps also traffic to pathogen vacuoles to potentially deliver microbicidal products to restrict intracellular replication . We have detected by ChIP-chip 3 IRF8 binding sites on the chromosome 3 cluster that are associated with regulation of several of these Gbps ( Gbp2 , 3 , 5 , 6 ) in response to IFNγ or M . tuberculosis infection . We have validated at the protein level the IRF8-dependence of basal and IFNγ/CpG inducible expression of Cd74 ( and its cleavage product SLIP ) ( Figure 6B ) , and Gbp1 in macrophages ( Figure S2 ) , with more modest effects noted at the protein level for Gbp2 , Gbp3 and Irgm 1 ( Figure S2 ) . Although transcriptional activation of the Irg and Gbp genes by IFNγ was previously associated with the presence of GAS ( Gamma interferon activation site ) and ISRE elements in their promoter region , and activation via the Jak/Stat pathway and IRF1 [41] , our results strongly suggest that IRF8 may additionally be involved in this regulation . Moreover , the study of Irgm3−/− mutant mice has identified defects in antigen cross-presentation in these mice [51] . Interestingly , it has been proposed that IRG proteins ( and possibly Gbps ) may not only be involved in the delivery of lysosomal cargo to bacterial-containing phagosomes , but may also be involved in facilitating transport of antigen containing lipid droplets for antigen cross-presentation by Class I MHC molecules [51] . Although speculative , this proposal is in agreement with the observed role of IRF8 in directing transcriptional networks associated with antigen presentation by Class I and Class II MHC molecules . Finally , a recent study [31] used a combination of IRF8 ChIP-chip and expression profiling in IRF8 knocked down human myelomonocytic leukemia THP-1 cells to identify primary and secondary IRF8 targets in these cells . In agreement with the de novo IRF8 binding motif described herein ( Figure 3A and Figure S1A ) , previous gene specific studies [52]–[55] and our binding motif association study ( Figure 3C ) , these authors demonstrated a significant overlap between IRF8 and PU . 1 ChIP-chip binding locations . They identified development and differentiation genes affected by the loss of IRF8 , but also immune response genes as direct targets . The list of 84 IRF8 primary targets was compared to our results ( Table S8 ) . This list shows a 43% overlap with our list of IRF8 binding peaks , a 21% overlap with our list of genes regulated by IFNγ/CpG in a IRF8-dependent fashion , and 70% overlap with the list of genes differentially regulated during pulmonary tuberculosis in vivo . Therefore , the overlaps between the IRF8 target genes identified in both studies is fairly important . Together , these studies emphasize the predominant role of IRF8 in myeloid cell functions .
C57BL/6J ( B6 ) mice were purchased from the Jackson Laboratory ( Bar Harbor , ME ) . Recombinant inbred BXH2 mice were originally derived by B . Taylor at the Jackson Laboratory [56] and subsequently maintained as a breeding colony at McGill University . BXH2 males were used to generate [BALB/cJ×BXH2]F1 mice , which were then inter-crossed to produce an F2 progeny . F2 littermates homozygote for either the wt ( R294; IRF8+/+ ) or mutant ( C294; IRF8R294C/R294C ) IRF8 allele were identified and used in the transcript profiling experiments . IRF8 alleles were identified by genotyping for the proximal marker D8Mit13 using oligonucleotide primer pairs 5′-CCTCTCTCCAGCCCTGTAAG-3′ and 5′-AACGTTTGTGCTAAGTGGCC-3′ , which distinguishes between BALB/cJ and C57BL/6J , the strain background of the IRF8 genomic segment onto which the R294C mutation appeared in BXH2 [9] . The isolation of genomic DNA , and the genotyping for D8Mit13 alleles were carried out as described [9] . Male and female mice 8 to 12 weeks of age were used for all experiments , according to guidelines and regulations of the Canadian Council on Animal Care . The mouse macrophage cell line J774 was grown in Dulbecco's modified Eagle's medium ( DMEM , Sigma ) supplemented with 10% heat-inactivated fetal bovine serum ( HI-FBS , GIBCO ) , 100 U/ml penicillin , and 50 µg/ml streptomycin ( Invitrogen ) at 37°C , in 5% CO2-containing humidified air . BMDMs were isolated from femurs of 8- to 12-week-old mice and were cultured in DMEM ( Sigma ) containing 10% heat-inactivated fetal bovine serum ( HI-FBS ) , 20% L-cell-conditioned medium ( LCCM ) , 100 U/ml penicillin , and 100 µg/ml streptomycin in bacteriological grade dishes ( Fisher ) at 37°C in a humidified atmosphere containing 5% CO2 . Seven days later , cells were harvested by gentle washing of the monolayer with phosphate-buffered saline containing citrate . Cells were plated in 150-mm tissue culture-grade plastic plates ( 18×106 cells per plate; Corning ) in DMEM containing 10% HI-FBS , 10% LCCM , 100 U/ml penicillin , and 100 µg/ml streptomycin . In some experiments , macrophages were primed with IFNγ ( 50 U/ml ) for 18 hrs , prior to stimulation ( 3 hrs ) with recombinant mouse IFNγ ( Cell Sciences , Canton , MA ) , and CpG DNA oligonucleotides ( 5′-TCCATGACGTTCCTGACGTT-3′ ) used at a concentration of 400 U/ml and 1 , 5 µg/ml , respectively . IFNγ and CpG stimulate both the IFNγ receptor and Tlr9 , and engagement of both receptors stimulates IRF8 expression , via STAT1 , NFKB and possibly other pathways [11] . Total RNA was extracted from BMDMs obtained from 6 individual mice per experimental group , either prior to or 3 hrs following stimulation of BMDMs with IFNγ ( 400 U/ml ) and CpG DNA ( 1 , 5 µg/ml ) . IFNγ/CpG-stimulated macrophages were initially primed with IFNγ ( 50 U/ml ) for 18 hrs . Purified RNAs were analyzed for integrity by gel electrophoresis , and were then hybridized to microarrays ( Illumina Mouse WG-6 v2 . 0 Expression BeadChip ) according to the manufacturer's recommended experimental protocol . To minimize technical variability , RNA processing steps ( RNA extraction , probe labeling and microarray hybridization ) were executed in parallel for all samples . The GeneSifter™ microarray data analysis system ( Geospiza Inc . , Seattle , WA , USA ) was used to examine data generated from comparisons between control ( unstimulated ) and IFNγ/CpG-stimulated ( 3 hrs ) groups . Log transformed data were normalized and transcripts showing differential expression were identified by pairwise , or two-way ( 2×2 interaction ) Anova analysis with a t test p value of 0 . 05 and a fold-change cutoff of 1 . 5X . Hierarchical clustering based on complete linkage method was applied to evaluate the effect of the different sources of variability ( IRF8 alleles , treatments , host specific responses ) . Complete microarray data ( accession no . E-MEXP-2962 ) has been deposited in the ArrayExpress database ( www . ebi . ac . uk/microarray-as/ae/ ) . The expression of individual mRNAs was measured by quantitative PCR ( qPCR ) amplification of cDNA transcripts generated by reverse transcriptase ( RT ) . Briefly , RNA samples ( n = 6 ) used for transcriptional profiling were pooled and 3 µg of pooled RNA was converted to cDNA using Moloney murine leukemia virus ( Invitrogen ) in a 20 µl reaction according to the manufacturer's recommended experimental protocol . PCR amplification was performed using Quantitech SYBR Green PCR kit ( Qiagen ) , and all samples were measured in duplicate . Each reaction contained 2 µl of cDNA template , 1 µl of the target-specific primer pair ( each primer at 5 µM ) , 9 . 5 µl of RNase-free water and 12 . 5 µl of Quantitech SYBR Green PCR master mix . PCR amplification included an initial denaturation step ( 10 min at 95°C ) followed by 50 cycles of amplification ( 15 s at 95°C , 30 s at 57°C , and 33 s at 72°C ) , and was performed using the 7500 Real Time PCR system ( Applied Biosystems ) . PCR primers were designed to generate amplicons ranging from 100 to 150 bp . The Hprt gene was used as a constitutively expressed internal control to normalize the mRNA levels of target genes . BMDMs were obtained from wt ( B6 ) and IRF8 mutant ( BXH2 ) mice , either prior to or following stimulation ( 4 and 24 hrs ) with IFNγ ( 400 U/ml ) and CpG DNA ( 1 , 5 µg/ml ) . Whole cell extracts ( 75 µg per lane ) were subjected to 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis ( SDS–PAGE ) , followed by electroblotting and overnight incubation with the monoclonal anti-Cd74 ( Ii ) antibody ( clone In-1 purchased from BD Pharmingen ) ( used at 1∶200 ) . Immune complexes were revealed with a horseradish peroxidase-conjugated goat anti-rat antibody ( used at 1∶3000 ) and visualized by enhanced chemiluminescence ( SuperSignal West Pico kit , Thermo Scientific , Rockford , IL ) . Intact p41 and p31 Ii , as well as SLIP Ii fragment ( contains the NH2-terminal portion of Ii ) are all detected by the In-1 mAb [57] . Antibodies , dilutions and source dilutions for the immune GTPases were: Irgm1 ( A19 , 1∶200 ) , Irgm3 ( M14 , 1∶200 ) , Irga6 ( G20 , 1∶200 ) , Irgb6 ( A20 , 1∶200 ) , Gbp1 ( M18 , 1∶200 ) , Gbp2 ( M15 , 1∶1000 ) , Gbp5 ( L12 , 1∶500 ) were from Santa Cruz Biotechnology; Gbp3 ( Abcam , 1∶200 ) , Beta actin ( Sigma , 1∶1000 ) . ChIP assays were performed on J774 macrophages stimulated with IFNγ/CpG for 3 hours , according to a method previously described [58] , [59] . Stimulated cells were treated with formaldehyde ( 1% final; 10 min , 20°C ) , washed with ice-cold PBS , and cross-linked cells were harvested by centrifugation . The cell pellet was resuspended in 1 mL of cell lysis buffer ( 1% SDS , 10 mM EDTA , 50 mM Tris-HCl pH 8 ) supplemented with a cocktail of protease inhibitors , followed by sonication on ice . Chromatin was recovered by centrifugation ( 13 , 000 g , 7 min , 4°C ) , and resuspended in ChIP dilution buffer ( 0 . 5% Triton X-100 , 2 mM EDTA , 100 mM NaCl , 20 mM Tris-HCl pH 8 . 1 ) followed by pre-clearing using a 50% slurry of salmon sperm DNA/protein G agarose beads ( Upstate/Millipore ) for 2 . 5 hrs at 4°C . IRF8-DNA complexes were immunoprecipitated ( 4°C , 16 hrs ) using an anti-IRF8 antibody ( sc-6058x; Santa Cruz ) , followed by addition of 50% slurry of salmon sperm DNA/protein G beads ( 600 µL; 3 hr , 4°C ) on a rotating device . Control and anti-IRF8 immunoprecipitates were washed ( 10 min ) sequentially with each of the following buffers: low salt Buffer I ( 1% Triton X-100 , 0 . 1% SDS , 150 mM NaCl , 2 mM EDTA pH 8 . 0 , 20 mM Tris-HCl pH 8 . 1 ) , high salt Buffer II ( 1% Triton X-100 , 0 . 1% SDS , 500 mM NaCl , 2 mM EDTA pH 8 . 0 , 20 mM Tris-HCl pH 8 . 1 ) and Buffer III ( 1% IGEPAL , 0 . 25 mM LiCl , 1% Na-deoxycholate , 1 mM EDTA pH 8 . 0 , 10 mM Tris-HCl pH 8 . 1 ) , and a brief final wash in TE buffer ( 10 mM Tris-HCl pH 7 . 5 , 1 mM EDTA pH 8 ) . DNA was recovered from immunoprecipitated IRF8 chromatin complexes by incubation in a buffer containing 1% SDS and 0 . 1 M NaHCO3 ( 65°C , 16 hrs ) , and further purified using the QIAquick PCR purification kit ( Qiagen ) . Sample preparation for hybridization to promoter arrays was carried out as recommended in Agilent Mammalian ChIP-on-chip protocol , with minor modifications . Briefly , the ChIP DNA was amplified by ligation-mediated PCR ( LM-PCR ) following DNA blunting and linker ligation . The LM-PCR samples were purified on QIAquick purification columns and submitted to 18 additional rounds of amplification in the presence of aminoallyl-dUTP ( final concentration 300 µM; Sigma ) . The LM-PCR samples containing aminoallyl-dUTP were purified ( QIAquick PCR purification columns ) and labeled with Cy3 and Cy5 dies . The DNA amount was calculated by using the OD at 260 and 320 , and the Cy3 and Cy5 incorporation was also determined . Samples were hybridized to Agilent 244K mouse extended promoter arrays containing ∼17 , 000 of the best-defined mouse transcripts as defined by RefSeq spanning the regions from −5 . 5 kb upstream to +2 . 5 kb downstream of the transcription start site . The procedure was done according to the Agilent mammalian ChIP on chip protocol version 9 . 2 . Following the hybridization at 65°C for 40 hrs , the arrays were washed and scanned using a GenePix 4000B scanner and data was extracted from the images using Agilent Feature Extraction software as described in the mammalian ChIP on chip protocol ( Agilent , v . 10 ) . Data from ChIP-chips were normalized and averaged using ChIP Analytics 1 . 3 software . Data was processed in ChIP Analytics using the intra-array Lowess normalization , Whitehead Error Model v1 . 0 and Whitehead Per-Array Neighbourhood Model v1 . 0 for peak detection and evaluation . The default parameters were used to identify significant binding events ( 1000 bp as the maximum distance for 2 probes to be considered neighbors in a probe set , probe set p-value<0 . 001 for a “bound” probe ) . We retrieved from the UCSC genome browser 500 bp sequences centered on each 319 IRF8 ChIP-chip and performed de novo binding motif analyses with 3 different algorithms: MEME [32] , MDscan [33] and AlignACE [34] . The resulting matrices were compared to the Transfac v11 . 3 known binding motif database using the STAMP web-tool [35] . The schematic representations of the IRF8 de novo binding motif were generated with WebLogo [60] . The 319 IRF8 binding regions were queried for all known binding motifs on 1000 bp sequences using the optimized matrix threshold from MatInspector software ( Genomatix ) . Then , we searched for the same motifs on five sets of 319 randomly chosen 1000 bp sequences , selected from Agilent 244K mouse extended promoter array oligos . Thereafter , we calculated enrichment of binding motifs between the IRF8 binding regions and the mean of motif occurrence in random sequence sets ( Table S5 ) . We used the DAVID ( database for annotation , visualization and integrated discovery ) website to calculate GO and KEGG ( Kyoto encyclopedia of genes and genomes ) pathways enrichment in our different ChIP-chip and expression datasets ( DAVID threshold set to p value≤0 . 001 ) [36] , [37] . The Agilent 244K mouse extended promoter array or Illumina Mouse WG-6 v2 . 0 Expression BeadChip complete gene lists were used as reference respectively for enrichment evaluation . | IRF8 is a member of the Interferon Regulatory Factor family that is expressed in myeloid cells such as macrophages and dendritic cells and that activates or represses gene transcription upon stimulation with interferon gamma ( IFNγ ) , lipopolysaccharide ( LPS ) , and other microbial stimuli . IRF8 plays an important role in several aspects of myeloid cells , including differentiation and maturation of early progenitor cells , expression of intrinsic anti-microbial defenses , and production of the interleukin-12 ( IL12 ) cytokine , which is essential for priming of early T cell– mediated immune response . IRF8 mutant mice are susceptible to a number of intracellular infections including pulmonary tuberculosis . The transcriptional and cellular pathways regulated by IRF8 and essential for resistance to infections were studied by a combination of genome-wide methods , including transcriptional profiling and chromatin immunoprecipitation ( ChIP-chip ) . These studies identified phagosome maturation , antigen processing , and antigen presentation as critical pathways in early host–pathogen interactions regulated by IRF8 in macrophages exposed to IFNγ/CpG and in lung tissues infected with Mycobacterium tuberculosis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"genome",
"expression",
"analysis",
"animal",
"genetics",
"genetic",
"mutation",
"immune",
"cells",
"antigen-presenting",
"cells",
"immunity",
"to",
"infections",
"immunology",
"microbiology",
"host-pathogen",
"interaction",
"bacterial",
"diseases",
"genome",
"analysis",
"tools",
"immune",
"defense",
"bacterial",
"pathogens",
"infectious",
"diseases",
"mycobacterium",
"major",
"histocompatibility",
"complex",
"gene",
"expression",
"microbial",
"pathogens",
"biology",
"molecular",
"biology",
"immune",
"response",
"immune",
"system",
"genetics",
"of",
"the",
"immune",
"system",
"immunity",
"bone",
"marrow",
"genetics",
"genomics",
"molecular",
"cell",
"biology",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics"
]
| 2011 | Interferon Regulatory Factor 8 Regulates Pathways for Antigen Presentation in Myeloid Cells and during Tuberculosis |
Genome-wide association studies ( GWASs ) have discovered numerous single nucleotide polymorphisms ( SNPs ) associated with human complex disorders . However , functional characterization of the disease-associated SNPs remains a formidable challenge . Here we explored regulatory mechanism of a SNP on chromosome 9p21 associated with endometriosis by leveraging “allele-specific” functional genomic approaches . By re-sequencing 1 . 29 Mb of 9p21 region and scrutinizing DNase-seq data from the ENCODE project , we prioritized rs17761446 as a candidate functional variant that was in perfect linkage disequilibrium with the original GWAS SNP ( rs10965235 ) and located on DNase I hypersensitive site . Chromosome conformation capture followed by high-throughput sequencing revealed that the protective G allele of rs17761446 exerted stronger chromatin interaction with ANRIL promoter . We demonstrated that the protective allele exhibited preferential binding affinities to TCF7L2 and EP300 by bioinformatics and chromatin immunoprecipitation ( ChIP ) analyses . ChIP assays for histone H3 lysine 27 acetylation and RNA polymerase II reinforced the enhancer activity of the SNP site . The allele specific expression analysis for eutopic endometrial tissues and endometrial carcinoma cell lines showed that rs17761446 was a cis-regulatory variant where G allele was associated with increased ANRIL expression . Our work illuminates the allelic imbalances in a series of transcriptional regulation from factor binding to gene expression mediated by chromatin interaction underlie the molecular mechanism of 9p21 endometriosis risk locus . Functional genomics on common disease will unlock functional aspect of genotype-phenotype correlations in the post-GWAS stage .
With the advent of genome-wide association studies ( GWASs ) , a large number of single nucleotide polymorphisms ( SNPs ) associated with human complex diseases have been discovered [1] . The findings from GWASs have provided previously unsuspected biological pathways relevant to the diseases [2] such as complement system in age-related macular degeneration [3] and autophagy in Crohn’s disease [4] . Furthermore , the annotation of the disease-associated SNPs to plausible susceptibility genes and pathways can be translated into drug discovery and repositioning [5] . However , functional characterization of the identified SNPs remains a formidable challenge [6] . Most of the identified SNPs are located on intergenic regions and introns rather than coding regions [7] , suggesting that these non-coding SNPs are associated with the disease risk through regulation of expression levels of nearby genes . This argument has been supported by the fact that the disease-associated SNPs were overrepresented in expression quantitative trait loci ( eQTL ) of genes [8–10] . The ENCODE and Roadmap Epigenomics projects have explored functional elements across the human genome in a wide variety of cell types [11 , 12] . The integrative approaches for linking the regulatory elements with the SNPs identified by GWASs highlighted the importance of DNase I hypersensitive sites ( DHSs ) in the genotype-phenotype correlations [11 , 13] , which signify accessible open chromatin regions . A large proportion of the SNPs identified by GWASs or tagged by them were located on DHSs in cell types relevant to the corresponding diseases and overlapped transcription factor ( TF ) recognition sequences [14] . These systematic analyses present a plausible model in which the disease-associated SNPs alter activities of cis-regulatory elements including promoters , enhancers , insulators , and silencers . There is a growing body of evidence supporting this view from functional genomic studies in coronary artery disease ( CAD ) [15] , cholesterol levels [16] , obesity [17] , type 2 diabetes [18] , cancers [19–21] , erythroid traits [22] , and pigmentation traits [23–25] . It has been also reported that mutations on cis-regulatory elements cause severe developmental disorders showing a Mendelian pattern of inheritance [26] . Current GWAS platforms contain now up to 1M SNPs , which are designed to efficiently surrogate known common variants covering the human genome . Therefore , there is a possibility that the SNPs identified by GWASs are merely surrogate markers for causal variants . Although a comprehensive list of SNPs in the associated loci has been genotyped , the association signals for many correlated SNPs within a linkage disequilibrium ( LD ) block are statistically indistinguishable . Under these circumstances , it can be a straightforward way to prioritize candidate causal SNPs by searching for SNPs that are in strong LD with the SNPs identified by the GWASs and located on DHSs in cell types relevant to the disease for subsequent functional genomic approaches . Endometriosis is a common gynecological disorder that is characterized by the presence of uterine endometrial tissue outside the normal location . The prevalence in women of reproductive age is estimated to be 6–10% [27] . Endometriosis is a risk factor for several subtypes of ovarian cancer [28] . GWASs of endometriosis in Japanese and European populations have identified several susceptibility loci [29–33] . Uno and colleagues reported that rs10965235 , locating in an intron of ANRIL ( antisense non-coding RNA in the INK4 locus or CDKN2B-AS1 ) on chromosome 9p21 , was significantly associated with endometriosis in Japanese population [29] . rs10965235 has been reported to be most strongly associated with the risk of endometriosis in Japanese population ( per allele odds ratio of 1 . 44 ) but rare or absent in European descent populations [29] . This association was replicated in an independent GWAS in Japanese [30] . Meta-analysis of European and Japanese GWAS data sets identified rs1537377 at 49kb downstream of ANRIL [32] . rs1537377 is common both in European descent and Japanese populations and associated with modest increase of the risk for endometriosis ( per allele odds ratio of 1 . 15 ) [32] . These two SNPs on 9p21 were shown to be independent association signals [32] , but their functional roles have not been characterized . Here we investigated regulatory mechanism of the endometriosis risk locus on 9p21 . Coupled with target re-sequencing of 9p21 region and DNase-seq data from the ENCODE project , we prioritized candidate causal variants that were in perfect LD with the SNP identified by the original GWAS and located on DHSs . Subsequent functional genomic approaches revealed that the SNP site functioned as a cis-regulatory element of ANRIL through allele-specific long-range chromatin interaction driven by preferential bindings of TCF7L2 and EP300 . Furthermore , we demonstrated that expressions of ANRIL and CDKN2A/2B were closely associated via Wnt signaling pathway . These results suggest that the 9p21 risk locus is involved in the development of endometriosis by modulating the expression level of ANRIL and CDKN2A/2B .
In order to assemble a comprehensive set of genetic variants on 9p21 region , we re-sequenced 1 . 29 Mb interval encompassing two endometriosis-associated SNPs ( rs10965235 and rs1537377 ) in 48 Japanese individuals with the average depth of 186 . 6 ( S1–S3 Figs ) . We detected 4 , 215 single nucleotide variants ( SNVs ) and 664 insertions and deletions ( indels ) with a high degree of confidence ( S4–S6 Figs ) . We compiled a list of variants that were in strong LD with rs10965235 or rs1537377 ( Fig 1A and S1 Table ) . There were 24 SNPs and two indels exhibiting strong LD with rs10965235 ( r2 > 0 . 8 ) . Among them , 16 SNPs and two indels were in perfect LD ( r2 = 1 . 0 ) , and therefore the association cannot be distinguished . For rs1537377 , seven SNPs and one indel were detected . The intervals containing the variants exhibiting strong LD with rs10965235 and rs1537377 are located in the 3’ region of ANRIL ( Fig 1B ) . We confirmed that rs10965235 and rs1537377 were in weak LD each other ( r2 = 0 . 02 ) , and in weak and moderate LD with the 9p21 SNPs associated with other diseases ( S7 Fig ) , indicating that these two associations were independent endometriosis-specific signals . We explored DHSs to prioritize candidate causal variants . Among the ENCODE cell lines with DNase-seq data , we focused on endometrial carcinoma cell lines ( Ishikawa and ECC1 ) as a cell type potentially relevant to endometriosis because endometrial carcinoma arose from the endometrium . Additionally , we examined six cell lines , consisting of lymphoblastoid ( GM12878 ) , chronic myeloid leukemia ( K562 ) , H1 embryonic stem cell ( H1-hESC ) , hepatoblastoma ( HepG2 ) , cervical cancer ( HeLa-S3 ) , and umbilical vein epithelial cell ( HUVEC ) with high priority in the ENCODE project ( Tiers 1 and 2 ) . We detected a distinct set of DHSs across 101 kb interval containing all the variants showing strong LD with rs10965235 and rs1537377 ( chr9: 22072730–22173676 ) in these cell lines ( Figs 1B and S8 ) . The two SNPs identified by the original GWAS did not lie on DHSs in the endometrial carcinoma cell lines and other cell lines analyzed ( Fig 1B and 1C ) ; therefore , we excluded these two variants from candidate . We identified a site harboring two SNPs ( rs17761446 and rs17834457 ) where significant DNase-seq signals were consistently detected in the endometrial carcinoma cell lines ( Fig 1B and 1C ) . rs17761446 and rs17834457 are closely located ( 76 bp apart ) and in perfect LD with rs10965235 ( r2 = 1 . 0 ) . The DHS harboring rs17761446 and rs17834457 seems to be cell-type specific open chromatin site ( Ishikawa , ECC1 , and HeLa-S3 ) rather than ubiquitous one ( Fig 1C ) . We could not identify any SNPs that were in strong LD with rs1537377 and coincided on DHSs . Therefore , we focused on the DHS containing rs17761446 and rs17834457 for subsequent functional genomic studies . The DHS harboring rs17761446 and rs17834457 located on an intron of long isoforms of ANRIL ( Fig 1B ) , approximately 123 , 143 , and 109 kb apart from the transcription start site of ANRIL , CDKN2A , and CDKN2B , respectively . Therefore , we hypothesized that the DHS was a distal regulatory element contacting with the promoter of ANRIL , or CDKN2A/2B through chromatin looping interaction . We investigated long-range interactions using chromosome conformation capture ( 3C ) assay [34] . As a first step , we examined chromatin loops formed between the restriction fragment containing the candidate causal SNPs ( rs17761446 and rs17834457 ) and the consecutive restriction fragments around ANRIL , CDKN2A , and CDKN2B ( S2 Table ) . The consistent PCR amplifications of the interacting fragments between the candidate causal SNPs and the promoter of ANRIL were observed in replicated experiments from HEC251 and HEC265 ( S9A Fig ) . Additionally , Sanger sequencing for the PCR product confirmed that the sequence for the ligation junction between fragments between the candidate causal SNPs and the promoter of ANRIL was successfully identified , supporting the presence of the chromatin interaction ( S9B Fig ) . Therefore , we focused attention on the chromatin interaction between the SNP site and the promoter of ANRIL ( Fig 2A ) . We developed a novel 3C-based approach to detect allele-specific chromatin interactions ( AS3C-seq ) ( Methods and S10 and S11 Figs ) . Briefly , the AS3C-seq quantifies a difference in chromatin interaction frequencies between two alleles of a SNP in the sequence reads of the fragmented 3C PCR product by using cell lines that are heterozygous for the SNP . In this study , we used HEC251 and HEC265 that were heterozygous for the candidate SNPs ( rs17761446 and rs17834457 ) . The substantial numbers of reads ( >1 , 000 ) were aligned to the ligation junction , supporting the presence of the chromatin interaction between ANRIL promoter and the candidate causal SNPs ( Figs 2B and S11 and S12 ) . rs17834457 and rs17761446 are in perfect LD and therefore constitute two haplotypes; risk and protective haplotypes . The proportions of the read tags supporting the protective haplotype over the total mapped tag counts were 59 . 4% and 75 . 1% in HEC251 and HEC265 , respectively ( Fig 2B ) . These were significantly deviated from the expected proportion of 50% ( P < 10−16; binomial test ) . Additionally , HEC265 was heterozygote for a SNP ( rs3731197 ) within the fragment containing the promoter of ANRIL . The re-sequencing of the 1 . 29 Mb interval and haplotype phasing showed that the C allele of rs3731197 resided on the protective haplotype in HEC265 . The AS3C-seq result showed the proportion of the C allele of rs3731197 was similar to those at the two candidate causal SNPs ( 71 . 7%; P < 10−16 ) ( Fig 2B ) . The consistent results in both of the interacting fragments suggest the quantification of allele-specific read counts in the AS3C-seq is reliable . In the pooled data set from HEC251 and HEC265 , the proportion of allele-specific read counts supporting the protective haplotype was 66 . 5% ( P < 10−16; likelihood ratio test ) , indicating that the protective haplotype tagged by the G allele of rs17761446 forms two times greater chromatin interactions with the promoter of ANRIL compared to the risk haplotype ( Fig 2C ) . One reservation on allele specific functional genomic studies based on next generation sequencing technologies is mapping bias [35] . In order to evaluate mapping bias on our AS3C-seq analysis , we performed simulation and sensitivity analyses ( Methods and S10C Fig ) . According to the results of the simulation and sensitivity analyses , we confirmed that mapping bias does not affect our result ( Fig 2D and 2E; Methods ) . We examined whether rs17761446 and rs17834457 overlapped with consensus TF binding site within the DHS by bioinformatics analysis ( S13 and S14 Figs and Methods ) . As a result , we found that rs17761446 lay adjacent to the core motif of HMG ( high mobility group ) class of TFs including TCF/LEF and SOX families ( Figs 3A and S15 ) . According to the position weight matrices ( PWMs ) in the HOCOMOCO database [36] , rs17761446 lies within the binding motifs of TCF7 and TCF7L2 , which favors the G allele over the T allele ( Fig 3A ) . Additionally , the G allele matches with the consensus binding motif of drosophila TCF family member ( AAGATCAAAGG ) [37] . Chromatin immunoprecipitation ( ChIP ) -seq data from the ENCODE project has demonstrated the binding of TCF7L2 to rs17761446 site in HeLa-S3 , which harbors a DHS in this SNP site ( Figs 1C and 3B ) . Additionally , the bindings of EP300 ( a coactivator for LEF/TCF family of transcription factors ) and H3K27ac ( a histone modification characteristic of active enhancer ) in HeLa-S3 cells indicate that the SNP site facilitates a distal enhancer activity ( Fig 3B ) . Additionally , colocalization of TATA-binding protein ( TBP ) with the enhancer marks at rs17761446 seems to reinforce the presence of the enhancer-promoter interaction ( Fig 3B ) . On the other hand , we could not find consistent results between predicted binding motifs overlapping with rs17834457 and evidence of bindings from the ENCODE ChIP-seq analyses . Accordingly , we focused attention on rs17761446 , and then demonstrated the bindings of TCF7L2 , EP300 , H3K27ac , and RNA polymerase II to the SNP site in HEC251 cells by ChIP assays ( Fig 3C ) . We quantified allele-specific factor bindings at rs17761446 using TaqMan-based allelic discrimination assay . The allelic ratios in immunoprecipitated and input chromatins were determined by fitting their log2 transformed VIC/FAM ratios to the standard curve constructed from DNA mixtures with known allelic ratios ( S16 and S17 Figs ) . As a result , we detected significant allelic imbalances in the occupancies of TCF7L2 in two distinct antibodies ( Cell Signaling , P = 0 . 019; OriGene , P = 0 . 012; combined P = 5 . 7×10−5 ) and EP300 ( P = 7 . 0×10−3 ) ( S17 Fig ) . These factors selectively bound to the G allele over the T allele of rs17761446 , in which the occupancies of the G allele reached ~70% ( TCF7L2 Cell Signaling , 68 . 3%; OriGene , 66 . 0%; and EP300 , 65 . 7% ) ( S18 Fig ) . This indicates that the G allele exerts 2 . 06- and 1 . 95-fold greater binding affinities to TCF7L2 and EP300 , respectively ( Fig 3D ) . RNA polymerase II strongly bound to the G allele by a factor of 1 . 55 ( Fig 3D ) , although the allelic imbalance did not reach to significance level ( P = 0 . 14 ) . On the other hand , H3K27ac did not exhibit allele-specific bindings ( P = 0 . 50 ) . We evaluated whether rs17761446 worked as a cis-regulatory element of ANRIL , CDKN2A , and CDKN2B with the allele specific expression ( ASE ) analysis using eutopic endometrial tissues and endometrial carcinoma cell lines . We determined the haplotype phases with re-sequencing data for 9p21 region , and then screened the samples that were heterozygous both for rs17761446 and SNPs within transcribed regions of 9p21 genes with the same haplotypes ( Fig 4A and Methods ) . The allele-specific paired-end read tag counts were measured by the deep sequencing of the reverse transcription PCR products . Genomic DNAs from the corresponding samples were also deeply sequenced as control . The average of the paired-end read tag counts was greater than 20 , 000 . While the average proportion of the G allele of rs10965215 , which was located on an exon of ANRIL and resided on the protective haplotype tagged by the G allele of rs17761446 , was 49 . 4% over the genomic DNA samples , the paired-end read tags supporting the protective haplotype were over-represented in the RNA samples ( on average 64 . 0% ) . The difference between the genomic DNA and RNA samples was statistically significant ( P = 0 . 026 ) ( Fig 4B ) . This result indicates that rs17761446 was a cis-regulatory element of ANRIL and the protective allele was associated with 1 . 78-fold higher expression of ANRIL than the risk allele . On the other hand , the two alleles were not differentially transcribed in CDKN2A ( P = 0 . 20 ) and CDKN2B ( P = 0 . 61 ) ( Fig 4B ) . These results seem to be consistent with the result that the G allele of rs17761446 showed stronger chromatin interaction with the promoter of ANRIL rather than CDKN2A/2B . We confirmed that the result of our ASE analysis was robust against mapping bias by using simulation and sensitivity analyses ( Fig 4C and 4D; Methods ) . According to the results showing that the cis-regulatory element at rs17761446 regulated ANRIL expression and a key TF for canonical Wnt signaling , TCF7L2 [38] , differentially bound to the SNP alleles , we hypothesized that the induction of Wnt signaling changed expression level of ANRIL . Wnt signaling was induced by CHIR99021 ( CHIR ) in HEC251 cell ( S19 Fig ) . CHIR is a small molecule inhibitor of GSK3 , and therefore can activate Wnt pathway through decreased phosphorylation and subsequent stabilization of β-catenin [38 , 39] . We examined the effects of CHIR treatment in HEC251 cells on the expression of long and short isoforms of ANRIL ( S20A Fig ) . Compared to cells treated with vehicle only ( DMSO ) , the CHIR treatment significantly increased the expression levels of long and short isoforms of ANRIL by 33 . 2% ( P = 5 . 7×10−3 ) and 53 . 9% ( P = 1 . 6×10−3 ) , respectively ( S20 Fig ) . The effect was greater on the short isoform of ANRIL . We confirmed the increase in overall ANRIL transcripts using the primer pair amplifying both long and short isoforms ( 29 . 4% , P = 9 . 4×10−4 ) ( S20B Fig ) . ANRIL is thought to regulate the INK4/ARF tumor suppressor locus [40 , 41]; therefore , we examined expressions of two CDKN2A transcripts ( p16INK4A and p14ARF ) and one CDKN2B transcript ( p15INK4B ) . The expression levels of p16INK4A and p15INK4B significantly increased by 38 . 7% ( P = 5 . 5×10−6 ) and 35 . 7% ( P = 5 . 0×10−4 ) , respectively , with the CHIR treatment ( S20B Fig ) . On the other hand , p14ARF did not change ( P = 0 . 24 ) . This result implies that the induction of Wnt signaling activates ANRIL expression , which in turn leads to increased expressions of G1 cell-cycle inhibitors , p16INK4A and p15INK4B . In order to present a more direct evidence linking the candidate SNP and the change in ANRIL expression by inducing Wnt signaling via CHIR treatment , we evaluated whether degree of allelic imbalance of ANRIL expression increased after inducing Wnt signaling via CHIR treatment . For this purpose , we performed ASE analysis of ANRIL by using CHIR- and DMSO-treated endometrial carcinoma cell line ( HEC251 ) . Although we replicated ASE of ANRIL where protective haplotype was associated with increased expression ( CHIR , P = 2 . 8×10−4; DMSO , P = 7 . 8×10−4; one sample t-test ) , there was no difference in ASE between two treatments ( P = 0 . 83; Welch’s t-test ) ( S21 Fig ) .
The major challenge in the post-GWAS stage is to elucidate the functional aspect of genotype-phenotype correlations , i . e . , molecular mechanisms and biological pathways , affected by the identified genetic variants . The post-GWAS analysis is essential for the translation of genetic knowledge to clinical application . Functional studies focusing only on the SNPs identified by GWASs may not be promising because the identified SNPs are most likely surrogate markers for causal variants . The information about regulatory elements distributed across the risk locus provided by the ENCODE project can be a powerful tool to prioritize putative causal variants . Additionally , the regulatory mechanism modulated by candidate causal variant must be experimentally unraveled . In this study , we addressed these two steps to elucidate transcriptional regulation of the endometriosis-associated variants on chromosome 9p21 . We demonstrated that integrative analysis of the LD profile created by re-sequencing of the risk locus with information about regulatory elements from DNase-seq data provided by the ENCODE consortium was effective to prioritize putative causal variants . As a consequence , we found that the SNP identified by the original GWAS did not coincide with DHSs in the endometrial carcinoma cell lines and other representative cell types . On the other hand , we identified two SNPs ( rs17761446 and rs17834457 ) showing strong DNase-seq signals in the endometrial carcinoma and HeLa cells ( Fig 1 ) . We leveraged “allele-specific” functional genomic approaches to full advantage for the identification of allelic imbalances in distinct stages of transcriptional regulation including binding affinities of TF and its coactivator , chromatin interactions , and gene expressions . The advantage of allele-specific functional genomic approaches is that the use of tissues or cell lines that were heterozygote for a SNP of interest allows to directly compare functional difference between the two alleles of the SNP under the same cellular environment [42] . To the best of our knowledge , only a few studies addressed allelic differences in the frequencies of chromatin interactions between two alleles of causal SNPs [21–23] . In this study , we developed a method to detect allele-specific chromatin interactions ( AS3C-seq ) ( S10 Fig ) . The AS3C-seq utilizes the next-generation sequencing technology to quantify the difference in chromatin interaction frequencies between the two alleles of a SNP of interest . By using AS3C-seq , we detected an allelic imbalance in the frequencies of chromatin interactions between candidate causal SNP and the promoter of ANRIL . The AS3C-seq showed that a DHS in an intron of ANRIL containing two candidate causal SNPs formed chromatin looping with the promoter of ANRIL and the interaction frequencies differed according to the SNP alleles ( Fig 2 ) . One of these two SNPs ( rs17761446 ) overlapping with binding motifs of HMG class of TFs including TCF/LEF and SOX families was shown to alter the binding affinities to TCF7L2 and EP300 ( Fig 3 ) . The differential bindings of TCF7L2 and EP300 between the two alleles of rs17761446 seem to result in the observed allelic imbalance in chromatin interactions between the SNP site and the promoter of ANRIL , because the DNA bending properties of TCF/LEF lead to the formation of DNA loops favoring the concentration of distant transcriptional complexes [43] and their coactivator EP300 facilitates as a bridge between DNA-binding TFs and basal transcription machinery such as TFIIB and TBP [44] . The ASE analysis showed the allelic imbalance in the transcription of ANRIL between the SNP alleles by using eutopic endometrial tissues and endometrial carcinoma cell lines , suggesting that rs17761446 work as a cis-acting eQTL of ANRIL ( Fig 4 ) . The results of this study sheds light on regulatory mechanisms underlying 9p21 endometriosis risk locus , in which preferential bindings of TCF7L2 and its coactivator EP300 to the protective G allele of rs17761446 lead to stronger chromatin interaction with the promoter of ANRIL , which in turn activate transcription of the non-coding RNA . Although we initially focused on binding of TCF7L2 according to consistent results between the motif search and the ENCODE ChIP-seq analyses , SOX family members can be a good candidate for a regulator of the enhancer ability at the 9p21 endometriosis risk locus . Indeed , we found that SOX4 strongly bound to the candidate SNP site ( S22 Fig ) . Co-localization of SOX4 with TCF7L2 at the SNP site seems to be intriguing , because SOX4 has been reported to regulate Wnt signaling by direct binding to TCF family members and β-catenin [45] . According to the result of motif search , other SOX family members ( SOX5 , SOX9 , SOX13 , SOX15 , and SRY ) were also predicted to bind to the SNP site ( S13 and S15 Figs ) . It has been reported that some SOX factors work as agonist and promote Wnt signaling , others act as antagonist [46] . Therefore , further analyses for evaluating bindings of a set of SOX family members at the SNP site together with allele-specific binding analyses may be useful to uncover regulatory mechanism of the 9p21 endometriosis risk locus . Motivated by the fact that TCF7L2 is a key TF in canonical Wnt pathway [46] , we demonstrated that expression level of ANRIL was increased by inducing Wnt signal by CHIR treatment . At the same time , we found increased expressions of CDKN2A ( p16INK4A ) and CDKN2B ( p15INK4B ) ( S20B Fig ) . As shown in Fig 4 , rs17761446 did not exert cis-regulatory effects on CDKN2A and CDKN2B . Therefore , it is speculated that ANRIL modulates expressions of CDKN2A and CDKN2B through trans-regulation , though we could not rule out the possibility that other regulatory mechanisms could account for this observation . These findings imply that lower expression level of ANRIL in the risk allele is associated with reduced expressions of G1 cell-cycle inhibitors , p16INK4A and p15INK4B , which may contribute to a proliferative phenotype of ectopic endometrial cells and promote the development of endometriosis . To uncover modular regulatory mechanisms of the candidate SNP , Wnt signaling , and ANRIL expression , which leads to the development of endometriosis , further functional studies are required . A straightforward approach is to evaluate effect of disrupting the binding motif of HMG class of TFs at the SNP site on ANRIL and genome-wide expression profiles and cellular phenotypes , because it has been reported that ANRIL regulates gene expression network in a trans-acting manner [47 , 48] . Genetic variants on 9p21 have been reported to be associated with numerous diseases such as CAD [49] , intracranial aneurysm [50] , type 2 diabetes [51] , glaucoma [52] , and several types of cancers [e . g . , 53] . Harismendy et al . demonstrated that a 9p21 CAD associated SNP rs10757278 disrupted a binding site for STAT1 , and this enhancer site modulated ANRIL expression via IFN-γ stimulation [15] . Interestingly , the binding of STAT1 at the enhancer site exerted cell-type specific regulations of ANRIL ( repression in lymphoblastoid cells lines; activation in HUVEC ) [15] . ANRIL is transcribed into several different splicing variants . The risk haplotype for CAD have been shown to be consistently associated with decreased levels of isoforms containing the 5’ exons [54] , which is similar to our ASE result . The elucidation of tissue-specific regulatory mechanisms of ANRIL isoforms may clarify the pleiotropic effects of 9p21 region . Limitations of our study should be noted . The methods of prioritizing two SNPs ( rs17761446 and rs17834457 ) for subsequent functional studies were based on several criteria: degree of LD with the original GWAS SNPs; DNase-seq data from ENCODE cell types; and detection of significant DHSs . Although the inclusion criteria were set in order to elucidate more generalized functional aspect , we can not rule out a possibility that the variants that do not fulfill the criteria can be responsible for the 9p21 endometriosis risk locus . We prioritized rs17761446 over rs17834457 based on the motif search and the ENCODE ChIP-seq analyses . As both of these two analyses are incomplete , there is a possibility that uncharted TFs bind to rs17834457 . It is also plausible that both rs17761446 and rs17834457 facilitate as a cis-regulatory haplotype in a coordinated fashion . In order to evaluate whether the haplotype constructed by these variants or single variant is truly functional , it is useful to create mutant cell lines that harboring these variants separately or some combinations by using genome editing technique such as CRISPR/Cas9 system [55 , 56] . Among the SNPs identified by GWASs of endometriosis , most replicated SNP is located on chromosome 1p36 in a region close to WNT4 [31–33] . Together with our results , these findings raise the possibility that Wnt signaling pathway is involved in the pathogenesis of endometriosis . The regulatory mechanisms of these susceptibility loci might converge to shared molecular network underlying the pathogenesis of endometriosis , which will provide clues for potential targets exploited for new drug development .
Study subjects were Japanese and recruited at Niigata University and Nippon Medical School . All the participants provided written informed consent . The Ethics Committee of National Institute of Genetics ( nig1128 , 2012 . 10 . 9 ) , Niigata University ( 488 , 2014 . 1 . 27 ) , and Nippon Medical School ( 24-12-273 , 2014 . 3 . 15 ) approved the study protocols . We selected 1 . 29 Mb interval of chromosome 9 ( chr9:21299764–22590271 , hg19 ) as a target region for next-generation sequencing ( S1 Fig ) . The method for selecting target region is described in S1 Fig . We used NimbleGen SeqCap EZ choice system as a target enrichment method ( Roche Diagnostics ) . A DNA probe set complementary to the target region was designed by NimbleDesign ( https://design . nimblegen . com ) . Genomic DNA was prepared from 48 Japanese patients with endometriosis using QIAamp DNA Blood Maxi Kit ( QIAGEN ) according to the manufacturer’s protocol as previously described [30 , 57] . For each sample , 1 μg of genomic DNA was sheared into fragments with a mode length of about 200 bp on the Covaris ( Covaris ) . Sequencing libraries were constructed with the Illumina TruSeq DNA sample Preparation Kit with 12 different indexed adapters ( Illumina ) . Then , the target enrichment was performed with NimbleGen SeqCap EZ choice system according to the standard protocol ( Roche Diagnostics ) . The libraries were sequenced on four runs of the Illumina MiSeq platform with 2×150-bp paired-end module ( Illumina ) . The reads containing the Illumina adapter sequences were trimmed by using Trimmomatic version 0 . 32 [58] . After the quality control step for excluding or trimming low quality sequences , the sequence reads were aligned to human reference genome ( hg19 ) via BWA version 0 . 7 [59] . The aligned reads were processed for removal of PCR duplicates and erroneous reads by Picard tools version 1 . 111 , and for local realignment and base quality recalibration by GATK version 3 . 2 . 2 [60 , 61] . SNVs and indels were detected with the HaplotypeCaller and VariantRecalibrator of the GATK version 3 . 2 . 2 [60 , 61] . We analyzed DNase-seq data generated by Duke University in the ENCODE consortium to explore the distribution of DHSs . The sequence alignment files in the BAM format were downloaded from the ENCODE public download repository at http://genome . ucsc . edu/ENCODE/downloads . html . The peak call for the identification of regions in which aligned reads were significantly enriched was implemented via F-Seq version 1 . 84 with the use of background model for 20 bp sequences [62] . Endometrial carcinoma cell lines ( HEC251 [JCRB1141] and HEC265 [JCRB1142] ) were grown in DMEM ( Sigma-Aldrich ) with 10% FBS . These cell lines were purchased from JCRB Cell Bank . We confirmed that HEC251 and HEC265 were heterozygous for rs17761446 by Sanger sequencing of the PCR product with the primer pair listed in S3 Table . The 3C libraries from HEC251 and HEC265 were generated with the established protocol ( S10A Fig ) [34 , 63] . HEC251 and HEC265 cells ( 1×108 cells per experiment ) were crosslinked with 1% formaldehyde for 12 min at room temperature followed by adding glycine to a final concentration of 0 . 125M to stop further crosslinking . The crosslinked cells were lysed with the buffer ( 10 mM Tris-HCl , pH = 8 . 0 , 10 mM NaCl , 0 . 2% IGEPAL CA-630 Protease Inhibitor ( EDTA free ) ) using the Dounce homogenizer . The cell pellets were treated with corresponding buffers to restriction enzymes and 1% SDS , and then incubated at 65°C for 10 min to remove non-crosslinked proteins from the DNA . Placed on ice , 10% Triton X-100 was added to quench the SDS . For digestion , we used “6 bp-cutter” restriction enzymes ( EcoRI or HindIII ) followed by incubation at 37°C overnight with rotation . T4 DNA ligase was used to create intra-molecule ligations . The crosslinks were degraded by proteinase K treatment at 65°C overnight followed by an additional 2 h incubation at 42°C . The 3C DNA was purified by two rounds of phenol-chloroform ( pH = 7 . 9 ) method followed by five ethanol washes . The unidirectional primer design [63] was used for examining chromatin loops formed between the restriction fragment containing candidate causal SNPs and the consecutive restriction fragments around ANRIL , CDKN2A , and CDKN2B ( S10B Fig ) . The designed primers are shown in S2 Table . The workflow of the AS3C is illustrated in S10 Fig . We used HEC251 and HEC265 cells that were heterozygous for rs17761446 . By using HindIII , we could design the primers whose product contained the ligation junction and the SNP site at tractable size ( 1 , 225 bp; S11 Fig and S2 Table ) . The PCR product was gel-purified and then subjected to transposase-based library construction with the Nextera DNA Sample Preparation Kit ( Illumina ) , which allowed simultaneous DNA fragmentation and adaptor ligation . The DNA libraries were pooled according to their molar concentrations evaluated by the Agilent High Sensitivity DNA Kit and 2100 Bioanalyzer ( Agilent Technologies ) . The DNA libraries were sequenced on the MiSeq platform with 2×150-bp paired-end module ( Illumina ) . The bioinformatics analysis in AS3C-seq is described in S10C Fig . For the sequence reads aligned to “artificial chromosome” that was an exact sequence of the 3C PCR product , the number of paired-end read tag counts supporting reference and alternative alleles was measured . The allelic imbalance between two alleles of the SNP in the AS3C-seq reads was evaluated by assuming that allele-specific paired-end read tag counts follows a binomial distribution . For each experiment , the deviation from balanced allele-specific paired-end read tag counts ( 50:50 ) was examined by the binomial test ( two-sided ) . When results from multiple experiments were combined , the likelihood ratio test was examined as follows: LRT=2logL ( π1 ) L ( π0 ) =2∑i[niAlogπ1π0+ ( ni−niA ) log1−π11−π0] , where LRT statistic follows a chi-square distribution with degrees of freedom of 1 , π0 and π1 are the proportion of allele-specific reads supporting “A” allele for the SNP site under the null and alternative hypotheses , respectively . For ith experiment , ni is the total number of reads mapped to the SNP site and niA is the number of reads supporting “A” allele . Under the null hypothesis , π0 is set to be 0 . 5 . π1 is defined as a weighted proportion across experiments: π1=∑iniA/∑ini . 61-nt sequences of which candidate SNP sites were located in the middle were retrieved from hg19 . Two sequences with reference or alternative allele were generated for each candidate SNP . The PWMs were constructed based on the HOCOMOCO database [36] . MotifSuite was used to detect motifs within the query sequences showing better fits to the HOCOMOCO PWMs over the background model for human genome [64] . ChIP-seq data in HeLa-S3 cells were downloaded from the ENCODE public download repository at http://genome . ucsc . edu/ENCODE/downloads . Significant peaks for the bindings of TCF7L2 , EP300 , and TBP were determined based on the ENCODE narrow peak regions . The enrichment of a histone modification , H3K27ac , was defined as the ENCODE broad peak regions . ChIP assays for HEC251 cells were performed with SimpleChIP Plus Enzymatic Chromatin IP Kit ( Cell Signaling ) according to the manufacturer’s protocol . Briefly , cells were crosslinked with 1% formaldehyde for 10 min at room temperature , then treated with micrococcal nuclease to obtain fragments with a length of approximately 150–900 bp . The fragmented chromatin was subjected to immunoprecipitation with antibodies against TCF7L2 , EP300 RNA polymerase II , H3K27ac , and SOX4 . Normal rabbit IgG was used as negative control . The purified immunoprecipitated chromatin , input chromatin , and mock-IP ( IgG ) was subjected to PCR amplification of the candidate SNP site by using oligonucleotides primers ( S3 Table ) . The binding affinities were measured by real-time quantitative PCR with the KAPA SYBR FAST qPCR kit ( KAPA Biosystems ) on the 7900HT sequence detection system ( Applied Biosystems ) . The binding affinity in the immunoprecipitated chromatin was normalized to that in the corresponding input chromatin as follows: 2− ( CtIP−CtInput ) . The enrichment of the normalized binding affinity for the candidate SNP site or positive control region was normalized by dividing by that for negative control region . We used MYC promoter and α satellite repeat element as positive and negative control regions , respectively ( S3 Table ) . Details on the antibodies are shown in S4 Table . The allele-specific TF binding at rs17761446 was determined using the TaqMan-based allelic discrimination assay . We used TaqMan SNP genotyping assay for rs17761446 ( Applied Biosystems , Assay ID: C__33349228_10 ) . In order to measure allelic ratios for the immunoprecipitated and input chromatins from HEC251 cells , a standard curve of the VIC/FAM ratios for the samples with known genotypes was generated as follows . Calibration of the allelic discrimination assay was determined by mixing three pairs of DNA samples with GG and TT homozygous genotypes for rs17761446 at the following proportions: 50:50 , 60:40 , 70:30 , 80:20 , and 90:10 . DNA samples with GG , GT , and TT genotypes were also analyzed in the calibration assay . The standard curve was constructed by regressing the log2 transformed ratios for the VIC/FAM intensity on the log 2 transformed ratios of the two alleles . The allelic imbalance was examined by comparing the VIC/FAM ratios between the immunoprecipitated and input chromatins with the paired t-test ( two-sided ) . We performed the ASE analyses of rs17761446 for ANRIL , CDKN2A , and CDKN2B by using eutopic endometrial tissues and endometrial carcinoma cell lines . The eutopic endometrial tissues were provided by patients with ovarian cysts who underwent a total hysterectomy . All the specimens were snap-frozen in liquid nitrogen and stored at −80°C . Six patients with ovarian cysts and two endometrial carcinoma cell lines ( HEC251 and HEC265 ) that were heterozygous for rs17761446 were used for further analyses . We performed target re-sequencing of the 9p21 region with the same methods described in the previous section . The haplotype phasing for the genotype data of these 8 samples along with the 48 samples in the previous section was implemented via Beagle 4 . 0 [65] . We found that the five patients with ovarian cysts and two endometrial carcinoma cell lines harbored heterozygous genotypes both for rs17761446 and a SNP within ANRIL ( rs10965215 ) with the same haplotypes . In total , six and seven samples were available for the ASE analyses by using rs3814960 and rs3217992 as transcribed SNPs in CDKN2A and CDKN2B , respectively . LD structure for these SNPs are shown in S23 Fig . RNA was extracted from eutopic endometrial tissues and endometrial carcinoma cell lines using AllPrep DNA/RNA Mini kit ( QIAGEN ) according to the manufacturer's instructions . 1 μg of the total RNA was reverse transcribed by using ReverTra Ace-α kit ( Toyobo ) with random primer . The synthesized cDNA was subject to PCR amplification by using oligonucleotide primers for the three 9p21 genes ( S3 Table ) . The genomic DNA was also subjected to PCR amplification by using oligonucleotides primers ( S3 Table ) , and used as control . The obtained PCR products were converted into indexed libraries by using NEBNext Ultra DNA Library Prep Kit for Illumina ( New England Biolabs ) followed by sequencing on the MiSeq platform with 250-bp paired-end module ( Illumina ) . The generated reads were aligned to “artificial chromosomes” , which were exact sequences of the PCR products . The reads mapped with a high confidence ( MAQ > 30 ) were used for the ASE analyses . The allele-specific paired-end tag counts were measured by using only high confidence base calls ( base quality > 20 ) at the transcribed SNP positions . The allelic imbalance was examined by comparing the log 2 transformed ratios of the allele-specific paired-end tag counts between cDNA and genomic DNA with the paired t-test ( two-sided ) . CHIR ( Tocris Bioscience ) was dissolved in DMSO ( Sigma ) . HEC251 cells were treated for 24 h with either DMSO ( vehicle ) or CHIR ( 5 μM ) , and then harvested for gene expression analysis . RNA was extracted from DMSO- and CHIR-treated HEC251 cells using AllPrep DNA/RNA Mini kit ( QIAGEN ) according to the manufacturer's instructions . 1 μg of the total RNA was reverse transcribed by using ReverTra Ace-αkit ( Toyobo ) with random primer . The expression levels were measured by real-time quantitative PCR with the KAPA SYBR FAST qPCR kit ( KAPA Biosystems ) on the 7900HT sequence detection system ( Applied Biosystems ) . Three 9p21 genes ( ANRIL , CDKN2A , and CDKN2B ) were evaluated . Expression was normalized to ACTB . The oligonucleotides primers used are shown in S3 Table . The fold change was calculated by dividing expression level in CHIR-treated cells by that in cells treated with vehicle only ( DMSO ) . Six replicates were performed . Difference in gene expression level between the two treatments was examined by testing whether log2 transformed fold change was different from zero with one sample t-test ( two-sided ) . Immunofluorescence analysis was performed with CHIR- and DMSO-treated HEC251 cells . Cells grown on microscope slides were rinsed in Dulbecco’s PBS and then fixed in Bouin’s fixative ( Sigma-Aldrich ) for 5 min . The cells were incubated with the primary antibody against polyclonal β-catenin ( S4 Table ) over-night in a humidified chamber at 4°C . After washing , the cells were incubated with the secondary antibody ( Alexa-Flour 488 goat anti-rabbit IgG; Life Technologies ) in a humidified chamber for 1 h at room temperature . Finally , the cells were mounted with Vetashield HardSet Mounting Medium supplemented with DAPI ( Vector Laboratories ) . Images acquisition was performed using Olympus FluoView™ FV1200 ( Olympus ) . One reservation on allele specific functional genomics studies based on next generation sequencing technologies is mapping bias , in which reads with reference allele are more likely to be mapped than reads with alternative allele ( i . e . , bias toward preferential mapping to reference allele ) [35] . Therefore , we addressed mapping bias on our AS3C-seq and ASE analyses by performing simulation and sensitivity analyses described as below . ( i ) Simulation analysis We performed a simulation analysis for evaluating mapping bias by patterning after three studies [35 , 66 , 67] . For AS3C-seq , we generated single-end reads based on the human reference genome ( hg19 ) . For all possible segments overlapping rs17761446 or rs17834457 at a defined sequence length , we generated a set of reads containing reference and alternative alleles . We considered random sequencing errors in the simulation , in which each base in the reads was substituted to a different randomly selected base with a Bernoulli probability of 0 . 001 , 0 . 005 , and 0 . 01 . The error rate was determined based on Loman et al . reporting that MiSeq produced 0 . 1 mismatches per 100 bases [68] . We considered three simulation scenarios of different sequence lengths ( 50 , 75 , and 100 bp ) . We arbitrarily assigned the best base quality score of 41 to each base of the simulated reads . rs17761446 and rs17834457 are closely located ( 76 bp apart ) and in perfect LD ( r2 = 1 . 0 ) . We incorporated this fact into our simulation . When the sequence length was 100 bp , some of the sequence segments overlapped both of these SNPs . In such case , the haplotype structure observed in Japanese population was reflected in the simulation: reference T allele at rs17761446 resided on the same haplotype with reference C allele at rs17834457 . The simulated reads were mapped to the human reference genome rather than the artificial chromosome . The SNP sites were distant from a ligation junction between fragments containing the SNPs and ANRIL promoter ( >700 bp ) , indicating that any single reads overlapping the SNPs did not reach to the ligation junction . Therefore , mapping to the reference genome are thought to be sufficient in order to evaluate effects of mapping bias at the SNP sites . BWA with default setting was used . For the reads with mapping quality > 30 , we measured allele-specific read counts at the SNP sites and calculated proportion of the reads supporting reference allele . For each parameter combination , 10 , 000 simulations were carried out . For allele specific expression analyses , we conducted simulations by using the sequences of transcripts rather than genome . The results of the simulations for allele specific chromatin interaction and allele specific expression analyses are shown in Figs 2D and 4C , respectively . The results show that when the sequence length is shorter and the error rate is higher , degree of preferential mapping to reference allele increases . The averaged proportion of mapped reads supporting reference allele did not surpass 52% , even when the error rate is set to be 0 . 01 and the sequence length is 50 bp , in which the error rate is ten times as large as a literature value for Illumina platform [68] . When we consider more realistic scenarios ( i . e . , the error rate of 0 . 001 and 0 . 005 ) , the average proportion of mapped reads supporting reference allele was at most 50 . 03% and 50 . 06% , respectively . The results of the simulation showed very subtle preferences toward reference alleles , indicating that the mapping bias expected by the simulations does not affect our results . ( ii ) Sensitivity analysis In order to evaluate effects of mapping bias not covered by the simulation study , we performed a sensitivity analysis in which a base at SNP site in the reference sequence is changed to alternative allele . If a SNP site is susceptible to mapping bias , change from reference to alternative allele may diminish a preference to reference allele and exaggerate a preference to alternative allele . We aligned the reads generated by our allele specific chromatin interaction and allele specific expression analyses to the “original” and “changed” reference sequences with the same method . Then , differences in allelic imbalance were examined between two reference sequences . The results of the sensitivity analyses for allele specific chromatin interaction and allele specific expression analyses are shown in Figs 2E and 4D , respectively . For the SNP sites analyzed , differences in allelic imbalance were small between two reference sequences ( at most 0 . 15% ) . According to the two analyses , we confirmed that mapping bias does not affect our results . Additionally , both of the significant results from our AS3C-seq and ASE exhibited increased proportions of alternative alleles . Therefore , it is most likely that the observed allelic imbalances in our study were not caused by mapping bias . Sequence data for re-sequencing of 9p21 region has been deposited at the European Genome-phenome Archive ( EGA ) , which is hosted by the EBI and the CRG , under accession number EGAS00001001741/EGAD00001001942 . Nucleotide sequence data for AS3C-seq are available in the DDBJ Sequenced Read Archive under the accession numbers DRX051835 and DRX051836 . | A large number of variants associated with human complex diseases have been discovered by genome-wide association studies ( GWASs ) . These discoveries have been anticipated to be translated into the definitive understanding of disease pathogeneses; however , functional characterization of the disease-associated SNPs remains a formidable challenge . Here we explored regulatory mechanism of a variant on chromosome 9p21 associated with endometriosis , a common gynecological disorder . By scrutinizing linkage disequilibrium structure and DNase I hypersensitive sites across the risk locus , we prioritized rs17761446 as a candidate causal variant . The results of our “allele-specific” functional genomic approaches sheds light on regulatory mechanisms underlying 9p21 endometriosis risk locus , in which preferential bindings of TCF7L2 and its coactivator EP300 to the protective G allele of rs17761446 lead to stronger chromatin interaction with the promoter of ANRIL , which in turn activate transcription of the non-coding RNA . Motivated by the fact that TCF7L2 was a key transcription factor of Wnt signaling pathway , we postulated that the induction of Wnt signaling activated expression levels of ANRIL and cell cycle inhibitors , CDKN2A/2B . Functional genomics on common disease will unlock functional aspect of genotype-phenotype correlations in the post-GWAS stage . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"sequencing",
"techniques",
"genome-wide",
"association",
"studies",
"medicine",
"and",
"health",
"sciences",
"population",
"genetics",
"carcinomas",
"cancers",
"and",
"neoplasms",
"endometrial",
"carcinoma",
"alleles",
"oncology",
"genome",
"analysis",
"sequence",
"motif",
"analysis",
"epigenetics",
"molecular",
"biology",
"techniques",
"population",
"biology",
"chromatin",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"genomics",
"sequence",
"alignment",
"chromosome",
"biology",
"artificial",
"gene",
"amplification",
"and",
"extension",
"gene",
"expression",
"gynecological",
"tumors",
"molecular",
"biology",
"genetic",
"loci",
"haplotypes",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"evolutionary",
"biology",
"polymerase",
"chain",
"reaction",
"human",
"genetics"
]
| 2016 | Allelic Imbalance in Regulation of ANRIL through Chromatin Interaction at 9p21 Endometriosis Risk Locus |
In insects , precisely timed periodic pulses of the molting hormone ecdysone control major developmental transitions such as molts and metamorphosis . The synthesis and release of ecdysone , a steroid hormone , is itself controlled by PTTH ( prothoracicotopic hormone ) . PTTH transcript levels oscillate with an 8 h rhythm , but its significance regarding the timing of ecdysone pulses is unclear . PTTH acts on its target tissue , the prothoracic gland ( PG ) , by activating the Ras/Raf/ERK pathway through its receptor Torso , however direct targets of this pathway have yet to be identified . Here , we demonstrate that Drosophila Hormone Receptor 4 ( DHR4 ) , a nuclear receptor , is a key target of the PTTH pathway and establishes temporal boundaries by terminating ecdysone pulses . Specifically , we show that DHR4 oscillates between the nucleus and cytoplasm of PG cells , and that the protein is absent from PG nuclei at developmental times when low titer ecdysone pulses occur . This oscillatory behavior is blocked when PTTH or torso function is abolished , resulting in nuclear accumulation of DHR4 , while hyperactivating the PTTH pathway results in cytoplasmic retention of the protein . Increasing DHR4 levels in the PG can delay or arrest development . In contrast , reducing DHR4 function in the PG triggers accelerated development , which is caused by precocious ecdysone signaling due to a failure to repress ecdysone pulses . Finally , we show that DHR4 negatively regulates the expression of a hitherto uncharacterized cytochrome P450 gene , Cyp6t3 . Disruption of Cyp6t3 function causes low ecdysteroid titers and results in heterochronic phenotypes and molting defects , indicating a novel role in the ecdysone biosynthesis pathway . We propose a model whereby nuclear DHR4 controls the duration of ecdysone pulses by negatively regulating ecdysone biosynthesis through repression of Cyp6t3 , and that this repressive function is temporarily overturned via the PTTH pathway by removing DHR4 from the nuclear compartment .
The development of higher organisms is fundamentally dependent on the precise progression of specific gene programs , and even minor differences in the timing of these events can be fatal [1] , [2] . The regulation of simultaneous developmental programs generally relies on systemic signals—typically hormones—that coordinate these activities . In most , if not all multicellular organisms , steroid hormones function as precisely timed cues that control diverse gene programs to advance development or synchronize physiological changes . In humans , for example , the onset of puberty—while ultimately under neuroendocrine control—is governed by the action of steroid hormones that coordinate the developmental and behavioral changes associated with sexual maturation [3] . Typically , the release of steroid hormones from their respective glands is temporally controlled , resulting in systemic pulses of defined duration [4] . This raises the interesting question as to how onset , size , and duration of hormone pulses are regulated , since all these variables will affect target tissue responses . The insect molting hormone ecdysone represents an excellent model to address these questions , allowing us to study the dynamic effects of a steroid hormone in the context of a developing organism . In Drosophila , major and minor ecdysone pulses ( corresponding to high and low hormone titers , respectively ) occur throughout development [5] . The major pulses of ecdysone control the molts , the onset of metamorphosis , and the differentiation of adult tissues , while the minor pulses are linked to changes in physiology , such as the termination of feeding behavior . During larval development , ecdysone is mainly synthesized in the PG , and following its release into the hemolymph the hormone is converted to its biologically active form , 20-Hydroxyecdysone ( 20E ) , in peripheral tissues [6] . In Drosophila melanogaster , at least three minor ecdysone pulses occur in third instar ( L3 ) larvae . These pulses are required for important changes in physiology and behavior , including the commitment of a larva to a pupal fate ( critical weight checkpoint ) , the induction of the Sgs ( “glue” ) genes that serve to attach the pupa to a solid substrate , and the switch from feeding to wandering behavior [7]–[11] . Critical weight is a physiological checkpoint that determines whether the larva has acquired sufficient resources to survive metamorphosis and it commits the animal to undergo puparium formation . Once critical weight is attained , the larva will pupariate within normal time , regardless of whether nutrients are scarce or abundant [12]–[14] . The accurate timing of the major and minor ecdysone pulses suggests that the molecular mechanisms by which the hormone is synthesized , released , and degraded are tightly regulated . The past decade has provided considerable insight into the biosynthetic pathway that converts dietary cholesterol to 20E . Six genes linked to the Halloween mutations encode two different enzyme classes that act in the ecdysone/20E biosynthetic pathway , the cytochrome P450 monooxygenases ( disembodied , phantom , shadow , shade , and spook ) [15]–[20] , and a short-chain dehydrogenase/reductase ( shroud ) [21] . Two non-Halloween genes have been shown to also play a role in ecdysone synthesis , neverland , which encodes a Rieske electron oxygenase [22] , [23] and spookier , a paralog of spook [24] . The first and final biochemical steps of ecdysteroid biosynthesis are reasonably well understood , however relatively little is known about the enzymes that catalyze the intermediate steps , namely the conversion from 7-dehydrocholesterol to 5β-ketodiol . This succession of uncharacterized reactions is generally referred to as the “Black Box” and is believed to harbor the rate-limiting step for the production of ecdysone [25] . Recent work suggests that shroud and spookier ( spook in Manduca and Bombyx ) act in the “Black Box” [21] , [24] , however it is still unknown which enzymatic steps are catalyzed by these enzymes , and whether other enzymes also act in the Black Box . Some Halloween genes are transcriptionally upregulated when ecdysone levels are high; in Bombyx and Manduca , for example , the expression levels of phantom ( phm ) mRNA correlate well with the two ecdysone peaks preceding pupation [17] , [20] , [26] . In Drosophila , phm , shadow ( sad ) , and shade ( shd ) are induced roughly 12 h before puparium formation , concurrent with the major pulse of ecdysone that triggers this event . However , transcript and protein levels of these three Halloween genes appear to be relatively constant during the first 36 h of the L3 [27] , suggesting that the three low-titer 20E pulses are not simply a consequence of modulating gene expression of the Halloween genes . Therefore , it is unclear whether the generation of minor ecdysone pulses involves transcriptional control of any of the known or hitherto unidentified Halloween genes or whether regulation at the transcriptional level plays a role at all . In this study , we show that DHR4 appears to act as a transcriptional repressor of Cyp6t3 , a cytochrome P450 gene with a previously unknown role in ecdysteroid biosynthetic pathway . The neuropeptide PTTH is believed to control the timing of all major ecdysone peaks during larval and pupal development [28] , however it is unclear whether low-titer ecdysone pulses are also controlled in this manner . Seminal studies conducted in Bombyx resulted in the identification of PTTH as a brain-derived neuropeptide , which triggers the production of ecdysone in the PG [29] , however the gene encoding PTTH was only recently identified in Drosophila [30] . PTTH is sufficient to upregulate some ecdysteroidogenic genes in cultured Bombyx PGs [31] , [32] , and ablation of PTTH-producing neurons in Drosophila results in reduced expression of Halloween genes [30] . Surprisingly , ablating PTTH-producing neurons did not abrogate molting or metamorphosis , but instead caused substantial developmental delays with long larval feeding periods that resulted in large animals . This suggests that PTTH acts as a timer that coordinates key developmental transitions and final body size of the developing Drosophila larva , but that it is not essential for molting and metamorphosis per se . Interestingly , the same report demonstrated that PTTH mRNA displays an unusual cyclic pattern during the L3 , with transcript levels peaking every 8 h . Whether this unexpected transcriptional profile translates into corresponding changes of PTTH peptide levels is unknown , but it is plausible that these PTTH mRNA oscillations are causally linked to the minor ecdysone pulses that occur during the L3 in Drosophila . A recent report showed that Drosophila PTTH binds to the Receptor Tyrosine Kinase encoded by the torso gene [33] . Torso expression is highly specific to the PG and disruption of torso function via tissue-specific RNA interference ( RNAi ) phenocopies the PTTH ablation experiments . Importantly , the authors also demonstrate that PTTH stimulates ecdysone production through ERK/MAPK , Ras , and Raf , and that loss-of-function of these pathway components ( ERK , Ras85D , and dRaf ) via RNAi yield large , delayed animals comparable to those seen in the PTTH ablation or torso RNAi experiments . Conversely , when a constitutively active form of Ras ( RasV12 ) [34] was expressed specifically in the PG , larval development was accelerated , resulting in small , precocious pupae . We previously described a strong hypomorphic mutation in the DHR4 gene ( DHR41 ) that results in prepupal lethality and defects in developmental timing [35] . DHR41 mutants spend less time feeding than controls and exhibit precocious wandering behavior as well as premature pupariation , ultimately resulting in smaller body sizes . DHR4 encodes an orphan nuclear receptor most similar to the vertebrate receptor Germ Cell Nuclear Factor [36] , [37] . In our original report [35] , we found that DHR4 is highly abundant in the cytoplasm of PG cells in mid and late L3 larvae , with little or no detectable protein in the nucleus . We found no DHR4 expression in the two neighboring endocrine tissues of the ring gland ( RG ) , the corpus allatum , or the corpora cardiaca . We show here that nuclear import of DHR4 is developmentally regulated , and that the protein exhibits a precise nucleocytoplasmic oscillatory pattern in L3 larvae . We also provide evidence that this oscillatory behavior is controlled by PTTH signaling and that DHR4 counteracts the stimulating activity of this neuropeptide by repressing ecdysone pulses . Furthermore , we demonstrate that the accelerated switch from feeding to wandering behavior is consistent with a precocious rise in ecdysone concentrations due to the loss of this repressive function when DHR4 is mutated or knocked down . Based on RG-specific microarrays , we show that Cyp6t3 is a cytochrome P450 gene specifically expressed in the RG , and that the gene is normally repressed by DHR4 . Further , we demonstrate that Cyp6t3 plays a key role in ecdysone biosynthesis and that disruption of its function results in low ecdysone titers as well as developmental timing and molting phenotypes . These phenotypes can be rescued by providing an ecdysteroid precursor , 5β-ketodiol , as well as with feeding 20E or ecdysone . We propose a model by which DHR4 inhibits ecdysone synthesis through the regulation of cytochrome P450 genes , and whereby PTTH activity triggers the translocation of DHR4 from the nucleus to the cytoplasm to temporarily relieve this inhibition , thus allowing for ecdysone pulses to occur .
In a previous report , we demonstrated that DHR41 larvae engage in wandering behavior much earlier than controls , which in turn results in precocious pupariation [35] . We also showed that DHR41 mutants die as prepupae that are smaller than controls ( Figure 1A ) . Here , we describe an additional growth phenotype—the dwarf larva—that affects ∼5% of the mutant population ( Figure 1B ) . These remarkably small L3 larvae do not feed , are unable to pupariate , and eventually perish as almost transparent larvae due to depleted fat stores ( Figure 1B ) . This extreme growth defect is likely caused by a very early onset of wandering behavior , which is in line with the observation that we consistently observe a small percentage of wandering second instar ( L2 ) larvae in populations of DHR41 mutants ( unpublished data ) . In our previous report we also showed that DHR41 mutants commit earlier to puparium formation than controls . Specifically , when young L3 larvae are transferred to cycloheximide-containing media ( which inhibits growth by blocking protein synthesis ) , DHR41 mutants are able to pupariate under these conditions , while controls fail to do so , suggesting that the critical weight checkpoint occurs earlier in DHR41 mutants than controls . This in turn results in early wandering behavior and the ability to pupariate since feeding/cycloheximide uptake has stopped . These two observations raise the possibility that at least some ecdysone pulses occur precociously in DHR41 mutants , thus triggering early critical weight assessment , early onset of wandering behavior , and precocious pupariation . DHR4 is expressed in the PG throughout larval stages and in the fat body prior to molts ( Figure S1 ) , raising the question of which of the two tissues is critical for DHR4-dependent regulation of development timing . To examine whether it is the expression of DHR4 in the fat body or the PG that is linked to the defects in the timing of wandering behavior and puparium formation , we used the Gal4 system to induce tissue-specific RNAi against DHR4 in either of these tissues . We found that single copies of the Gal4 driver and responder transgenes were insufficient to elicit reliable phenotypes . However , when we used lines homozygous for both the driver ( RG: P0206-Gal4 , fat body: Cg-Gal4 ) and the responder ( UAS-DHR4-RNAi ) , we found that DHR4 RNAi in the ring gland triggers phenotypes consistent with early wandering , while fat body-specific RNAi resulted in prepupal lethality in ∼10% of the population ( compared to 0% in controls , unpublished data ) . In particular , P0206-Gal4 driven DHR4-RNAi in the ring gland resulted in early pupariation ( Figure 1C ) and small precocious prepupae ( Figure 1A ) , while fat body-specific interference of DHR4 expression yielded phenotypes consistent with a defect in the ecdysone hierarchy , namely prepupal lethality , the failure to evert anterior spiracles ( unpublished data ) , incomplete or absent head eversion , and incorrect location of the gas bubble ( Figure 1A ) . Importantly , we did not observe any timing defects in Cg>DHR4-RNAi animals , since we found normally sized pupae that pupariate with similar timing as controls ( Figure 1C ) . This suggests that DHR4 expression in the fat body is important for prepupal development , while it functions in the PG to control the timing of wandering behavior and puparium formation . The phenotypes observed with fat body and ring gland-specific DHR4 RNAi lines recapitulate the phenotypes we observed for the DHR41 mutation , suggesting that DHR4 function is most critical in these two tissues . Since critical weight is determined in early L3 larvae , we tested whether DHR4 function is required during this time to ensure appropriate timing of wandering behavior . Specifically , we used a heat-inducible RNAi line ( hsDHR4-RNAi ) to activate DHR4 RNAi either 4 h prior to or 4 h after the L2/L3 molt . To examine if either of these heat treatments affected the timing of wandering behavior , we determined the percentage of clear-gut larvae at different time points during the L3 stage . Gut clearing occurs in wandering L3 larvae , typically 30–36 h after the molt and is completed around 4–6 h prior to puparium formation . Although we induced DHR4 RNAi just 8 h apart , we only observe precocious wandering behavior when the heat treatment is applied in the late L2 , but not in the early L3 ( Figure 1D , E ) . In addition , we found a small percentage of dwarf larvae when RNAi was induced in late L2 larvae , while no dwarf larvae were found with the later heat shock ( unpublished data ) . These data suggest that DHR4 function around the L2/L3 molt is necessary for the correct timing of wandering behavior , which corresponds to the time window when critical weight is determined [38] . To test whether early wandering in P0206>DHR4-RNAi larvae ( Figure 1C ) is caused by precocious ecdysone signaling , we examined the expression levels of the glue gene Sgs-4 via quantitative real-time PCR ( qPCR ) . Sgs-4 represents one of the first identified ecdysone-inducible genes in Drosophila [39] , [40] . Roughly 24 h after the molt to the L3 ( ∼96 h after egg deposition ) , Sgs-4 is induced in salivary glands by a low titer ecdysone pulse [10] , [11] , after which the gene maintains high expression levels until it is abruptly turned off at pupariation . At 24 h after the molt , we observed feeding and wandering larvae in P0206>DHR4-RNAi populations , while controls show no sign of wandering behavior . Sgs-4 expression was drastically higher in the 16 h and 24 h RNAi populations relative to controls , and ∼2-fold higher when one compares the wandering with the feeding larvae in the RNAi group at 24 h ( Figure 2A ) . This indicates that the wandering cohort has received the 20E pulse that induces Sgs-4 earlier than the feeding cohort . Even at 16 h after the molt , when all P0206>DHR4-RNAi larvae are still feeding , we observe much higher expression levels of Sgs-4 compared to controls , suggesting that the corresponding 20E pulse has occurred already before this time point , thereby preceding the wild type 20E peak ( ∼20 h after the molt ) by at least 4 h . We also examined whether heat-induced DHR4-RNAi in late L2 would trigger precocious Sgs-4 induction , since this treatment triggers premature wandering behavior ( Figure 1D ) . In contrast to PG-specific DHR4-RNAi , we did not observe induction at 16 h after the L2/L3 molt , however at the 20 h mark we found that RNAi larvae had ∼3-fold higher Sgs-4 mRNA levels than controls ( Figure 2B ) . However , when we examined Sgs-4 levels at 24 and 28 h after the molt , we found higher expression of the gene in controls , suggesting that Sgs-4 is induced precociously , but submaximally in hsDHR4-RNAi animals . Finally , when heat shocked in early L3 , we do not observe differences in Sgs-4 expression between hsDHR4-RNAi and wild type larvae ( Figure S2 ) , consistent with our observation that only a heat treatment in late L2 triggers early wandering behavior . To complement these findings , we analyzed the expression profiles of two isoforms of the E74 gene , which are both ecdysone-regulated [41] , [42] . This approach has been used previously [43] and is based on the idea that the A and B isoforms of the E74 gene respond inversely to ecdysone concentrations . In particular , if the ecdysone titer is high , E74A is induced and E74B is repressed , but when ecdysone titers drop to intermediate concentrations , E74A is turned off , while E74B mRNA levels rise . Therefore , by measuring both isoforms using qPCR , we can infer whether ecdysone concentrations have fallen or risen . When hsDHR4-RNAi animals are heat shocked as late L2 , we find that E74A levels start to rise at 24 h after the molt and remain roughly 2-fold higher than controls at the 24 , 28 , 32 , and 36 h time points , while E74B levels show a corresponding drop at 28–36 h ( Figure S3 , left panels ) . These findings suggest a preceding rise in ecdysone concentrations , consistent with the precocious induction of Sgs-4 discussed above . When animals received a heat shock 4 h after the L2/L3 molt , we observed no significant differences in E74A and E74B transcript levels between hsDHR4-RNAi animals and controls ( Figure S3 , right panels ) , agreeing with our finding that a later induction of RNAi fails to trigger precocious wandering behavior ( Figure 1E ) . To test whether the precocious ecdysone signaling in the mid-L3 larvae could be linked to the inappropriate timing and/or duration of a preceding ecdysone pulse , we measured ecdysone titers during the first 24 h of the L3 using a 20E EIA immunoassay ( Cayman Chemical ) . The antibody appears to have similar affinities for 20E and its immediate precursor , ecdysone ( E ) ( Naoki Yamanaka , personal communication ) , and therefore titers likely reflect a combination of both ecdysteroids . We find that knocking down DHR4 in the RG overall results in significantly higher ecdysteroid levels at all time points we examined ( Figure 2C ) . Importantly , while we can resolve two minor ecdysteroid peaks in the control , we observe no recession of the first L3 pulse in P0206>DHR4-RNAi larvae ( Figure 2C , 16 h time point ) . Rather , the first and the second L3 pulse appear to be fused in RNAi animals , demonstrating that the first pulse was not properly repressed . It is likely that the combination of higher hormone levels and the inability to repress the first pulse causes the premature effects observed for Sgs-4 and E74 transcripts , as well as the acceleration of wandering behavior and pupariation . The derepression of ecdysone titers and the premature ecdysone signaling in DHR4-RNAi larvae suggest that the wild type function of this nuclear receptor is to inhibit ecdysone production and/or release . To test this idea , we reasoned that increasing DHR4 expression specifically in the PG using the phmN1- and P0206-Gal4 drivers should maintain low systemic ecdysone levels and prevent ecdysone pulses from occurring . Indeed , we found that DHR4 blocks molting when overexpressed in the PG , however the penetrance of this phenotype is dependent on the driver/responder combination being used , as well as the chromosomal location of the transgene , suggesting that this effect is dose-sensitive . For instance , when UAS-DHR4 ( inserted on 2nd or 3nd chromosome ) is used in combination with phmN1-Gal4 , animals cannot progress beyond the L1 stage ( Figure 2D ) . Similarly , using P0206>DHR4/3 ( inserted on the 3rd chromosome ) results in all larvae being trapped in the L2 stage ( Figure 2E ) , while P0206>DHR4/2 ( inserted on the 2nd chromosome ) allows 5%–10% escapers to reach the pupal stage . To test whether the DHR4-mediated block in molting could be rescued by ecdysone , we added 20E to the medium and scored the percentage of animals that progress to later stages . Regardless of the driver used , we observed significant rescue when 20E was added to the diet . In the presence of the hormone , ∼80% of the phmN1>DHR4/3 animals progress to the L2 stage , with another 20% reaching the L3 ( Figure 2D ) . Similarly , most of the P0206>DHR4/3 animals molt to the L3 in the presence of 20E , and ∼4% of the population pupariate ( Figure 2E ) . Next , we tested whether the ability of DHR4 to block molting was specific to the PG . Since the phmN1-Gal4 driver shows some expression in the fat body , we expressed DHR4 specifically in the fat body using the Cg-Gal4 driver . Similar to the results obtained with the PG drivers , we observe a developmental arrest in the L1 and L2 stages . In contrast to overexpressing DHR4 in the PG , however , the developmental arrest caused by fat body-driven expression of DHR4 cannot be rescued with 20E ( Figure 2F ) . Taken together , these results demonstrate that expressing DHR4 in the PG blocks molting and interferes with systemic 20E levels in a DHR4-dose-dependent manner , where a stronger Gal4 driver leads to animals trapped in earlier stages . We further characterized the ability of phmN1>DHR4 expression to block larval development . Closer examination revealed that these animals survive and continue to grow for up to 10 d as first instar ( L1 ) larvae . These L1 larvae grow very large , accumulate lipids in their fat bodies , and have larger organs than controls due to continued proliferation ( Figure 2G , H ) . This demonstrates that expression of DHR4 in the PG specifically blocks molting and does not trigger lethality . Rather , it appears that these animals simply lack the ecdysone pulse to molt to the next stage and that all other aspects of larval life function normally . We previously reported that DHR4 is highly enriched in the cytoplasm of PG cells , even though the protein is nuclear in fat body cells in late L3 larvae [35] . Therefore , it appears that the subcellular localization of DHR4 is differentially controlled in these two tissues , raising the question as to whether DHR4 can enter the PG nucleus at all and , if so , how this translocation is regulated . To test whether DHR4 could enter the nucleus of PG cells at certain times during larval development , we stained ring glands isolated from carefully staged L3 larvae ranging from 0 to 36 h after the molt with affinity-purified DHR4 antibodies . Using this approach , we found that the subcellular localization of DHR4 changes periodically in PG cells during the L3 . DHR4 appears to be entirely nuclear at 0 , 8 , 24 , and 36 h , completely cytoplasmic at 4 , 12 , and 20 h , and present in both compartments at 16 , 28 , and 32 h after the L2/L3 molt ( Figure 3A ) . During the first 36 h of the L3 , DHR4 completes at least three cycles: It shifts from the nucleus to the cytoplasm and back during the first 8 h after the molt , while the next two cycles last 16 and 12 h , respectively ( Figure 3B ) . These three oscillations show an intriguing correlation with the occurrence of the three low titer pulses during the L3 . In particular , based on direct measurements of 20E titers [8] and indirect assessments based on ecdysone-regulated gene profiling in L3 larvae [11] , the three minor 20E pulses have been mapped to 8 , 20 , and 28 h after the L2/L3 molt . It should be noted that 20E constitutes the final and active form of the molting hormone , and that a biosynthetic profile in the PG of its immediate precursor , ecdysone , would therefore have to precede the depicted 20E curve . Taking this into account , it appears that DHR4 is cytoplasmic during a minor pulse , but nuclear between these peaks , consistent with the idea that DHR4 regulates the timing of these peaks . In this context it is also important to note that PTTH mRNA was shown to cycle with an 8-h periodicity in staged L3 larvae [30] , raising the possibility that a causal link exists between the cyclic behaviors of PTTH expression and DHR4 localization . PTTH acts through Ras signaling , and larvae that express constitutively active Ras in their prothoracic glands ( phm22>RasV12 ) display shortened larval stages and small pupae [33] , which are strikingly similar to DHR4 loss-of-function phenotypes . We therefore investigated whether DHR4 acts in the PTTH pathway . To examine the impact of altered PTTH activity on the subcellular distribution of DHR4 , we analyzed the location of the DHR4 protein in PGs isolated from 0- to 8-h-old L3 larvae , which were carefully staged at the L2/L3 molt . We reasoned that this stage would not only allow us to follow DHR4 protein through an entire cycle , but also ensure that these animals are as precisely timed as possible . Manipulating components of the PTTH pathway affects developmental timing , which makes later time points difficult to compare between the different genotypes . To test the effects of genetically removing PTTH function , we ablated PTTH-producing neurons in ptth>grim transgenic animals . In PTTH-abolished animals , DHR4 accumulated in the nucleus , with some residual protein residing in the cytoplasm ( Figure 4 ) . Ring glands from later L3 time points look comparable ( unpublished data ) , strongly suggesting that nuclear export and/or degradation of DHR4 is abrogated when PTTH signaling is disrupted . To validate these findings , we examined the effects of torso RNAi , which targets the PTTH receptor . Very similar to the PTTH ablation line , we observed nuclear enrichment of DHR4 and loss of oscillation in phm22>torso-RNAi animals ( Figure 4 ) . We surmised that hyperactivating the PTTH pathway via constitutively active RasV12 [34] should result in cytoplasmic rather than nuclear accumulation of DHR4 . To test this , we carried out DHR4 antibody stains on ring glands isolated from phm22>RasV12 larvae . In contrast to the nuclear accumulation observed in larvae without intact PTTH signaling , we found strong cytoplasmic enrichment of DHR4 in PG cells when RasV12 was expressed ( Figure 4 ) , indicating that Ras activity dictates whether DHR4 can accumulate in the nucleus or not . Note that constitutively active RasV12 results in overproliferating PGs , explaining the large and malformed glands we observe . Taken together , our data demonstrate that the PTTH pathway controls DHR4 and that nuclear accumulation of the protein is only permitted when the pathway is inactive . This suggests that PTTH regulates DHR4 activity by controlling its subcellular localization , thereby permitting or preventing access of DHR4 to its target genes . To further examine the ability of RasV12 to prevent DHR4 from entering the nucleus , we tested whether RasV12 could abolish DHR4 nuclear localization in larval fat body cells , where DHR4 was shown to be restricted to the nucleus once translation has occurred , at least in L2 and L3 larvae ( Figures 5A , S1 ) [35] . When we specifically drive expression of UAS-RasV12 in the fat body using Cg-Gal4 , we found DHR4 to be virtually absent from the nuclei and to be strongly enriched in the cytoplasm instead ( Figure 5B ) , indicating that constitutively active Ras is sufficient to trigger cytoplasmic retention of DHR4 . These findings corroborate our results obtained from the PG , providing strong evidence that nucleocytoplasmic distribution of DHR4 is controlled by PTTH signaling . The dependence of DHR4 on Ras is intriguing because RasV12 is—to the best of our knowledge—the only other known genetic alteration besides the DHR41 mutation that results in accelerated larval development and small pupae [33] , [43] . We also observe the occurrence of dwarf larvae in RasV12 animals ( Figure 1B ) . These findings are consistent with our observation that RasV12 prevents DHR4 from accumulating in the nucleus , thus disrupting its nuclear functions , similar to DHR41 mutants or DHR4-RNAi animals . Based on this observation , one would predict that some of the effects of RasV12 could be blocked if one increases the level of DHR4 protein in the same tissue . We therefore asked whether DHR4 was epistatic to Ras or , more specifically , if we could rescue RasV12-induced phenotypes by overexpressing DHR4 specifically in the PG . First , we determined the average time to puparium formation when RasV12 or DHR4 , or both together , are expressed in the ring gland using the P0206-Gal4 driver . As mentioned previously , ∼5%-10% of P0206>DHR4/2 larvae reach the prepupal stage , allowing us to determine their timing profiles . As previously reported by others , P0206>RasV12 animals develop much faster than controls [33] , [43] , preceding the appearance of control prepupae by ∼20 h ( Figure 5C ) . In contrast , P0206>DHR4/2 prepupae form with a ∼20-h delay compared to controls . However , when RasV12 and DHR4 are expressed together in the PG , we observe normal timing of pupariation , and a partial rescue of the DHR4-mediated lethality ( Figure 5C ) . Since RasV12 overexpression results in a hyper-proliferation phenotype ( Figures 4 , 5B ) , we wondered whether this aspect of Ras activity could also be rescued by DHR4 overexpression . For this , we examined the morphology of ring glands isolated from staged L3 larvae that express RasV12 and/or DHR4 . P0206>RasV12 larvae have very large ring glands , while DHR4 expression using the same driver results in slightly smaller ring glands compared to controls ( Figure 5D ) . Importantly , when DHR4 is co-expressed with RasV12 , hyper-proliferation of the ring gland appears to be repressed , strongly suggesting that increasing levels of DHR4 in this tissue blocks Ras activity . PTTH acts on the PG by ultimately activating ERK , a MAP kinase , via phosphorylation . Upon activation , ERK can enter the nucleus and phosphorylate nuclear target proteins such as transcription factors or other kinases [44] . Since a key role of PTTH is the induction of ecdysone biosynthesis , and DHR4 appears to have the opposite function , we would predict that PTTH activity would remove nuclear DHR4 via activated ERK entering the nucleus . We therefore stained ring glands from staged early L3 larvae to examine the subcellular localization of ERK at various time points . At 0 and 8 h after the L2/L3 molt , we found ERK evenly distributed between nucleus and cytoplasm , however at the 4-h time point , ERK is strongly enriched in the nucleus ( Figure 5E ) . We conclude that ERK shows an inverse relationship to DHR4 , at least at the examined time points: ERK accumulates in the nucleus when DHR4 is enriched in the cytoplasm , and when DHR4 is abundant in the nucleus , we found no particular increase of nuclear ERK over cytoplasmic ERK . These data are consistent with the idea that ERK plays a role in the displacement of DHR4 from the PG nuclei in response to PTTH , in line with our findings that the subcellular localization of DHR4 is regulated by the PTTH/Ras/Raf/ERK signaling pathway . To identify possible target genes of DHR4 , we triggered RNAi in late L2 using hsDHR4-RNAi and carried out microarray analysis of ring gland RNA isolated from larvae staged at 4 and 8 h after the molt . To reduce the number of false positives , we analyzed two adjacent time points , 4 and 8 h after the L2/L3 molt , allowing us to select for genes that exhibit significant expression changes at both time points . Using a stringent filtering approach we identified 54 genes whose transcript levels showed a greater than 4-fold difference between controls and DHR4-RNAi animals at both time points ( Figure 6A ) . Selected genes from these 54 genes are shown in Figure 6B and C . Intriguingly , among these 54 genes are four cytochrome P450 genes , an enrichment that is highly unlikely to occur by chance ( p value = 2 . 4E-11 ) . Two of the P450 genes are downregulated ( Cyp6a17 and Cyp9c1 ) , while the other two show increased expression ( Cyp6t3 and Cyp6w1 ) , and the effects are very similar between the two time points ( Figure 6D , left panels ) . We also found a short-chain dehydrogenase/reductase ( CG2065 ) among the affected genes , which belongs to the same enzyme family as the Halloween gene shroud . Finally , CG16957 , which encodes a protein with a cytochrome b5 domain , is also affected by DHR4-RNAi . This protein family is functionally related to cytochrome P450 enzymes because both enzyme classes act as oxidoreductases and carry heme groups . To validate some of these observations , we analyzed the expression of all four affected cytochrome P450 genes in brain-ring gland complexes isolated from hsDHR4-RNAi animals as well as DHR41 mutants that were staged at 4 h after the L2/L3 molt . We also included the analysis of the Halloween genes sad , dib , and phm as additional controls , in case these genes were affected , but not identified by the microarray approach . As expected , all four cytochrome P450 genes identified by the array display very similar profiles in the qPCR validation experiments ( Figure 6D , right panels ) . When we analyzed the samples derived from the DHR41 mutants , we confirmed that Cyp6t3 and Cyp6w1 are significantly higher when DHR4 function is impaired . However , we found that Cyp9c1 displayed higher rather than lower levels compared to controls . In addition , we were unable to detect Cyp6a17 in DHR41 mutants or in the corresponding parental line , P427 , suggesting that both Cyp9c1 and Cyp6a17 expression varies substantially between different genetic backgrounds . Future experiments will address whether Cyp9c1 or Cyp6a17 are dependent on DHR4 function . Our qPCR analysis revealed no substantial effects on the tested Halloween genes ( Figure 6D , all panels ) , confirming the microarray results . Taken together , our microarray data identified two cytochrome P450 genes , Cyp6t3 and Cyp6w1 , which display significantly higher expression levels in the ring glands of hsDHR4-RNAi animals as well as in brain-ring gland complexes isolated from DHR41 mutants , indicating that DHR4 normally represses these genes . Of the two cytochrome P450 genes reproducibly affected by loss-of-DHR4 function , Cyp6t3 and Cyp6w1 , we chose to analyze Cyp6t3 in more detail for two reasons . First , Cyp6t3 , but not Cyp6w1 ( unpublished data ) , is specifically expressed in the ring gland based on qPCR analysis , which shows ∼20-fold enrichment of Cyp6t3 transcripts in the ring gland compared to whole body ( Figure S4 ) . We confirmed this by in situ hybridization , which demonstrates that Cyp6t3 is specifically expressed in the prothoracic glands and the corpus allatum ( Figure 7E ) . Bleed-through of the tyramide-amplified signal did not allow us to determine whether Cyp6t3 is also expressed in the corpora cardiaca . Second , when we examined the changes in gene expression at 0 , 4 , 8 , and 12 h after the L2/L3 molt , we found that Cyp6t3 levels , but not Cyp6w1 levels , oscillate during this time window , where lower concentrations of Cyp6t3 correlate with nuclear DHR4 , consistent with the idea that DHR4 represses this gene ( compare Figures 7G and S5 with Figure 3A ) . For these reasons we examined whether interfering with Cyp6t3 function in the PG via RNAi results in any developmental defects . Specifically , we expressed Cyp6t3 RNAi ( VDRC #109703 ) [45] in the PG by generating phm22>Cyp6t3-RNAi animals . Disrupting Cyp6t3 in this manner generates phenotypes typically observed in mutants that have defects in the ecdysone synthesis pathway [30] , [38] , [46] , [47] . For instance , we observe very large pupae ( similar in size to phantom ( phm ) and disembodied ( dib ) RNAi pupae , Figure 7A ) , double larval mouthhooks ( a common molting defect , Figure 7A inset ) , and L2 prepupae ( Figure 7A ) . The latter phenotype occurs when larvae forego the molt to an L3 and directly molt from an L2 to a prepupa . This phenotype is relatively rare and has been only associated with mutations in E75 [46] , dre4 [48] , and itpr [49] , all of which have dramatically lowered ecdysone levels . We also tested a second independently generated Cyp6t3 RNAi line ( VDRC #30896 , Figure S6 ) , which is based on a smaller dsRNA construct . In this RNAi line , we also observe very large pupae , consistent with a longer feeding period , but failed to identify any L2 prepupae . Feeding 20E to these animals completely rescues the large body size ( Figure S6 ) . Since the VDRC line #109703 was stronger than the #30896 line , we continued our studies with the former line . No phenotypes are observed when Cyp6t3-RNAi is expressed in the fat body ( unpublished data ) , indicating that the phenotypes induced by phm22>Cyp6t3-RNAi are specific to the PG . Next , we examined whether Cyp6t3-RNAi animals have lower ecdysone titers . To test this , we compared ecdysteroid concentrations at multiple time points between L3 control larvae and delayed L2 larvae of the same absolute age . The latter ultimately develop into L2 prepupae . As expected , Cyp6t3-RNAi larvae have severely reduced ecdysteroid titers compared to controls , but generate a small pulse before they form L2 prepupae ( Figure 7B , 100 h L2 time point ) . In addition , we also measured ecdysteroid concentrations in earlier stages and found that Cyp6t3-RNAi animals have lower hormone levels at all larval stages , but not as embryos ( Figure S7 ) . One would predict that feeding ecdysone to phm22>Cyp6t3-RNAi animals should rescue at least some of the phenotypes , and indeed we observed that the occurrence of L2 prepupae is completely rescued when 20E is added to standard medium ( Figure 7C ) , corroborating our finding that phm22>Cyp6t3 RNAi affects ecdysone production . To further characterize at which step in the ecdysone biosynthetic pathway Cyp6t3 might act , we examined which 20E precursors might also result in a rescue . For this , we took advantage of the fact that the Cyp6t3-RNAi phenotype was more pronounced on an instant medium ( 4–24 , Carolina Biological Supply Company ) , hereafter referred to a “C424 . ” This medium is naturally low in cholesterol and other sterols , and has been used by us for sterol rescue studies before [50] . On this medium , phm22>Cyp6t3-RNAi animals very rarely progress beyond the L2/L3 molt ( <0 . 5% ) , either dying as L2 larvae or L2 prepupae . When we supplemented C424 with carrier only ( ethanol ) , cholesterol , or 7-dehydrocholesterol ( 7dC ) , we failed to see any rescue , defined by larvae developing to L3 larvae or later stages . In contrast , adding E or 20E to C424 medium resulted in >60% rescue , while supplementation with 5β-ketodiol rescued ∼15% of the Cyp6t3-RNAi population past the L2/L3 molt ( Figure 7D ) , with some animals reaching the pupal stage ( unpublished data ) . The lower percentage of rescued animals with 5β-ketodiol likely reflects the fact that this compound has to enter the PG , while E and 20E can act directly on the target tissues . This strongly suggests that Cyp6t3 plays a role in the black box , since mutations affecting enzymes acting downstream cannot be rescued with 5β-ketodiol [21] , [24] . While our data demonstrate that Cyp6t3 is specifically expressed in the ring gland , it appears to be expressed at a fairly low level . Based on our microarray and qPCR experiments we estimate that in the ring gland , Cyp6t3 transcript levels are two orders of magnitude lower than those of phm , dib , and sad . A possible explanation for this is that Cyp6t3 forms part of a “bottleneck” for ecdysone production and that a low abundance of transcripts renders Cyp6t3 more susceptible to transcriptional regulation than its Halloween counterparts . We therefore wondered whether ( a ) overexpression of Cyp6t3 would be sufficient to alleviate this bottleneck and accelerate ecdysone synthesis and development and ( b ) if loss of Cyp6t3 function in a DHR41 mutant background is necessary for accelerated development of these animals . In the first experiment , we generated a phm22>Cyp6t3 line that expresses a Cyp6t3 cDNA at high levels in the PG [51] . This resulted in no obvious phenotypes with respect to timing or overall morphology ( unpublished data ) , suggesting that Cyp6t3 is not sufficient to accelerate developmental timing via an increased rate of ecdysone production . In the second experiment we found that DHR41; phm22>Cyp6t3-RNAi larvae display delayed instead of accelerated development relative to w1118 controls ( Figure 7F ) . This indicates that upregulation of Cyp6t3 in DHR41 mutants ( Figure 6D ) is necessary for the accelerated development in these animals , and that lowering the levels of Cyp6t3 via RNAi effectively abolishes this effect . We conclude that Cyp6t3 is necessary but not sufficient for accelerating development , suggesting that if Cyp6t3 is indeed part of a bottleneck , it is not acting alone in this rate-limiting step .
A series of reports have provided ample evidence that PTTH utilizes the Ras/Raf/ERK pathway to regulate ecdysone biosynthesis in Bombyx , Manduca , and Drosophila [33] , [52]–[56] . In the present study , we demonstrate that a critical readout of this pathway is the nuclear receptor DHR4 . We suggest a simple model where PTTH represses DHR4 activity via its removal from the nucleus , while DHR4 in turn represses the occurrence of low-titer ecdysone peaks when nuclear ( Figure 8 ) . In laboratory fly cultures , PTTH was shown to be a non-essential gene , however when PTTH-producing neurons are ablated , development is substantially delayed and animals have a concomitant increase in body size [33] . We show here that disrupting DHR4 specifically in the PG results in opposite phenotypes to loss-of-PTTH function , where animals are smaller and develop faster than controls . Like PTTH ablation lines , animals homozygous for P0206>DHR4-RNAi can be kept as a viable stock , consistent with the idea that DHR4 functions in the PG as a PTTH-dependent , non-essential developmental clock to generate appropriately timed ecdysone pulses . It is of interest to note that the expression of constitutively active Ras in the PG of P0206>RasV12 animals results in accelerated larval development and small pupae , very similar to DHR41 mutants and P0206>DHR4-RNAi animals . P0206>RasV12 animals are also viable , and we showed that L3 larvae of this genotype accumulate DHR4 in the cytoplasm of PG cells ( Figure 4 ) . This strongly suggests that P0206>RasV12 larvae display these phenotypes precisely because DHR4 protein is prevented from entering PG nuclei , thereby mimicking the loss-of-function phenotypes observed in DHR4 RNAi or mutant larvae . We have demonstrated that the PTTH pathway controls the subcellular location of DHR4 . Loss of PTTH signaling results in nuclear presence of DHR4 , while constitutively activating this pathway leads to cytoplasmic localization of the protein . It is unclear at this point whether the DHR4 oscillations represent shuttling or involve cycles of degradation and synthesis . Shuttling would require a stable DHR4 protein that moves in and out of the nucleus , however this would be difficult to reconcile with the fact that DHR4 RNAi works well in our hands , since a continuously shuttling protein would be impervious to RNAi . It is evident that sufficient turnover of the DHR4 protein must occur , at least around the L2/L3 molt , when we conducted our heat-induced DHR4-RNAi experiments ( Figures 1D–E , 2B , and S3 ) . This raises the question of whether DHR4 mRNA levels are oscillating , given that the degraded protein must be replaced periodically . When we conducted a time course microarray of wild type ring glands , we observed very low and constant levels of DHR4 mRNA , which does not support the idea that DHR4 transcripts levels are oscillating ( Ou et al . , manuscript in preparation ) . Based on these data , we suspect that DHR4 mRNA is highly stable in PG cells and translated when needed in L3 larvae . Alternatively , our current approach might be too insensitive to detect periodic changes in DHR4 mRNA levels . It was shown in mammalian cell cultures that the ERK pathway controls the subcellular localization of the nuclear receptor PPARγ . Specifically , mitogenic stimulation of resting cells causes the binding of nuclear MEK1 to PPARγ , which is followed by rapid export of the protein complex from the nucleus [57] , [58] . Our study does not provide direct evidence that DHR4 is phosphorylated . However , our data are suggestive of the idea that ERK plays a role in removing DHR4 from the nucleus , since we found that ERK changes its nucleo-cytoplasmic distribution in early L3 larvae , in an apparent inverse relationship to DHR4 , consistent with the notion that it acts upstream of this nuclear receptor ( Figure 5E ) . Future experiments will have to examine whether DHR4 is a direct target of ERK and whether phosphorylation plays a role in the nucleo-cytoplasmic oscillations of DHR4 . An intriguing finding is that Drosophila PTTH mRNA levels oscillate with an apparent 8-h cycle time throughout the 48-h duration of the L3 stage [30] . DHR4 , on the other hand , displays an 8-h , 16-h , and 12-h cycle time for the first 36 h of the L3 stage , raising the question as to how these ultradian periods are established . What could account for the difference in these cycle times ? A simple possibility is that we were unable to detect all DHR4 cycles and that the DHR4 oscillations are well aligned with the PTTH cycles . In this study , we chose to conduct a time course based on a 4-h step size , because we consider 4 h a robust time interval that should compensate for the inherent asynchrony that exists in developing Drosophila larvae . We examined 15–20 ring glands per time point , and only found some discrepancies among ring glands from later time points , likely due to the asynchronous development of the population . However , it appears that during some time points , such as 16 h after the L2/L3 , DHR4 is detected in the nucleus and the cytoplasm , and it is possible that this reflects a transition phase of a cycle we might have missed . Future studies could attempt a time course with a 2-h step size , ideally in combination with a Sgs3-GFP reporter line to re-stage animals in the mid third instar [8] . Alternatively , the differences in cycle duration between DHR4 and PTTH could reflect the possibility of another cyclic process that may contribute to the timing of DHR4 periodicity . An attractive possibility is that circadian rhythms are superimposed on the PTTH oscillations to determine DHR4 cycle times . Anatomical evidence indicates that the central circadian pacemaker cells found in the Drosophila brain , the Lateral Neurons , indirectly innervate the PG [59] . A critical effector of these neurons is the neuropeptide PDF ( Pigment Dispersing Factor ) , and a mutation in pdf alters the periodicity of PTTH mRNA oscillations [30] . Future experiments will address whether PDF and other components of the circadian clock impinge on the oscillatory behavior of DHR4 in PG cells . The mechanism by which PTTH regulates ecdysone biosynthesis has been the subject of intense research for the last 35 years . PTTH triggers a complex array of signaling events in the PG that precede the synthesis of ecdysone . These include second messengers like Ca2+ influx and the synthesis of cAMP , followed by the stimulation of Protein kinase A , the activation of p70S6K and the concomitant phosphorylation of ribosomal protein S6 , a myriad of tyrosine phosphorylations , as well as the activation of the ERK pathway discussed in this report [56] . Clearly , PTTH triggers a range of events , which , among others , results in the transcriptional upregulation of genes required for ecdysteroid biosynthesis . The first evidence that PTTH is sufficient to increase transcript levels of an ecdysteroidogenic gene is based on the observation that the Bombyx disembodied gene ( dib-Bm ) is upregulated by administering PTTH to cultured prothoracic glands [32] . The upregulation of phm and spo appears to be more moderate [31] , [32] , while sad transcription appears not to be induced under these conditions [32] . Attempts to generate functional recombinant Drosophila PTTH have so far been unsuccessful [33] , but loss-of-function studies have confirmed a role for PTTH in the transcriptional regulation of ecdysteroidogenic genes . Specifically , ablation of PTTH-producing neurons resulted in a strong reduction of Drosophila dib ( ∼10-fold down ) and had a moderate effect on phm , sad , and spok ( 2–3-fold down ) . While similar studies with a torso RNAi line have not been published , indirect results come from a recent paper where the authors knocked down dSmad2 function in the PG , which results in a strong downregulation of torso [60] . Concomitantly , spok and dib are strongly reduced in phm>dSmad2-RNAi larvae , but similar to the above findings , no effect was seen for phm and sad . Taken together , the findings from Bombyx and Drosophila seem to suggest that the transcriptional effect of PTTH is most clearly established for dib , while phm and sad appear to be less dependent on PTTH signaling . One aspect of PTTH-mediated transcriptional regulation that has not been satisfactorily addressed is whether some of the ecdysteroidogenic genes require PTTH to reach high expression levels in the first place or whether the hormone provides a “boost” to elevate transcript levels even further . According to our RG microarray and qPCR data presented here , we conclude that expression levels of the Halloween genes are very high in L3 larvae , comparable to that of ribosomal genes . We also conducted a microarray time course ( Ou et al . , manuscript in preparation ) , which suggests that the Halloween genes are expressed at very high levels during the first 36 h of the L3 , without much fluctuation in their expression levels . Therefore , it would seem unlikely , at least according to our data , that transcriptional downregulation of the highly expressed ecdysteroidogenic genes like dib or phm can be the mechanism by which the three minor ecdysone peaks are generated , which is in line with our finding that knocking down DHR4 in the RG has no effect on phm , sad , and dib ( Figure 6D ) . Rather , it appears that DHR4 negatively regulates Cyp6t3 and possibly other uncharacterized genes that play critical roles in the production of ecdysone ( Figure 8 ) . DHR4 appears to mainly act as a repressor [35] , similar to what has been reported for its vertebrate ortholog GCNF [61] . It remains to be seen whether the genes that are downregulated in our hsDHR4-RNAi microarrays are indirect targets of DHR4 or whether this nuclear receptor can act as an activator on some promoters . We have previously shown that DHR4 acts downstream of the 20E receptor EcR as a key component of the ecdysone hierarchy during puparium formation [35] . It therefore appears that DHR4 acts upstream of ecdysone in the PG , but downstream of the hormone in target tissues ( Figure 8 ) , which nicely reflects the duality of DHR41 mutant phenotypes: the mutation affects developmental timing ( due to its role in the PG ) and puparium formation ( due to its role in the fat body ) . While some transcription factors have been identified for their roles in the ecdysone production pathway , none have been reported to repress ecdysteroidogenesis , and none have been directly linked to PTTH . A mutation in the woc ( without children ) gene , which encodes a putative zinc finger transcription factor , causes low ecdysone levels . This phenotype can be rescued with feeding 7-dehydrocholesterol ( 7dC ) , suggesting that Woc might control the conversion of cholesterol to 7dC by regulating the genes required for this step [62] . The nuclear receptor βFTZ-F1 is expressed in the PG of late L3 larvae and is required for normal levels of Phm and Dib protein [27] . Another nuclear receptor , E75A , acts in a feed-forward loop to maintain normal ecdysone levels , possibly by acting upstream of βFTZ-F1 in the PG [46] . Our ring gland microarray and qPCR data revealed that Cyp6t3 transcript levels are significantly elevated in hsDHR4 RNAi animals and DHR41 mutants . In addition , we showed that Cyp6t3 expression levels oscillate during the first 12 h after the L2/L3 molt , with lower levels of Cyp6t3 when DHR4 is nuclear ( Figure 7G ) . It therefore appears plausible that Cyp6t3 is a direct transcriptional target of DHR4 , however direct evidence for this is lacking . No DNA recognition sites have been identified for DHR4 , nor is it known whether this nuclear receptor acts as a homodimer , heterodimer , or monomer . According to our microarray data , as well as judging by the cycle numbers required to detect Cyp6t3 in qPCR experiments , we estimate that transcript levels of this gene are relatively low , probably by two orders of magnitude lower than the Halloween mRNAs for phm , dib , spookier , and shadow in PG cells ( unpublished data ) . This is consistent with the finding that Cyp6t3 was previously neither detected by in situ hybridization in any larval tissue , nor amplifiable from larval cDNA , while the aforementioned 4 Halloween genes showed strong expression under the same conditions in the PG of Drosophila larvae [63] . However , using tyramide amplification coupled to in situ hybridization , we were able to validate the ring gland-specific expression of Cyp6t3 ( Figure 7E ) . According to our wild type microarray study , Cyp6t3 is one out of nine cytochrome P450 transcripts that have a higher than 10-fold enrichment in the ring gland compared to the whole body signal ( Ou et al . , manuscript in preparation ) , supporting the idea that this gene has an important role in ecdysteroid biosynthesis . The fact that Cyp6t3 is expressed at very low levels raises the possibility that the Cyp6t3 enzyme is scarce and therefore rate-limiting with respect to the production of ecdysone . An attractive model is that DHR4-mediated repression of Cyp6t3 suffices to reduce ecdysone production to basal levels . If true , one would predict that Cyp6t3 turnover is controlled so that transcriptional control of the Cyp6t3 gene becomes a relevant factor in controlling ecdysone synthesis . Conversely , derepression of Cyp6t3 due to loss-of-DHR4 function could result in faster accumulation of ecdysone , which would account for the timing defects we observe . However , when we overexpressed a Cyp6t3 cDNA specifically in the PG , we did not observe any obvious phenotypes or effects on development timing , suggesting that changing levels of Cyp6t3 alone is not sufficient for this response ( unpublished data ) . In contrast , we could show that Cyp6t3 function is necessary for the accelerated developmental phenotype of DHR41 mutants , strongly supporting the notion that Cyp6t3 is a key target of DHR4-mediated repression of ecdysone pulses ( Figure 7F ) . The degree by which Cyp6t3 is repressed might directly correspond to the amount of DHR4 protein allowed to enter the nucleus . In support of this idea we find that overexpression of DHR4 cDNA in the PG results in varying degrees of larval arrest , depending on the strength of the Gal4 driver being used ( Figure 2D–F ) , which in turn suggests that the strength of the phenotype depends on how much DHR4 can enter the nucleus . Therefore , it is conceivable that the nuclear functions of DHR4 are dose-sensitive , giving rise to the idea that the oscillations of this nuclear receptor do not necessarily represent an all or nothing response , but may in fact fine-tune the expression levels of target genes instead . An interesting question is whether Cyp6t3 has clearly identifiable orthologs in other insects species , which is the case for the Halloween genes [64] . To address this question we conducted a series of BLAST searches with the Drosophila melanogaster Cyp6t3 amino acid sequence as a query . We limited our search to insect species with sequenced genomes and aligned the top hits from this search ( Figure S8 ) and analyzed their phylogenetic relationship using the program Seaview ( Figure S9 ) [65] . Due to the large size of the Cyp450 family 6 [66] and the sequence similarities within this family , we were not able to identify a definitive Cyp6t3 ortholog outside the Drosophila genus . We reached this conclusion based on a reverse BLAST search strategy , which revealed that the top hits from the D . melanogaster query did not retrieve Cyp6t3 as the best hit when used as queries themselves . Despite this , Cyp6t3 is highly similar to sequences found in other insect species , and it is likely that very similar paralogs are masking the true “functional” ortholog . Often , reverse BLAST searches fail to reveal definitive orthologs . For instance , the best characterized nuclear receptor in Drosophila , EcR , is most similar to both FXR and LXR , depending on whether one uses the DNA-binding domain or the ligand-binding domain as a query [67] . Future studies in other insect species will have to address the question of whether Cyp6t3 orthologs are also important for ecdysone production .
w1118 ( #3605 ) and UAS-RasV12 ( #4847 ) were ordered from the Bloomington stock center . Gal4 drivers were obtained from labs indicated by the references . Ring gland: P0206-Gal4 , UAS-mCD8-GFP [68] , phm22-Gal4 [33]; phmN1-Gal4 [69] . Fat body: Cg-Gal4 [70] . PTTH-Gal4 driver and PTTH ablation line: UAS-Grim/CyO-act-GFP & ptth-Gal4/Ser-act-GFP [30]; DHR41/FM7h , & hsDHR4-RNAi[35]; RNAi lines were obtained from Vienna Drosophila RNAi Center [45] . UAS-Torso-RNAi ( VDRC #4300 & #1016 ) ; UAS-Cyp6t3-RNAi ( VDRC #109703 & #30896 ) ; UAS-phm-RNAi: ( VDRC #100811 ) , UAS-dib-RNAi ( VDRC #101117 ) . Before embryos were collected on grape juice agar plates , flies were allowed to lay eggs twice for 2 h in order to reduce egg retention . After 2-h egg collection intervals at 25°C , eggs were transferred to petri dishes containing fresh yeast paste and reared at 25°C . Pupariation was scored in 2-h intervals . For the hsRNAi experiments , larvae were reared on yeast until the late L2 stage , at which w1118 controls and hsDHR4-RNAi L2 larvae were heat shocked for 35 min at 37 . 5°C . After a 4-h recovery , newly molted L3 larvae were transferred to yeast paste supplemented with 0 . 05% bromophenol blue to monitor their gut clearing status [71] , [72] . All qPCR data shown here are based on 3–4 biological samples each tested in triplicate . Whole larvae were collected in distilled water and snap-frozen in liquid nitrogen , while dissected tissue samples were prepared in ice-cold PBS , rinsed twice with fresh PBS , transferred to TRIzol ( Invitrogen ) , and snap-frozen in liquid nitrogen . Total RNA of whole larvae was isolated following a modified TRIzol protocol , where we substituted sodium acetate with lithium chloride for RNA precipitation . Total RNA from tissue samples was extracted using the RNAqueous-Micro Kit ( Ambion ) following the manufacturer's instruction . RNA samples ( 0 . 5–2 µg/reaction ) were reverse transcribed using ABI High Capacity cDNA Synthesis kit , and the synthesized cDNA was used for qPCR ( StepOnePlus , Applied Biosystems ) using PowerSYBR Green PCR master mix ( Applied Biosystems ) with 5 ng of cDNA template with a primer concentration of 200 nM . Samples were normalized to rp49 based on the ΔΔCT method . All primer sequences can be found in Table S1 . To generate pUAST-DHR4 cDNA , a 6 . 3 kb fragment containing the full-length synthetic cDNA of DHR4 [35] was cut with Eco RI and Xba I from Litmus 28 and cloned into pUAST digested with the same enzymes . For the pUAST-DHR4-RNAi construct , the same inverted repeat used in the hsDHR4-RNAi [35] was used to clone the fragment into pUAST using Xba I for all restriction cuts . Transgenic flies were generated by injecting DNA at a concentration of 0 . 5 µg/µl along with 0 . 1 µg/µl helper plasmid pΔ2–3 into embryos following standard procedures [73] , [74] . Tissues were dissected from larvae in PBS , fixed in 4% paraformaldehyde ( EMS #15710 ) in PBST ( PBS containing 0 . 3% Triton-X 100 ) for 20 min at room temperature ( RT ) , and washed in PBST . Tissues were then blocked for 2 h at RT or overnight at 4°C in PBST/5% NGS . Primary antibodies were incubated at 4°C overnight , while the secondary antibody was either incubated overnight at 4°C or 4 h at RT . Nuclei were stained with DAPI ( 1∶5000 ) . After several wash steps , tissues were mounted in SlowFade Gold Antifade Reagent ( Invitrogen ) . Images were captured on a Nikon C1 plus confocal microscope . Anti-DHR4 antibody was used at a dilution of 1∶500 , and anti-ERK antibody was used at a dilution of 1∶100 ( Cell Signaling #4695 ) . Secondary antibodies ( anti-rabbit Cy3 ) were used at a dilution of 1∶200 ( Rockland #611-104-122 ) . DHR4-RNAi and w1118 populations were heat shocked as late L2 larvae for 35 min at 37 . 5°C . To carefully stage larvae at the L2/L3 molt , L3 larvae were discarded 4 h after the heat treatment , and L3 larvae that molted in the following hour were allowed to feed for either 4 h or 8 h before their ring glands were dissected in ice-cold PBS . Ten ring glands were dissected and washed twice in PBS before being transferred to ice-cold TRIzol reagent ( Invitrogen ) . The lysates were then vortexed for 5 s at RT , flash frozen , and stored at −80°C . Total RNA was isolated by Ambion RNAqueous-Micro Kit . Isolated RNA was quantified by RiboGreen Quanti Kit ( Invitrogen ) and RNA integrity was analyzed by Agilent Bioanalyzer Pico Chips . RNA linear amplification was based on the MessageII RNA Amplification kit ( Ambion ) : First-strand cDNA synthesis was carried out by a T7- ( dT ) primer and ArrayScript reverse transcriptase using 50 ng RNA of each ring gland sample . Second-strand cDNA synthesis was performed according to the provided protocol . Purified cDNA was then fed into the IVT reactions . The amplified RNA ( aRNA ) was column-purified and analyzed by Agilent Bioanalyzer Nano chips . 1 µg of aRNA was used for double-stranded cDNA synthesis ( Invitrogen SuperScript One-Cycle cDNA Kit ) and 1 µg of the purified cDNA was Cy3-labeled by Roche NimbleGen one-color cDNA labeling kit . From this , 4 µg of Cy3-labeled cDNA was hybridized on a Drosophila 12×135K Array ( NimbleGen ) . Each condition was analyzed by three independent biological samples . Chip hybridization and scanning was performed by the Alberta Transplant Applied Genomics Center . Raw data were normalized with the NimbleScan software ( NimbleGen ) using the RMA algorithm [75] , and data were analyzed with Arraystar 4 . 0 ( DNAstar ) as well as Access ( Microsoft ) . Larvae were collected in 1 . 5 ml tubes and stored at −80°C . Samples were then homogenized in methanol and centrifuged at maximum speed , after which the precipitates were re-extracted with ethanol . The extracts were pooled and dried with a SpeedVac centrifuge . The dried extracts were thoroughly dissolved in EIA buffer at 4°C overnight prior to the EIA assay . 20E EIA antiserum ( #482202 ) , 20E AChE tracer ( #482200 ) , Precoated ( Mouse Anti-Rabbit IgG ) EIA 96-Well Plates ( #400007 ) , and Ellman's Reagent ( #400050 ) were all purchased from Cayman Chemical , and assays were performed according to the manufacturer's instructions . DIG-labeled RNA probes were generated by in vitro transcription following the manufacturer's instructions ( Roche DIG RNA Labeling Mix , #11 277 073 910 ) . L3 larvae were dissected in ice-cold PBS and fixed in 4% paraformaldehyde for 20 min at RT . After treatment with 1% H2O2 , samples were stored in hybridization buffer at −20°C . Tissues were pre-hybridized in hybridization buffer for 3 h at 58°C and RNA probes were denatured for 3 min at 80°C h . Probe hybridization was performed for 18 h at 58°C , followed by extensive wash steps at 58°C . After cooling , tissues were blocked with PBTB buffer ( 2% NGS and 1% BSA ) for 1 h at RT before overnight incubation with mouse Anti-Digoxin antibody ( Jackson ImmunoResearch Cat . # 200-062-156 , 1∶500 dilution ) at 4°C . Tissues were then incubated with streptavidin-HRP conjugates ( Molecular Probes #S991 , 1∶400 dilution ) in PBTB for 1 h at RT , followed by six wash steps ( 1 h each ) in PBTB at RT . Before TSA amplification , tissues were washed in PBTB . Tyramide reagents ( PerkinElmer TSA Plus Cyanine 3 Kit , Cat . #NEL744001KT ) were diluted 1∶1000 in 1× amplification buffer provided by the kit . TSA reactions were performed for 40 min at RT and washed 6 times for 1 h in PBS at RT . Tissues were mounted in the ProLong Gold antifade reagent ( Invitrogen , P36934 ) and analyzed by confocal microscopy ( Nikon AZ-C1 Confocal Microscope System ) . | Steroid hormones play fundamental roles in development and disease . They are often released as pulses , thereby orchestrating multiple physiological and developmental changes throughout the body . Hormone pulses must be regulated in a way so that they have a defined beginning , peak , and end . In Drosophila , pulses of the steroid hormone ecdysone govern all major developmental transitions , such as the molts or the transformation of a larva to a pupa . While we have a relatively good understanding of how an ecdysone pulse is initiated , little is known about how hormone production is turned off . In this study , we identify a critical regulator of this process , the nuclear receptor DHR4 . When we interfere with the function of DHR4 specifically in the ecdysone-producing gland , we find that larvae develop much faster than normal , and that this is caused by the inability to turn off ecdysone production . We show that DHR4 oscillates between cytoplasm and nucleus of ecdysone-producing cells under the control of a neuropeptide that regulates ecdysone production . When the neuropeptide pathway is inactive , DHR4 enters the nucleus and represses another gene , Cyp6t3 , for which we show a novel role in the production of ecdysone . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"neurochemistry",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"molecular",
"development",
"neuropeptides",
"erk",
"signaling",
"cascade",
"neurochemicals",
"nuclear",
"receptor",
"signaling",
"biology",
"molecular",
"biology",
"biochemistry",
"signal",
"transduction",
"signaling",
"molecular",
"cell",
"biology",
"signaling",
"cascades"
]
| 2011 | Nuclear Receptor DHR4 Controls the Timing of Steroid Hormone Pulses During Drosophila Development |
Targeting key regulators of the cancer stem cell phenotype to overcome their critical influence on tumor growth is a promising new strategy for cancer treatment . Here we present a modeling framework that operates at both the cellular and molecular levels , for investigating IL-6 mediated , cancer stem cell driven tumor growth and targeted treatment with anti-IL6 antibodies . Our immediate goal is to quantify the influence of IL-6 on cancer stem cell self-renewal and survival , and to characterize the subsequent impact on tumor growth dynamics . By including the molecular details of IL-6 binding , we are able to quantify the temporal changes in fractional occupancies of bound receptors and their influence on tumor volume . There is a strong correlation between the model output and experimental data for primary tumor xenografts . We also used the model to predict tumor response to administration of the humanized IL-6R monoclonal antibody , tocilizumab ( TCZ ) , and we found that as little as 1mg/kg of TCZ administered weekly for 7 weeks is sufficient to result in tumor reduction and a sustained deceleration of tumor growth .
It is widely believed , based on increasing evidence , that a small population of tumorigenic cells , which are in many ways similar to normal adult stem cells , is responsible for the initiation and maintenance of malignant tumors [1–5] . This concept , termed the cancer stem cell ( CSC ) hypothesis , takes the view that tumors , like adult tissues , arise from multipotent cells that exhibit the ability to self-renew as well as give rise to differentiated tissue cells [4–7] . It is hypothesized that CSCs are responsible for tumor initiation , progression , resistance and recurrence [4 , 6 , 8] . Cancer stem cells have now been identified in a variety of malignancies , including tumors of the blood , breast , colon , brain , and head and neck [8] . Head and neck squamous cell carcinoma ( HNSCC ) , a highly invasive form of cancer , is the sixth most common cancer in the world , with over 600 , 000 new cases diagnosed globally each year [9] . The identification of cancer stem cells as drivers of the tumorigenic process in HNSCC [4] provides a rationale for the targeted elimination of these cells in HNSCC tumors . It is well known that growth and survival of CSCs is highly influenced by tumor micro-environmental factors and molecular signaling , initiated by cytokines and growth factors [10–13] . IL-6 is a pleiotropic cytokine , secreted by a variety of cell types , that is a key player in number of cellular processes including proliferation , survival , differentiation , migration and invasion [14] . It is also commonly overexpressed in most cancer types including HNSCC [8 , 14 , 15] . High IL-6 expression independently predicts tumor recurrence , tumor metastasis and poor survival in head and neck cancer patients [14] . IL-6 signaling is mediated by binding to its natural receptor , IL-6R and the universally expressed gp130 receptor . Once bound to IL6 , the IL-6R-gp130 complex results in the phosphorylation of STAT3 , which is indicative of stemness [8] . Recent evidence shows that IL-6R is overexpressed on CSCs and IL-6 secreted by both tumor cells and endothelial cells ( ECs ) enhances the survival , self-renewal and tumor initiation potential of cancer stem cells in HNSCC [8] . Given that HNSCC has a 5-year survival rate of less than 60% , which has improved little over the last 20 years [16] , these studies of the impact of IL-6 on CSCs provide strong motivation for the development of anti-IL-6 therapies for the targeted treatment of HNSCC . The fact that CSCs form only a small portion of the total tumor burden , but may play a disproportionately important role in determining tumor growth and treatment outcomes makes them an important cellular phenotype in need of further study . In this paper , we develop a predictive computational framework that aims to advance our current understanding of the differential impact of IL-6 on CSC self-renewal and HNSCC growth and investigate the mechanisms of tumor reduction associated targeted treatment with the anti-IL-6R antibody Tocilizumab . While numerous models of cancer stem cell driven tumor growth exist ( see [17] for a review ) , connecting mathematical models of the CSC hypothesis to experimental data , either at the molecular and cellular scale or at the clinical scale is far less common , for examples see [18–21] . Our model is unique in that it includes the molecular level details IL-6 signal initiation and its effect on tumor cell survival and proliferation , while also capturing the influence of IL-6 on the probability of self-renewal for cancer stem cells . To our knowledge , this is the first model of cancer stem cell driven tumor growth that operates across molecular and cellular scales . Our model allows for the quantification of the temporal changes fractional occupancies of bound IL-6 receptors and their impact on tumor growth dynamics , which is precisely the level of detail required to better understand targeted therapies that antagonize IL-6 signaling .
This mathematical model has as its foundation specific biological knowledge of the function of IL-6 signaling and the differential cellular responses to it . The pre-treatment model tracks the temporal evolution of three cancer cell types ( stem , progenitor , and terminally differentiated ) as well as IL-6 and membrane bound IL-6 receptors ( IL-6R ) in their free and bound forms as depicted in Fig 1 . Although a soluble form of IL-6R ( sIL-6R ) exists and can bind IL-6 with a similar affinity as the membrane bound form [22] , we choose to simplify our modeling approach by not including sIL-6R because there is evidence that its role is most important during trans signaling when cells lack membrane bound IL-6R [14 , 22 , 23] . Table 1 lists each model variable along with its units . Eq ( 7 ) describes the association , at rate kf , and dissociation , at rate kr , of IL-6 ( L ) to its cell-bound receptors ( RS , RE and RD ) on stem , progenitor and differentiated cells; respectively ( Fig 1B ) . An underlying assumption in this equation is that the binding rates are the same , independent of cell type . IL-6 is removed via natural decay at rate λL and is produced by tumor cells at rate ρ . d L d t = − k f L R S ︸ IL-6 binding to stem cells + k r C S ︸ IL-6 dissociation from stem cells − k f L R E ︸ IL-6 binding to progenitor cells + k r C E ︸ IL-6 dissociation from progenitor cells − k f L R D ︸ IL-6 binding to differentiated cells + k r C D ︸ IL-6 dissociation from differentiated cells − λ L L ︸ IL-6 natural decay + ρ ( S + E + D ) ︸ IL-6 Production by tumor cells ( 7 ) Eqs ( 8 ) , ( 9 ) and ( 10 ) model the temporal changes in free IL-6 receptors on each of the cell types that we are considering . The first two terms in each equation are the association and dissociation of IL-6 to IL-6R . The recycling terms describe the reactions by which IL-6 is used up in the processes of mediating its cellular response , and the free receptors are recycled back to the cell surface . Following the formulation in [38] , the last two terms in each equation describe the production of new free receptors as new cells are generated and the loss of these receptors as cells die . Definitions of P ( ⋅ ) and D ( ⋅ ) are provided in the following section . We note that when IL-6 binds to IL-6R , it subsequently recruits a GP130 molecule to form a ternary complex ( IL-6/IL-6R/GP130 ) [15] . However , we do not model GP130 explicitly here , instead , we indirectly account for its role in the binding parameters and recycling parameters . d R S d t = − k f L R S ︸ IL-6 binding to stem cells + k r C S ︸ IL-6 dissociation from stem cells + k p C S ︸ Recycling + R T S P S ( S , ϕ S ) ︸ Generation of new R S via cell proliferation − R S R S + C S R T S D S ( S , ϕ S ) ︸ Loss of R S via cell death ( 8 ) d R E E t = − k f L R E ︸ IL-6 binding to progenitor cells + k r C E ︸ IL-6 dissociation from progenitor cells + k p C E ︸ Recycling + R T E P E ( E , ϕ E ) ︸ Generation of new R E via cell proliferation − R E R E + C E R T E D E ( E , ϕ E ) ︸ Loss of R E via cell death ( 9 ) d R D D t = − k f L R D ︸ IL-6 binding to differentiated cells + k r C D ︸ IL-6 dissociation from differentiated cells + k p C D ︸ Recycling + R T D P D ( D , ϕ D ) ︸ Generation of new R D via cell proliferation − R D R D + C D R T D D D ( D , ϕ D ) ︸ Loss of R D via cell death ( 10 ) Eqs ( 11 ) , ( 12 ) and ( 13 ) are analogous to the ones above , as they describe changes in receptor-ligand complexes on each cell type . Similarly , in these equations , the internalization terms describe the reactions by which the complex is internalized and the free receptors are recycled to the cell surface . The last term in each equation describes the loss of these receptor complexes due to cell death . d C S d t = k f L R S ︸ IL-6 binding to R S − k r C S ︸ IL-6 dissociation from R S − k p C S ︸ Internalization − C S R S + C S R T S D S ( S , ϕ S ) ︸ Loss of C S via cell death ( 11 ) d C E d t = k f L R E ︸ IL-6 binding to R E − k r C E ︸ IL-6 dissociation from R E − k p C E ︸ Internalization − C E R E + C E R T E D E ( E , ϕ E ) ︸ Loss of C E via cell death ( 12 ) d C D d t = k f L R D ︸ IL-6 binding to R D − k r C D ︸ IL-6 dissociation from R D − k p C D ︸ Internalization − C D R D + C D R T D D D ( D , ϕ D ) ︸ Loss of C D via cell death ( 13 ) Input parameters necessary to characterize the dynamics of the CSC , progenitor and differentiated cell pools include the cell division and death rates as well as the probability of stem and progenitor cell self-renewal . The proportion of cancer stem cells ( CSCs ) within a tumor varies widely among cancer types and cell lines [39] . CSCs make up only a fraction of 1% of the proliferating cells in the bone marrow and approximately 1 − 10% of the proliferating cells in epithelial cancers . Parameter values for cancer stem cells ( including symmetric/asymmetric division rates ) also vary widely across tumor types . In [19] the cell-cycle length is approximated around Tc = 25 hours which is in agreement with the result given in [40] in which Tc is estimated to be varying between one and two days . Therefore , for our numerical simulations we use αE = ln2/1 . 04 and αS ∈ [ln2/1 . 04/2 , ln2/1 . 04] . The death rate of differentiated cancer cells , δD , has varied widely in a window between 0 . 01 per day to 15-18 per week in previous studies [17 , 27 , 41 , 42] . Finally , under this assumed hierarchical structure , CSCs live longer than both progenitor and differentiated caner cells [1] , so the maximum death rate of progenitor and differentiated cells ( δE , δD ) is chosen to be close to but larger than the death rate of CSCs , δS . The parameter values obtained from the literature are tabulated in Table 2 . For those parameters that there was little or no published information , we compute a best fit to experimental data to obtain reasonable estimates . In the Results section , we also perform sensitivity analysis to find the most influential parameters on the tumor growth , percentage of CSCs and the fractional occupancies of bound IL-6 receptors on CSCs . To begin to understand the impact of stromal IL-6 on the survival of CSCs , Krishnamurthy et al . [8] generated tumor xenografts by transplanting primary human cancer stem-like cells in severe combined immunodeficient mice . Specifically , immediately after surgical removal of the primary tumor from patients with HNSCC , ALDHHIGHCD44HIGH cells were sorted and transplanted into IL-6 +/+ or IL-6 -/- immunodeficient mice . This approach differs from scaffold experiments where tumor xenografts , vascularized with functional human microvessels , are generated in SCID mice . In that experimental setup , human tumor cells are seeded along with human dermal microvascular endothelial cells ( HDMECs ) in poly ( L-lactic ) acid biodegradable scaffolds , resulting in the growth of human tumors with human vasculature , and an additional source of human IL-6 ( the HDMECs ) . In the experimental setup modeled here , no human endothelial cells are implanted and the only source of human IL-6 are the tumor cells themselves . Another difference between the experimental approach modeled here and others in the literature is the use of primary tumor cells and not immortalized tumor cell lines . Fig 3 shows the relevant data presented in [8] . When 1 , 000 ALDHHIGHCD44HIGH were cells transplanted into the IL-6 +/+ mice , the result was more and larger tumors as compared to the transplantation of 1000 ( ALDHHIGHCD44HIGH ) into IL-6 -/- deficient littermates . We now extend our model by modifying it to include the therapeutic administration of Tocilizumab ( TCZ ) , an anti-IL-6R antibody , to study the response of tumor cells to this targeted treatment .
Numerical simulations of pretreatment tumor growth are presented in Fig 5-A . There is a strong correlation between the model output ( red ) and the experimental data in [8] ( blue ) . The green line in Fig 5-A represents the tumor volume over time when tumor cells do not produce IL-6 , thereby showing how much even low secretion rates of IL-6 ( ρ = 7 × 10 - 7 fMol cell × day ) influence tumor growth . In addition , we use the best-fit parameter values to predict the percentage of CSCs on the last day of the experiment ( see the Methods section for details on the experimental design ) . Fig 5-B shows the experimentally measured percentage of CSCs in primary tumors ( brown ) , the experimentally measured percentage of CSCs on day 121 for tumors grown in IL-6 +/+ mice ( blue ) , along with our mathematical model prediction ( red ) . The model is able to accurately capture the correct proportion of stem cells and Fig 5-D shows how the stem cell percentage evolves over time . The first step in the IL-6 signal transduction pathway is to binding to IL-6R . The IL-6-IL-6R complex then recruits GP130 . The complex of IL-6-IL-6R-GP130 activates signaling pathways ( such as STAT3 ) [55 , 56] that play a critical role in the self-renewal and survival of CSCs . Therefore , the fractional occupancies of bound receptors can be a useful tool for quantifying the influence of tumor cell-secreted IL-6 on the tumorigenic potential of CSCs and subsequently on tumor growth dynamics . Fig 5-C plots the fractional occupancy of IL-6R on CSCs over time for our baseline level of tumor secretion of IL-6 . The model suggests that a fractional occupancy of 12% on CSCs is sufficient to result in the experimentally observed tumor growth rate . In fact , because endothelial cells can secrete higher levels of IL-6 than tumor cells [8] , if we were to add endothelial cells to our model then we would expect even greater interdependencies among IL-6 , tumor growth dynamics and the tumorigenic potential of CSCs . Values for the IL-6 secretion rate by tumor cells , ρ , vary quite widely in the literature [45–47] . Our baseline value of ρ = 7×10−7 corresponds to 200 pg per ml per 106 HNSCC cells per day , which is within the ranges reported in [45–47] . Fig 6 shows that if we keep all the other parameters at their baseline values provided in Tables 2 and 3 , while varying the secretion rate of IL-6 , ρ , then relatively small increases in ρ ( from ρ = 7×10−7 to 5 . 35×10−6 fmol/cell/day ) lead to 90% fractional occupancy . This supports the idea that an IL-6 antagonist could temper the effects of IL-6-induced pathways , thereby impeding tumor growth . We used the treatment model ( see S1 Appendix ) together with the parameter values listed in Tables 2 , 3 and 5 to predict tumor response to weekly administration of 1mg/kg or 5mg/kg of TCZ for 7 weeks . Fig 10-A shows the model predictions for the amount of TCZ within the tumor for the two different doses . Our model predicts a 25% and 28% reduction in the tumor volume as compared with tumor volume without treatment for the two different dosing strategies ( Fig 10-B ) . This result is very similar to the experimentally observed tumor reduction shown in [54] for UM-HMC-3B , salivary gland cancer xenografts .
It is clear that IL-6 plays a critical role in the pathobiology of cancer , due in part to its impact on cancer stem cells . This has provided strong rationale for developing targeted inhibitors of IL-6 . This modeling study not only quantifies the influence on IL-6 on primary tumor xenografts; it also provides some explanations for the various effects of TCZ on tumor growth and CSC percentage . We are currently modifying this model to describe xenografts that include human endothelial cells that have been demonstrated to produce IL-6 in greater amounts than tumor cells . Preliminary results for this approach , some of which were described above , are promising . We will also extend the model to include combination therapies with traditional chemotherapeutic agents , like cisplatin . This extended model can be used to simulate different dose-scheduling regimens in order to investigate synergism between the two therapies . Continued modeling efforts in this direction have the potential to shed light on conditions under which TCZ sensitizes cancer cells for treatment with cisplatin and can be used to predict the optimal dose scheduling that will lead to maximal tumor response . | A small population of cancer stem cells that share many of the biological characteristics of normal adult stem cells are believed to initiate and sustain tumor growth for a wide variety of malignancies . Growth and survival of these cancer stem cells is highly influenced by tumor micro-environmental factors and molecular signaling initiated by cytokines and growth factors . This work focuses on quantifying the influence of IL-6 , a pleiotropic cytokine secreted by a variety of cell types , on cancer stem cell self-renewal and survival . We present a mathematical model for IL-6 mediated , cancer stem cell driven tumor growth that operates at the following levels: ( 1 ) the molecular level—capturing cell surface dynamics of receptor-ligand binding and receptor activation that lead to intra-cellular signal transduction cascades; and ( 2 ) the cellular level—describing tumor growth , cellular composition , and response to treatments targeted against IL-6 . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
]
| [
"cell",
"binding",
"cell",
"physiology",
"cell",
"death",
"tumor",
"stem",
"cells",
"medicine",
"and",
"health",
"sciences",
"endothelial",
"cells",
"cancer",
"treatment",
"cell",
"processes",
"cancer",
"stem",
"cells",
"cell",
"differentiation",
"epithelial",
"cells",
"oncology",
"developmental",
"biology",
"stem",
"cells",
"animal",
"cells",
"biological",
"tissue",
"carcinogenesis",
"cell",
"biology",
"anatomy",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
]
| 2018 | A mathematical model for IL-6-mediated, stem cell driven tumor growth and targeted treatment |
Accurate temporal control of gene expression is essential for normal development and must be robust to natural genetic and environmental variation . Studying gene expression variation within and between related species can delineate the level of expression variability that development can tolerate . Here we exploit the comprehensive model of sea urchin gene regulatory networks and generate high-density expression profiles of key regulatory genes of the Mediterranean sea urchin , Paracentrotus lividus ( Pl ) . The high resolution of our studies reveals highly reproducible gene initiation times that have lower variation than those of maximal mRNA levels between different individuals of the same species . This observation supports a threshold behavior of gene activation that is less sensitive to input concentrations . We then compare Mediterranean sea urchin gene expression profiles to those of its Pacific Ocean relative , Strongylocentrotus purpuratus ( Sp ) . These species shared a common ancestor about 40 million years ago and show highly similar embryonic morphologies . Our comparative analyses of five regulatory circuits operating in different embryonic territories reveal a high conservation of the temporal order of gene activation but also some cases of divergence . A linear ratio of 1 . 3-fold between gene initiation times in Pl and Sp is partially explained by scaling of the developmental rates with temperature . Scaling the developmental rates according to the estimated Sp-Pl ratio and normalizing the expression levels reveals a striking conservation of relative dynamics of gene expression between the species . Overall , our findings demonstrate the ability of biological developmental systems to tightly control the timing of gene activation and relative dynamics and overcome expression noise induced by genetic variation and growth conditions .
Normal development requires precise temporal control of differential gene expression , yet development must be robust to natural genetic variation and environmental changes . [1–3] . This resilience of developmental systems is important for keeping a wide genotypic pool adaptable in changing environmental conditions and thus , for the survival of the species [4 , 5] . Identifying how the control systems overcome genetic and environmental changes is important to the mechanistic understanding of developmental processes and their evolution [1 , 3 , 4] . Specifically , comparing different aspects of expression dynamics between individuals within the species and between closely related species can illuminate the range of variation in temporal expression that can still produce similar embryonic structures [1 , 6–8] . Comparative studies of interspecies differences in the kinetics of gene regulatory circuits can provide predictions for trans and cis evolutionary changes in circuit connectivity . The timing of gene expression depends on the temporal expression profiles of the inputs ( trans ) and the logic applied on the inputs by the cis-regulatory modules [9 , 10] ( S1A–S1C Fig ) . For example , if two inputs are activated sequentially and the target cis-regulatory element requires both of them ( necessary inputs , AND logic ) , the target gene will turn on only after the activation of the later input gene ( S1B Fig ) [9] . If the two inputs are additive ( OR logic ) , the target gene will turn on immediately after the activation of the earlier input gene [11] ( S1C Fig ) . Thus , evolutionary changes in cis-regulatory logic , e . g . from AND to OR , could result in changes in gene expression timing . Comparing the expression profiles of both input and target genes between two species can provide predictions for changes in input dynamics and in the target's cis-regulatory logic . Comparative studies of temporal variation of gene regulatory circuits between related species must rely on detailed experimentally-based models of the gene regulatory networks in these organisms . The current models of the gene regulatory networks that drive ectoderm , endoderm and mesoderm specification in the sea urchin embryo are among the most comprehensive of their kind and are based on experimental studies in a few main species . [12–16] . The purple sea urchin , Strongylocentrotus purpuratus ( Sp ) inhabits the Pacific coasts of North America while the sea urchin Paracentrotus lividus ( Pl ) inhabits the eastern Atlantic Ocean and the Mediterranean Sea . These species shared a common ancestor about 40 million years ago and the average similarity in their coding sequences is about 85% , which is similar to that found between human and mouse . The growth temperature of these two species is different , reflecting their different environments; Pl embryos will successfully develop over a temperature range that is higher than Sp ( standard lab temperatures 18°C versus 15°C , respectively ) . These species show apparent similarities in size , morphology , spatial gene expression patterns and gene regulatory networks , despite their genomic divergence and geographic distance ( Fig 1A ) [14–25] . High resolution studies of the temporal expression profiles of more than a hundred regulatory and differentiation genes that operate at different embryonic territories were performed for Sp [13 , 26] , but equivalent information for Pl is still limited [18] . Here , we perform high-resolution quantitative analysis of the transcriptional expression profiles of key regulatory genes in Pl , asses the temporal expression variation within the species and compare gene expression dynamics to those measured in Sp [26] . For these studies , we selected regulatory circuits that operate in five embryonic territories and contain common network motifs found in many other gene regulatory networks , such as positive feedback and feedforward structures . The positive feedback circuitry locks down a specification state within a cell ( intracellular , S1D Fig ) or within an embryonic territory ( intercellular , S1E Fig ) and is important for cell fate decision [15 , 27–29] . Coherent and incoherent feedforward motifs are used for the sequential activation of genes in a cell ( S1F Fig ) [30–32] . Our results portray a tight control of timing of gene activation that is highly conserved between the species despite their genetic and geographic distance . The developmental rates of the two species scale linearly , in agreement with the species’ different growth temperatures . When we scale the developmental rates of the two species , we reveal a remarkable conservation of relative expression dynamics . Thus our study illuminates the dynamic properties of biological regulatory systems and their ability to control relative dynamics accurately despite genetic and growth condition differences .
Comparing gene expression profiles between Pl and Sp can identify both conserved and diverged expression patterns and suggest similarity and changes in circuits’ connectivity . High resolution time courses in Sp were measured by nanostring up to 48 hpf in this species [26] , which includes the time interval 0–30 hpf in Pl ( Fig 1A ) . While comparing actual mRNA levels between species is difficult due to the different methods used [26] , comparison of initiation times and relative gene expression levels is possible . In Fig 3 , we present comparative expression profiles of the studied genes separated into five regulatory circuits that initiate the specification of the skeletogenic mesoderm ( Fig 3A–3C ) , the aboral non-skeletogenic mesoderm that form pigment cells ( Fig 3D–3F ) , the endoderm ( Fig 3G–3I ) , the aboral ectoderm ( Fig 3J–3L ) and the oral ectoderm ( Fig 3M–3O ) . Sp expression profiles are taken from Materna et al , 2010 , [26] ( running averages of two biological replicates measured by nanostring technique , the data is available at http://vanbeneden . caltech . edu/~m/cgi-bin/hd-tc/plot . cgi ) . The circuit diagrams are based on experimental validations that include perturbation and cis-regulatory analysis in Sp [13 , 15 , 27 , 33 , 34 , 36] . Below we discuss the level of conservation of each circuit between the species in the light of our temporal expression comparison and previous studies .
Embryo development generates similar morphologies despite natural genetic variation and within broad environmental conditions . This flexibility of the developmental program is essential for the survival of the species and keeping a wide genotypic pool adaptable in a changing environment . Understanding the properties of the regulatory control system that underlie cell fate specification is a key to the mechanistic understanding of this developmental stability . Here we studied the reproducibility and conservation of expression dynamics of regulatory circuits in two sea urchin species that shared common ancestor about 40 million years ago and inhabit distinct geographic habitats . Embryo size , cell types and morphologies of these two species are highly similar despite their genomic and geographic distance ( Fig 1 ) . Our studies illuminate tight control of gene activation timing within the species ( Fig 2 ) and a striking similarity of relative dynamics revealed by scaling the developmental rates of the two species and normalizing gene expression levels ( Fig 5 ) . The regulatory systems that enable this reproducibility and conservation are the underlying mechanisms of morphological similarity amidst genetic and environmental variation . The high resolution of our studies reveals tight control of initiation times that show lower variation than the variations in maximal mRNA levels between different individuals in the same species ( Fig 2 ) . Interestingly , lower variations of initiation time compared to the variation of expression levels were also detected in a comparative study of the developmental transcriptomes of two Xenopus species [35] . Previous studies in yeast provide a possible mechanistic explanation of these findings [56 , 57] . These studies show explicitly that the initiation of gene activation is highly similar for different levels of the activating input once the input level is above a certain threshold for long enough time [56] . On the other hand , once the gene is on , the level of gene expression is highly dependent of the level of the activating input . The molecular explanation for the different behavior of initiation timing and expression level was suggested by the same group several years before [57] . Their measurements and modeling of expression kinetics indicated that the timing of gene initiation is controlled by the slow rate of nucleosome removal from the DNA . Once the nucleosomes are removed , the level of gene expression depends on the affinity of the transcription factor binding sites and the concentration of the activating transcription factor that define the binding site occupancy and the rate of mRNA generation . Thus , the ability to buffer variations in expression level and still tightly control the timing of gene activation , possibly by using nucleosomal positioning as a threshold mechanism , could be a general property of eukaryote gene regulatory networks . Our interspecies comparison of temporal expression profiles of key regulatory circuits revealed a high conservation of the temporal order of gene activation within the circuits but also some cases of divergence ( Fig 3 ) . Integrating the differences in temporal profiles with available perturbation and spatial expression data provides predictions for specific cis-regulatory changes within the ectodermal circuits . The highest interspecies conservation of temporal ordering and the timing of gene activation are observed in the endoderm circuit ( Figs 2G–2I and 5C ) . This degree of conservation supports the conservation of both the architecture and the cis-regulatory logic of this circuit . The endodermal circuit is one of the most conserved circuits within echinoderms , with a similar architecture detected in the sea star that shared a common ancestor with the sea urchin ~500 mya [58 , 59] . The mesodermal and ectodermal networks show higher variation of circuit connectivity between the sea urchin and sea star [60–62] , emphasizing the strong developmental constraints on the endoderm circuit . The constraints that define this high degree of temporal conservation could be the requirement to initiate gastrulation and the invagination of the gut at the right developmental time . Thus , high-resolution comparison of circuits’ dynamics is a good tool for the prediction of conservation and changes in circuit connectivity when the general circuit structure is known at least in one of the species . We used gene initiation times measured in the two species to estimate a ×1 . 3 ratio between the molecular developmental rates in Pl and Sp ( Fig 4 ) . Apparently , a major contribution to the accelerated developmental rate in Pl is its higher culture temperature compared to the culture temperature of Sp ( 18°C in Pl vs . 15°C in Sp ) . A recent study had shown that when Sp embryos are cultured in 18°C their developmental rate increases by about ×1 . 24 fold based on morphological comparison , close to the ratio we obtained [3] . This is in agreement with recent studies in invertebrate and vertebrate embryos that show morphological and molecular scaling with temperature of diverse species [35 , 63] . A recent morphological comparison of ten Drosophila species shows that the rate of embryogenesis scales with temperature within a wide range of temperature ( 17 . 5°C-32°C ) [63] . A comparative study of two Xenopus species grown in different culture temperature ( 28°C vs . 22°C ) shows that the rate of embryogenesis scales with temperature based on morphology and on the timing of gene activation for most studied genes [35] . Thus , the ability to adapt to different temperatures by scaling the developmental rates without distinct morphological phenotypes is a common property to both vertebrate and invertebrate species . Our studies reveal remarkable interspecies conservation of expression dynamics when the developmental rates of the two species are scaled and gene expression levels are normalized ( Fig 5 ) . This demonstrates an impressive ability of biological developmental systems to tightly control gene activation timing and relative expression dynamics despite genetic and growth conditions differences . This raises the question: Is the observed conservation an outcome of a strong negative selection against genetic changes of regulatory circuits or due to the structure of regulatory circuits that buffers genetic and environmental changes ? We tend to support the second option and the ability of the regulatory system to overcome expression noise . This could be achieved by noise filtration mechanisms , e . g . , the threshold activation suggested above , or by the use of network motifs that define different levels of sensitivity to upstream variation . For example , computational studies show that positive feedback circuitry is more efficient than other architectures in buffering noise in the inducing signal while keeping high responsivity to the level of the signal [64 , 65] . On the other hand , incoherent feedforward motifs can generate consistent response to activating input that depends mostly on fold changes in input and not on noisy absolute protein levels [64 , 66–68] . Apparently , this flexible design of gene regulatory circuits enables them to conserve similar expression dynamics and specify similar cell types while allowing the species to keep a broad genotypic variance and survive through changing environmental conditions .
Adult sea urchins were supplied from a mariculture facility of the Israel Oceanographic and Limnological research in Eilat . Sea urchin eggs and sperm were obtained by injecting adult sea urchins with 0 . 5M KCl . Embryos were cultured at 18°C in artificial sea water . Total RNA was extracted using Qiagen mini RNeasy kit from embryos at indicated time points . 1 μg of total RNA from each time point of each three independent biological replicates was used to generate cDNA using BioRad i-script kit and subsequently used for QPCR . Initiation times , t0 , for both Fig 2D and Fig 4 were estimated by the use of the sigmoid function: Log ( mRNA ( t ) ) = a − b/ ( 1 + exp ( c ( t − t0 ) ) as in Yanai et el , 2011 [35] . The sigmoid was fit using Matlab’s Curve Fitting Toolbox , using the nonlinear least-squares method . For all genes R2>0 . 94 , except from PlDlx repeat #3 that had R2 = 0 . 88 . To calculate Pearson correlation between Sp and Pl time course , we first scaled the developmental rates in the two species using a factor of ×1 . 3 between Pl and Sp . The exact time points we compared are given in S1 Table . Then we calculated Pearson correlation between the averaged expression levels in Sp and Pl using the CORREL function in excel . | Embryonic development necessitates a delicate balancing act . On one hand , precise regulation of the expression of developmental genes is crucial for the maintenance of morphology and function . On the other hand , these same regulatory networks must allow normal development to proceed through genetic variation and environmental changes . To learn how regulatory circuits operate robustly within natural variation , we study the temporal expression profiles of key regulatory genes in the Mediterranean sea urchin , Paracentrotus lividus , and compare them to those of its Pacific Ocean relative , Strongylocentrotus purpuratus . These species shared a common ancestor about 40 million years ago and show highly similar embryonic morphologies . Our studies reveal highly reproducible gene initiation times that show lower variations than the variations in maximal mRNA levels within the species ( Pl ) . We observe high interspecies conservation of the temporal order of gene activation within regulatory circuits and some cases of divergence . This conservation was even more profound when expression levels were normalized and scaled to the different developmental rates between the species . Our findings highlight that , despite genetic variations and different growth conditions , expression dynamics in developmental gene regulatory networks are extremely conserved over 40 million years of evolution . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Comparative Study of Regulatory Circuits in Two Sea Urchin Species Reveals Tight Control of Timing and High Conservation of Expression Dynamics |
Buruli ulcer , caused by infection with Mycobacterium ulcerans , is a necrotizing disease of the skin and subcutaneous tissue , which is most prevalent in rural regions of West African countries . The majority of clinical presentations seen in patients are ulcers on limbs that can be treated by eight weeks of antibiotic therapy . Nevertheless , scarring and permanent disabilities occur frequently and Buruli ulcer still causes high morbidity . A vaccine against the disease is so far not available but would be of great benefit if used for prophylaxis as well as therapy . In the present study , vesicular stomatitis virus-based RNA replicon particles encoding the M . ulcerans proteins MUL2232 and MUL3720 were generated and the expression of the recombinant antigens characterized in vitro . Immunisation of mice with the recombinant replicon particles elicited antibodies that reacted with the endogenous antigens of M . ulcerans cells . A prime-boost immunization regimen with MUL2232-recombinant replicon particles and recombinant MUL2232 protein induced a strong immune response but only slightly reduced bacterial multiplication in a mouse model of M . ulcerans infection . We conclude that a monovalent vaccine based on the MUL2232 antigen will probably not sufficiently control M . ulcerans infection in humans .
Mycobacterium ulcerans causes Buruli ulcer ( BU ) , a disease of the skin and underlying subcutaneous tissue , which is reported from over 30 countries worldwide [1] . BU is most prevalent in West African countries and mainly affects children less than 15 years of age , living in remote , rural areas [2] . The natural reservoir of M . ulcerans has not been identified so far and also the mode of transmission of this pathogen remains unclear [3] . M . ulcerans produces a macrolide exotoxin called mycolactone , which induces apoptosis in mammalian cells and leads to the typical clinical presentation of ulcerative BU skin lesions after the overlying epidermis has collapsed [4] . Non-ulcerative forms of the disease are nodules or papules , oedema and plaques [5] . BU was traditionally treated by wide surgical excision of the affected skin and skin grafting if necessary . Since 2004 , treatment of patients for eight weeks with the antibiotics rifampicin and streptomycin is recommended as standard therapy by the World Health Organization ( WHO ) [6] . Even though the use of antibiotics has reduced recurrence rates to less than 2% [7–9] , patients are often left with scars and lifelong disabilities [10] . A vaccine against BU would therefore be of high value to prevent and treat the disease [11] . As opposed to M . tuberculosis and M . leprae , which are both intracellular pathogens for which T helper 1 ( TH1 ) cellular immune responses are essential for infection control [12] , M . ulcerans is found in extracellular clumps in the necrotic subcutaneous skin tissue of advanced lesions . However , there is evidence for an initial intracellular stage in macrophages during the early phase of infection [13 , 14] . Correlates of protection have not been identified , and in particular it is not known , whether antibodies specific for surface antigens of M . ulcerans have protective activity [15] . Vaccination with Bacille Calmette-Guérin ( BCG ) seems to confer a transient protection from BU and a shorter duration of ulcers [16–18] . In a mouse model of M . ulcerans infection , vaccination with either BCG or a mycolactone-negative M . ulcerans mutant strain delayed the onset of foot pad swelling [19 , 20] . A limited protective efficacy has also been achieved with monovalent DNA-based protein subunit vaccine formulations targeting either the mycolyl-transferase antigen ( Ag ) 85A from M . bovis or M . ulcerans [21 , 22] or mycolactone polyketide synthase domains that are encoded on the giant M . ulcerans plasmid pMUM001 [23] . Vesicular stomatitis virus ( VSV ) is a member of the virus family Rhabdoviridae and has a non-segmented single-stranded RNA genome of negative polarity . VSV is transmitted by insects and causes a vesicular-like disease in livestock and Flu-like symptoms in humans . The VSV seroprevalence is very low in the human population , indicating that human infections are rare . The non-persisting replication of VSV in the cytosol of the host cell , low virulence and low pre-existing immunity in humans , and the option of simple genetic manipulation [24] has recommended VSV as viral vector system for vaccination [25 , 26] . Several VSV-based vaccines for protection against viral and bacterial pathogens have been generated in recent years and have been evaluated in a number of animal models [27–30] . More recently , single-cycle VSV vectors have been developed , which lack the genetic information for the VSV glycoprotein G and are complemented with this protein in trans . Although these VSVΔG vectors are restricted to a single round of infection , they were shown to be as immunogenic as propagation-competent VSV vectors [31] . In the present study we used this safe vaccine vector platform to characterize the immunogenicity of two M . ulcerans protein antigens . In view of the mainly extracellular nature of M . ulcerans , we chose antigens known to be highly expressed on the surface of the bacteria . The 18 kDa small heat shock protein ( MUL2232 ) was identified amongst the most immunodominant antigens expressed by M . ulcerans when serum responses from people living in endemic areas were analysed [32] . A homologue of MUL2232 was found in the genome of M . leprae but neither in M . bovis nor in M . tuberculosis . Similar to the M . leprae homologue , MUL2232 is associated with the cell wall fraction of M . ulcerans [32 , 33] . The second antigen chosen , MUL3720 , is a 21 kDa protein with a putative lectin and a peptidoglycoan-binding domain that does not have homologues in M . leprae or M . tuberculosis [34 , 35] . It is highly expressed by M . ulcerans and may play a role in cell attachment and cell-cell interactions [35] , making it a candidate antigen for vaccine development . We generated recombinant single-cycle VSV vectors encoding the selected antigens individually and analysed their expression in mammalian cell lines . We further analysed the immunogenicity of the generated VSV vectors and evaluated two short immunization protocols . Finally , we investigated the protective potential of such immunization schemes in an experimental mouse model of M . ulcerans infection .
All animal experiments were conducted in compliance with the Swiss animal protection law and approved by the animal welfare committee of the Canton of Basel ( authorization number 2375 ) and the Canton of Vaud ( authorization number 2657 ) . Baby hamster kidney 21 ( BHK-21 ) cells were obtained from the German Cell Culture Collection ( DSZM , Braunschweig ) . Cells were grown in Earle’s minimal essential medium ( MEM , Life Technologies ) supplemented with 5% foetal bovine serum ( FBS , Biowest ) . BHK-G43 , a transgenic BHK-21 cell line expressing the VSV G protein in an inducible manner , was maintained as described previously [36] . Murine L929 fibroblasts ( ATCC , The Global Bioresource Center ) were cultivated in RPMI medium ( Gibco ) supplemented with 10% foetal calf serum ( FCS , Sigma ) , 2 mM glutamine ( Gibco ) and 0 . 05 mM β-mercaptoethanol ( Gibco ) . The potential protein vaccine candidate antigens MUL_2232 ( GenBank accession number 4550596 ) and MUL_3720 ( GenBank accession number 4553013 ) of M . ulcerans Agy99 were ordered as codon optimized genes for expression in humans ( GenScript ) and received in pUC57 plasmids . For generation of recombinant VSV replicon particles ( VRPs ) , codon optimized target antigens were amplified by PCR and inserted into the pVSV*ΔG plasmid using single MluI and BstEII restriction sites upstream and downstream of the fourth transcription unit , replacing the VSV G gene [37] . Sequence integrity of the resulting plasmids was confirmed by Sanger sequencing . Recombinant VRPs were generated as described previously [38] . In brief , BHK-G43 cells were infected with recombinant MVA-T7 virus expressing T7 RNA polymerase [39] . Subsequently , the infected cells were transfected with plasmids driving T7 RNA polymerase-mediated expression of the VSV proteins N , P , and L , and with pVSV*ΔG ( MUL2232 ) or pVSV*ΔG ( MUL3720 ) driving T7 RNA polymerase-mediated transcription of VSV antigenomic ( negative-sense ) RNA . Expression of the VSV G protein was induced by adding mifepristone ( Sigma ) to the cell culture medium . At 24 hours post transfection cells were detached with trypsin and seeded along with an equal number of fresh BHK-G43 cells into T-75 flasks . Cells were then incubated at 37°C for another 24 hours in the presence of mifepristone . Cell culture supernatant was clarified by low-speed centrifugation and by passage through a 0 . 2 μm pore filter . The VRPs in the clarified cell culture supernatant were further propagated on mifepristone-induced BHK-G43 cells and stored at -70°C . VRPs were titrated on BHK-21 cells taking advantage of the eGFP reporter protein . For indirect immunofluorescence analysis , BHK-21 cells were grown on 12 mm diameter cover slips ( 2 x 105 cells/well ) and infected with VRPs ( 106 infectious units/well ) for 6 hours at 37°C . Cells were fixed with 3% paraformaldehyde ( PFA ) and washed with PBS containing 0 . 1 M of glycine . The cells were permeabilized with 0 . 25% ( v/v ) Triton X-100 and subsequently incubated for 1 hour with anti-MUL2232 or anti-MUL3720 mAbs in appropriate dilutions in 1% bovine serum albumin ( Sigma ) . For detection of antigen-bound primary antibodies , cells were incubated with an Alexa Flour 546 labelled anti-mouse IgG secondary antibody ( 1/500; Molecular Probes , A-11018 ) . The cells were washed with distilled water , embedded in Mowiol 4–88 ( Sigma ) mounting medium , and analyzed with a Leica TCS SL confocal microscope and LCS software ( Leica Microsystems AG , Glattbrugg , Switzerland ) . ELISA plates ( Maxisorp; Nunc ) were coated with 10ug/ml purified recombinant MUL2232 ( rMUL2232 ) or MUL3720 ( rMUL3720 ) protein produced in Escherichia coli [40] . After blocking , plates were incubated with serially diluted sera from immunized mice . Alkaline phosphatase-conjugated goat anti-mouse antibody ( Sigma ) was used as secondary antibody and p-nitrophenyl phosphate ( Sigma ) served as substrate . The optical density ( OD ) of the reaction product was measured at 405 nm with a microplate reader ( Sunrise Absorbance Reader; Tecan ) . The threshold for endpoint titer determination was defined as the double of the mean measurements plus the mean standard deviation of a dilution series done without primary antibody and a dilution series done with pre-bleed serum . Individual serum dilution series were approximated with sigmoidal dose-response curves and the reciprocal dilution of the intersection between the curve and the threshold was defined as individual endpoint titer . All animal studies were conducted in 8 weeks old female BALB/c mice ( Janvier ) . VRPs ( 107 fluorescence-forming units ) were either applied subcutanously ( s . c . ) in the neck ( for evaluation of the vaccination protocol ) or intramuscularly ( i . m . ) into the right caudal tight muscle using a volume of 100 μl or 30 μl , respectively . When several immunizations were conducted , injections were performed in three week intervals . For prime-boost vaccination regimen , 30 μg of non-adjuvanted recombinant protein were applied s . c . into the neck . Blood was collected from the tail vein prior to every immunization as well as 5/6 and 14 days after the protein boost . Serum was prepared by centrifugation of the blood in SST Microtainer tubes ( Becton , Dickinson and Company ) . Immunized mice were euthanized 3 weeks after the second immunization . Heart blood was collected and spleens were aseptically removed and homogenized by passing through a 70 μm cell strainer ( BD Falcon ) . Cells were then pelleted and red blood cells lysed by incubating the pellet in red blood cell lysing buffer ( Sigma ) for 1 minute . Remaining cells were washed several times with stemline T-cell expansion medium ( Sigma ) supplemented with 4mM L-Glutamine ( Gibco ) , 1% Pen-Strep ( Gibco ) , 2 . 5 ug/ml Amphotericin B ( Sigma ) and 0 . 05 mM β-mercaptoethanol ( Gibco ) and finally adjusted to 4 x 106 white blood cells/ml . 7 . 2 x 105 cells per well were incubated in round-bottom microwell plates ( BD Falcon ) in a humidified CO2 incubator and stimulated with Concanavalin A ( 2 μg/ml , Sigma ) as positive control or recombinant proteins at 5 μg/ml . Supernatants were harvested after 24h for Interleukin 2 ( IL-2 ) and 96h for Interleukin 10 ( IL-10 ) and Interferon gamma ( IFNγ ) assays and stored frozen at -20°C until analysis for the selected cytokines . Amount of cytokines in supernatants was determined with Quantikine ELISA kits for IL-2 , IL-10 and IFNγ ( R&D Systems ) . The M . ulcerans strain ( S1013 ) used for the experimental infection of mice was isolated in 2010 from the ulcerative lesion of a Cameroonian BU patient [2] . Bacteria were cultivated for 6 weeks in Bac/T medium ( Biomerieux ) , recovered by centrifugation , and suspended in sterile PBS to 125 mg/ml wet weight corresponding to 2 . 8 x 105 CFU/ml as determined by plating serial dilutions on 7H9 agar plates ( Difco ) . Three weeks after the last immunization , mice were infected with 30 μl of M . ulcerans suspension ( 1/100 of the stock solution in PBS ) into the left hind foot pad . Development of the infection was followed by weekly measurements of the foot pad thickness with a caliper . At day 60 after experimental infection , mice were sacrificed and foot pads aseptically removed for enumeration of M . ulcerans bacteria or histopathology . Draining inguinal lymph nodes of designated animals were removed and fixed in formalin as well . All M . ulcerans infection experiments were conducted under BSL3 conditions . Mouse feet designated for enumeration of M . ulcerans bacteria were immediately removed above the ankle after euthanasia , shortly dipped into 70% ethanol , then dried under the laminar flow , cut in 4 pieces with a scalpel and transferred to 750 μl of Bac/T medium in reinforced hard tissue grinding tubes ( MK28-R , Precellys ) . Tissue homogenization was performed with a Precellys 24-Dual tissue homogenizer ( 3 x 20 s at 5000 rpm with 30 s break ) , the lysate was transferred to a new tube and the lysis tube still containing tissue remains was refilled with 750 μl of Bac/T medium . The remains were homogenized a second time and the two lysates pooled . Mouse feet designated for histopathological analysis were removed above the ankle and immediately transferred to 10% neutral-buffered Formalin solution ( approx . 4% formaldehyde , Sigma ) for fixation during 24 hours at room temperature . Subsequently , the feet were decalcified in 0 . 6 M EDTA and 0 . 25 M citric acid for 12 days at 37°C and transferred to 70% ethanol for storage and transport . The samples were dehydrated and embedded into paraffin . 5 μm thin sections were cut , deparaffinised , rehydrated , and stained with Haematoxylin/Eosin ( HE , Sigma , J . T . Baker ) or Ziehl-Neelsen/Methylene blue ( ZN , Sigma ) according to WHO standard protocols [44] . Stained sections were mounted with Eukitt mounting medium ( Fluka ) . Pictures were taken with a Leica DM2500B microscope or with an Aperio scanner . Differences of bacterial load in infected foot pads were statistically analysed by the Mann-Whitney test using Graph Pad Prism ( Version 6 . 03 ) . Results of cytokine production were subjected to log10 transformation and subsequently analysed with SAS software ( SAS Institute , Cary , USA , release 9 . 3 ) using linear mixed models adjusted for random effects . Image processing and picture panel assembly was performed with Photoshop software ( Adobe Photoshop CS6 Extended , version 13 . 0 . 1 ) .
Candidate vaccines were generated by replacing the VSV glycoprotein G gene with the M . ulcerans genes MUL2232 or MUL3720 ( Fig 1A ) . In order to ease virus detection and titration , the coding sequence for the enhanced green fluorescent protein ( eGFP ) was inserted as an additional transcription cassette downstream of the M . ulcerans genes . The recombinant viruses VSV*ΔG ( MUL2232 ) and VSV*ΔG ( MUL3720 ) as well as the control virus VSV*ΔG , which only contained the eGFP gene in place of the VSV surface glycoprotein G ( Fig 1A ) , were produced and propagated in the helper cell line BHK-G43 providing the VSV G protein in trans . As expected from other studies with similar constructs , the trans-complemented particles were able to infect a variety of different mammalian cell lines but were unable to release progeny viruses [27 , 45] . We thus refer to them as to virus replicon particles ( VRPs ) . In order to study the expression of MUL2232 and MUL3720 in infected cells , BHK-21 cells were infected with the three different VRPs generated and analysed by using MUL2232 and MUL3720-specific mouse monoclonal antibodies ( mAbs ) in immunofluorescence microscopy . Both antigens accumulated in the cytosol of the infected BHK-21 cells ( Fig 1B1 and 1B2 ) . While in the immunofluorescence analysis MUL2232 appeared to be expressed at lower levels than MUL3720 , staining intensities in Western blotting analyses were comparable for both proteins ( Fig 1C ) . MUL2232 was mainly detected in the soluble fraction of infected L929 fibroblasts and to a smaller part in the insoluble fraction ( Fig 1C1 ) while MUL3720 was found in the soluble fraction only ( Fig 1C2 ) . Both proteins co-migrated with the corresponding proteins in M . ulcerans lysate according to the predicted molecular mass ( Fig 1C ) . To characterize the immune responses elicited by the recombinant VRPs , we compared different immunization regimens in BALB/c mice . Humoral immune responses were assessed by ELISA on immobilized recombinant protein . Sub-cutaneous ( s . c . ) administration of 107 VRPs did not lead to measureable antibody responses against the target antigens ( Fig 2A and 2B ) . However , when VSV*ΔG ( MUL2232 ) was given i . m . , Immunoglobulin ( Ig ) G antibodies were elicited in all immunized animals . In contrast , only some mice produced antibodies following i . m . immunization with VSV*ΔG ( MUL3720 ) ( Fig 2A and 2B ) . Therefore induction of antibodies solely by i . m . immunization with VSV*ΔG ( MUL3720 ) was no longer pursued in subsequent experiments . Independently of the route of administration , both VRPs primed the immune system , as demonstrated by the fast humoral immune response to an adjuvant-free booster immunization ( s . c . ) with 30 μg of the corresponding recombinant protein ( Fig 2A and 2B ) . Six days after the booster injection , antibody titers were generally higher in mice primed i . m . than in mice primed s . c . ( Fig 2A and 2B ) . Therefore , only the i . m . route was employed for VRP administration in subsequent experiments . In addition , a dose of 107 VRPs per immunization was generally used . In a next step , we assessed the potential of a shorter prime-boost immunization strategy by immunizing the animals only once i . m . with VSV*ΔG ( MUL2232 ) or VSV*ΔG ( MUL3720 ) and three weeks later with 30 μg of rMUL2232 or rMUL3720 via the subcutaneous route in the absence of adjuvant . Five days after the rMUL2232 boost , the ELISA IgG titers were only marginally higher than those observed prior to the boost , but increased further in the subsequent two weeks ( Fig 3A ) . On the other hand , administration of rMUL3720 led to high antibody titers already eight days after the boost ( Fig 3C ) . Importantly , the elicited antibodies were not only cross-reactive with the recombinant proteins produced in E . coli but also with the target proteins expressed by M . ulcerans , as demonstrated by Western blotting on M . ulcerans lysate ( Fig 3B and 3D ) . To evaluate whether this immunization regimen would also elicit antigen-specific cellular immune responses , antigen-specific cytokine secretion of spleen cells was studied in vitro after primary immunization of mice with VSV*ΔG ( MUL2232 ) and a second immunization with either rMUL2232 or VSV*ΔG ( MUL2232 ) . Three weeks after the second immunization , spleen cells were stimulated with either rMUL2232 or the unrelated rMUL3720 as control . Stimulation with Concanavalin A ( ConA ) served as positive control for confirming the viability of the cultured spleen cells . The stimulated cell culture supernatants were analysed by ELISA for the production of the TH1 cytokine Interferon gamma ( IFNγ ) ( Fig 4A ) , the pleiotropic [46] cytokine Interleukin 2 ( IL-2 ) ( Fig 4B ) and the TH2 cytokine Interleukin 10 ( IL-10 ) [47] ( Fig 4C ) . Overall , no marked difference in terms of cytokine production was observed between the two immunization schedules . In both cases all three cytokines were produced in significantly higher amounts when cultured spleen cells were stimulated with the corresponding rMUL2232 antigen . In contrast , stimulation with the unrelated antigen rMUL3720 or mock stimulation had no significant effect ( Fig 4 ) . The most pronounced response was found for IL-10 ( Fig 4C ) , indicating a slight polarization towards a TH2 type response . The same experimental setup with VSV*ΔG ( MUL3720 ) was not successful due to technical problems . In a last step , we explored the protective efficacy of the VRP-based immunization in an experimental M . ulcerans infection model . Groups of six BALB/c mice were immunized according to the VSV*ΔG ( MUL2232 ) /rMUL2232 prime-boost regimen or two times i . m . with VSV*ΔG ( MUL2232 ) . Respective control groups were immunized either with the control VRP VSV*ΔG and an rMUL2232 boost ( prime-boost control ) or twice with VSV*ΔG ( Fig 5 ) . Three weeks after the last immunization , the left hind foot pad of the mice was infected with 8 . 4 x 103 M . ulcerans bacilli . The slowly progressing infection was followed by weekly measurements of the thickness of the infected foot pads with a caliper ( Fig 5A ) . To determine the bacterial multiplication , mice were euthanized at day 60 after infection and foot pads either processed for histopathological analysis or lysed for enumeration of M . ulcerans by standard CFU plating . Quantification was additionally performed by an adapted qPCR method , suitable for the detection of bacterial proliferation in mouse foot pads over the course of the infection ( Fig 5C ) . No differences in the kinetics of footpad swelling were observed between immunized mice and the respective control animals ( Fig 5A ) . However , 60 days post infection the number of CFU per foot pad was slightly but significantly lower in animals immunized with the VRP prime-protein boost regimen as compared to the prime-boost control immunized animals ( Fig 5B ) . Quantification of M . ulcerans DNA in the footpads by insertion sequence ( IS ) 2404 specific qPCR yielded similar results , i . e . a slight but significant reduction in bacterial multiplication caused by the VRP prime-protein boost immunization , but not by two subsequent immunization with VRPs ( Fig 5B and 5D ) . The same experiment with MUL3720 VRPs did not result in any difference between immunized and control groups ( S1 Fig ) . A histopathological analysis of representative foot pads confirmed the findings of a slight reduction of bacterial burden in the VSV prime-protein boost immunized animals as compared to the prime-boost control animals . The non-infected right foot pads served as control and appeared completely normal ( Fig 6A1 ) with intact muscle tissue ( Fig 6A2 ) and no apparent oedematous changes ( Fig 6A3 ) . In comparison , the infected left foot pads of control mice showed typical histopathological signs of BU with strong oedema ( Fig 6B1 and 6B4 ) , necrotic sole of foot ( Fig 6B2 ) , and inflammatory infiltration and extensive haemorrhages all over the foot pad ( Fig 6B3 ) . Ziehl-Neelsen/Methylene blue ( ZN ) staining revealed large clumps of acid fast bacilli ( AFB ) ( Fig 6C1 ) , not only located where they were initially injected , but also more towards the heel of the foot ( Fig 6C2 ) and in the oedematous tissue in the upper part of the foot pad ( Fig 6C5 ) . AFB were associated with remains of infiltrating immune cells ( Fig 6C3 ) and were also found as fibrous structures in completely necrotic tissue ( Fig 6C4 ) . In animals that received a VRP prime-protein boost immunization , a trend towards a reduction in the number of AFB clusters was observed ( Fig 6D1 ) . Furthermore , AFB appeared to be more often in close contact with infiltrating cells or were found intracellular ( Fig 6D3 ) .
Despite substantial control efforts and improvements in treatment and diagnosis , the socioeconomic impact of BU on affected communities remains high . Long treatment regimens with daily i . m . injections in rural settings of West Africa , late reporting to health facilities , scarring and resulting permanent disabilities create high morbidity that could be prevented by a vaccine against M . ulcerans . Although no such vaccine is available at the moment , reports on self-healing in patients with early stages of the disease [48 , 49] as well as sero-epidemiological studies in BU-endemic countries [32 , 50] indicate that the human organism is in principle capable of inducing a protective immune responses against BU . Furthermore , protein subunit vaccination approaches have demonstrated partial protection in a BU mouse infection model [21 , 22 , 51] . In the present study , we used a viral replicon particle system for delivering M . ulcerans protein antigens to the immune system . Immunization with a high number of VRPs induced M . ulcerans specific antibody responses with i . m . application being superior to s . c . application . Our VRP prime-recombinant protein boost regimen not only induced a humoral but also a cellular immune response . This is in line with previous work showing that VRPs are able to trigger humoral as well as cellular immune responses [52] . The bacterial proteins MUL2232 and MUL3720 were chosen as vaccine antigens because previous work suggested that these antigens are not only highly immunogenic but also associated with the outer surface of M . ulcerans [32 , 34 , 35] . Therefore , the strong humoral immune response observed following immunization of mice was expected to allow antibody-mediated opsonization of the bacteria with consequent enhanced phagocytosis , complement activation , or antibody-dependent cellular cytotoxicity . While we did not observe vaccination-induced reductions in foot pad swelling in our murine M . ulcerans infection model , assessment of the bacterial load by qPCR as well as CFU plating revealed a slight , but significant reduction in bacterial multiplication in VRP prime-protein boost immunized mice . While measurement of foot pad swelling over time is a parameter that can be followed without euthanizing mice [53 , 54] , our results illustrate that foot pad swelling not necessarily reflects the extent of bacterial proliferation . Despite a slight inhibitory effect on bacterial proliferation , immunization with only one target antigen in a VRP prime-protein boost regimen was not sufficient to confer full protection against the experimental infection . Several factors , like the choice or number of antigens included into the VRPs , could have let to this negative result . Since it is not known , which immune effector functions are relevant for protection against M . ulcerans disease [15] , it is not clear whether lack of a strong TH1-polarisation is of major relevance for the failure to achieve a strong protective efficacy with the immunization regimens tested . One of the advantages of the replicon system lies in the ability of VSV to tolerate incorporation of long stretches of foreign DNA into its genome [25 , 26] . As a modular system , it offers the possibility to design a multivalent subunit vaccine by combining several M . ulcerans proteins in one replicon . Furthermore the versatility of the system allows to engineer the location of the expressed protein in the infected cell , or to target specific immune cells by including genes for co-stimulatory molecules or receptors to be expressed on the surface of the infected cells [25] . Here we have demonstrated that RNA replicon particles are a very good delivery system for mycobacterial antigens , which is in particular encouraging future development of VRP-based multivalent subunit vaccines . | Infection with Mycobacterium ulcerans can lead to a slow progressing , ulcerative disease of the skin and underlying soft tissue called Buruli ulcer . The disease is most prevalent in rural African communities with limited access to health care facilities . The most efficient means to prevent the disease , a vaccine against Buruli ulcer is not available to date . In the present study we investigated the immunogenicity and protective potential of a single cycle virus system expressing the two M . ulcerans antigens MUL2232 and MUL3720 . Immunization of mice with those vesicular stomatitis virus replicon particles led to the induction of humoral as well as cellular immune responses in the immunized animals . Subsequent challenge experiments in a mouse model of M . ulcerans infection demonstrated only a limited reduction of bacterial burden in mice immunized with a prime-boost approach with MUL2232 . Most probably , a vaccine formulation with only one antigen will not be able to provide protection against Buruli ulcer in humans . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | Use of Recombinant Virus Replicon Particles for Vaccination against Mycobacterium ulcerans Disease |
Chikungunya is a mosquito-borne viral infection of humans that previously was confined to regions in central Africa . However , during this century , the virus has shown surprising potential for geographic expansion as it invaded other countries including more temperate regions . With no vaccine and no specific treatment , the main control strategy for Chikungunya remains preventive control of mosquito populations . In consideration for the risk of Chikungunya introduction to the US , we developed a model for disease introduction based on virus introduction by one individual . Our study combines a climate-based mosquito population dynamics stochastic model with an epidemiological model to identify temporal windows that have epidemic risk . We ran this model with temperature data from different locations to study the geographic sensitivity of epidemic potential . We found that in locations with marked seasonal variation in temperature there also was a season of epidemic risk matching the period of the year in which mosquito populations survive and grow . In these locations controlling mosquito population sizes might be an efficient strategy . But , in other locations where the temperature supports mosquito development all year the epidemic risk is high and ( practically ) constant . In these locations , mosquito population control alone might not be an efficient disease control strategy and other approaches should be implemented to complement it . Our results strongly suggest that , in the event of an introduction and establishment of Chikungunya in the US , endemic and epidemic regions would emerge initially , primarily defined by environmental factors controlling annual mosquito population cycles . These regions should be identified to plan different intervention measures . In addition , reducing vector: human ratios can lower the probability and magnitude of outbreaks for regions with strong seasonal temperature patterns . This is the first model to consider Chikungunya risk in the US and can be applied to other vector borne diseases .
Chikungunya fever ( CHIKF ) is a mosquito-borne viral infection first isolated in Tanzania in 1953 [1] , [2] . CHIKF is caused by Chikungunya virus ( CHIKV ) , an alphavirus with different variants endemic to countries in Africa and Southeast Asia [3] , [4] . Illness caused by CHIKV is usually diagnosed based on symptoms , and often confused with dengue given some overlapping symptomology [5] . One symptom specific for Chikungunya is a debilitating and prolonged joint pain , affecting the peripheral small joints [6] , that appears in conjunction with other nonspecific symptoms including fever , severe joint pain , muscle pain , headache , nausea , fatigue and occasionally rash [7]–[11] . CHIKF-related mortality is rare , but can occur , often in patients with other health conditions [1] , [9] , [11] , [12] . There is no specific treatment for the disease; consequently , treatment is focused on symptomatic care and mosquito vector control . No vaccines are currently available for prevention of CHIKV infection , although vaccine candidates currently are under investigation [13] . The onset of the symptoms occurs after an intrinsic incubation period in the human host of approximately 4 days post infection [1] , [8]–[11] , and viremia in infective individuals usually persists for a period of approximately 7 days [1] , [3] , [14] . During this period , mosquitoes may be infected with CHIKV when feeding on viremic hosts . After the acute stage of infection , severe joint pain may persist for long periods in affected individuals . Some people show mild to no overt signs of illness . Seroprevalence studies have demonstrated that 25% of infected individuals have mild symptoms or were asymptomatic [3] . Laboratory studies have demonstrated that CHIKV disseminates to the salivary glands in competent mosquitoes quickly , within 2 days ( range 1–14 days ) post-infection [15] . Once infectious , mosquito vectors are thought to remain infectious for their lifetime . Prior to 2000 , Aedes aegypti was the most important vector of CHIKV [6] , with Ae . albopictus considered a secondary vector [16] . Within the last decade several epidemics of CHIKF were reported . In 2005–2006 , a severe epidemic occurred in Réunion Island [3] , followed shortly after by epidemics in India [3] , Southeast Asia [17] and other Indian Ocean islands [1] , [18] . Sampling during the Reunion Island epidemic provided evidence for the role of Ae . albopictus as the main vector [19]–[21] . Sequencing of the envelope protein of the Reunion Island CHIKV isolates ( CHIKV 226OPY1 ) showed that the outbreak was caused by a new variant of the virus with a single adaptive mutation . This single amino acid change from Alanine to Valine in the E1 glycoprotein at position 226 [22] increased infection and dissemination in Ae . albopictus [23] . Another epidemic of CHIKV OPY1 genotype occurred in Italy in 2007 [24] , [25] . Epidemiological studies strongly implicate introduction of the virus from India by a traveler [21] , [26] . This unexpected outbreak is a striking example of disease introduction in an area recently colonized by Ae . albopictus [15] , [21] , [27] . Moreover , it highlights the fact that CHIKV outbreaks can originate from just one infective individual even in temperate areas with seasonal transmission of arboviruses [25] , [19] . The Asian tiger mosquito , Ae . albopictus , is an invasive urban mosquito native to East Asia [28] . It is a diurnally active species and thought to have a broader host range than Ae . aegypti , although in some regions it can be highly anthropophagic when human hosts are readily available [29] . In the past couple of decades this species has invaded many countries through the transport of goods , especially used tires , and increasing international travel [30]–[32] . Native to tropical regions of Asia , Ae . albopictus has successfully adapted to cooler climates within the 10°C isotherm [33] . Thus , eggs from strains in temperate regions are moderately tolerant to cold and can even tolerate short durations of freezing temperatures [34] , [35] . Female Ae . albopictus lay eggs in human-made and natural containers just above the waterline . Reported flight range of this species is typically less than 200 m [36] . Several laboratory studies on Ae . albopictus vector competence for the CHIKV LR 226OPY1 epidemic strain have now been conducted reporting a range of dissemination rates from 26–100% in various geographic strains of the vector [20] , [37]–[40] . A recent laboratory study , using salivary gland infection as a proxy for transmission , demonstrated transmission rates from above 67% for Galveston , TX strain , Florida strain , and a New Jersey/New York metropolitan strain ( Harrington , Sanchez-Vargas and Olson , unpublished data ) . Concerns for the role of Ae . albopictus as an active disease vector have been raised since its introduction and in the USA [41] , [42] . Since introduction , Ae . albopictus has become established in 26 states primarily in Southeast , gulf coast and mid-Atlantic regions . The species is currently expanding its range through New Jersey and into New York State [43] , [44] . Given the establishment of Ae . albopictus in these regions , travel related introductions of several arboviruses suggest a potential increase in epidemic risk for the USA . High numbers of CHIKF cases are periodically reported in US travelers [45] , [46] . However , local CHIKV outbreaks have not been detected in the US to date , presumably because of the asynchrony between the arrival of the exposed individuals and the abundance of the vectors [45] . In this study , we explicitly evaluated the risk of epidemic events by simulating the introduction of Chikungunya virus into three naïve US populations . Assuming established mosquito populations in each area , we introduced one exposed individual to evaluate the epidemic potential size of an outbreak , taking into account the population dynamics of the vector and its susceptibility to temperature regimes . We predicted low epidemic risk for disease introduction during periods of low vector abundance and high epidemic risk for certain critical periods that show increasing , or high , vector abundance . These results provide valuable additional information not only for early warning systems but also for the implementation of intervention strategies with the goal of reducing vector populations or human risk of exposure .
To study the dynamics of the introduction of CHIKV in an immunologically naive population we constructed a model with demographic stochasticity for mosquitoes and humans ( Figure 1 , see Material-S1 for model equations ) . Using a classical approach , the human host population was divided into susceptible ( S ) , exposed ( E ) , symptomatic infective ( IS ) , asymptomatic infective ( IA ) and recovered ( R ) classes . Analogously , the adult mosquito population was divided into susceptible ( S ) , exposed ( E ) and infected ( I ) classes . In addition , we considered the immature stages of mosquito population , including mosquito eggs ( G ) , larvae and pupae ( L ) and eggs undergoing diapause ( D ) . Vital mosquito rates in this model were temperature dependent ( Table S1 and Figure 2 ) . Density-dependent effects were added to both mosquito and human populations . It is worthwhile to note that although the net effect of having density-dependent terms in the model is to avoid uncontrolled population growth , they represent a broad range of factors from larval overcrowding effects to human behavior . The functional forms of the temperature forcing on the parameters for the dynamic of the vector population are presented in Figure 2 ( The mathematical forms are presented in Material S1 ) . Two temperature thresholds ( TsD and TeD ) were used to determine the diapause state . Eggs entered diapause ( i . e . , arrested development ) when temperature was below TsD , and eggs do not undergo diapause for temperatures above TeD . The proportion of eggs undergoing ( avoiding ) the diapause state linearly decreased ( increased ) with increasing temperature for environmental conditions between TsD and TeD . Although the determination of diapause periods follows a complex combination of factors — including temperature and photoperiod— temperature was used as a proxy for such combination in here . Dependence on temperature of both egg survival and development time was fitted to experimental data ( Harrington , unpublished data ) and reports from the literature [47] , [48] . Similarly , experimental and reported data were used for the fitting of temperature influence on larvae survival and development time [Harrington , unpublished data , [47] , [48] . Adult longevity dependence on temperature was fitted to experimental data using the general assumption that longevity declines linearly with temperatures under 10°C [49] . Although there is no experimental support reported in the literature , CHIKV extrinsic incubation period ( EIP ) was assumed to be reduced with increasing temperature ( up to 32°C ) as with other arboviruses such as DENV [50] . The reported minimum extrinsic incubation period for CHIKV in several studies is 2 days [15] . Hence , we modeled EIP as a linear function with a minimum duration of 1 . 5 days at 32°C and a maximum duration of 4 days at 10°C [15] , [51] . For the purposes of the current model we assumed transmission rates based on early experimental work [52] . Daily temperatures were calculated by applying a spline interpolation to the monthly mean temperature data of the last decade obtained from the Intergovernmental Panel on Climate Change ( http://www . ipcc-data . org ) ( Figure S1 ) . The model was run using temperature data from different locations to evaluate variability on epidemic risk with temperature patterns . Here , we present the results for three major US ports of entry that encompass a wide seasonal variation in temperature: New York , Atlanta , and Miami . Population sizes and carrying capacity parameters for human populations were estimated using the city size data reported in the last census ( http://www . census . gov ) . Mosquito population carrying capacity was estimated assuming a maximum number of vectors per host . We ran independent simulations for the three ports of entry changing this ratio ( we present here the results for 0 . 5 , 1 and 3 mosquitoes per human ) . The model was run for five years . Initial population sizes in the model were selected according to the expected equilibrium values . During the first year of simulation there was no disease present in the model and therefore both human and mosquito populations drifted to their respective equilibria . CHIKV was introduced into the model during the second year of simulation via one exposed individual , and the simulation was run until the end of the fifth year . We calculated the final number of infective individuals , the number of infected at the outbreak peak , and the time to reach the outbreak peak from the day of introduction for each one of 1000 Monte Carlo simulations . We ran the simulations systematically varying the day of introduction of the disease from January 1st to December 31st , which allowed us to express the outbreak probability as a function of the introduction day ( Figure 3 ) . Here , the probability of outbreak was defined as the frequency of cases where the chain of infection was functional ( i . e . , the number of infected individuals was bigger than 1 ) . In addition , we calculated the risk of an outbreak as the mean of the final epidemic size ( summation of all infected individuals ) over the average susceptible population size ( Figure 4 ) . These simulations were replicated not only varying the ratio of mosquitoes to humans ( considering values of 0 . 5 , 1 and 3 for that ratio ) but also reducing the mosquito feeding pattern for human blood from 100% to 25% .
Figure 3 displays the probability of outbreak for the different locations as function of the day of introduction and 0 . 5 mosquitoes to human ratio for each location , and a 100% human meal preference ( See Material S1 for results with other vector to host ratios and human blood meal patterns ) . The probability of an outbreak ( defined as at least one successful transmission event to a human ) for New York shows a peak around 38% for a CHIKV introduction in August , and is over 30% during the interval from August 6th to September 11th . In addition , there is a significant probability of outbreak after an introduction on June 15th and up to December . The probability of having an outbreak late in November is very small and it is a consequence of using mean monthly temperature data as a basis for the temperature patterns . Outbreaks also were seasonal for Atlanta , with no significant probability of outbreaks after introductions between January 12th and April 9th . Moreover , in Atlanta , the probability of outbreak was greater than 30% for a longer period , extending from June 6th to September 26th , with peak values similar to those in New York . In contrast , for Miami chances of a CHIKV outbreak were significant after an introduction at any time during the year . Our model only demonstrated the occurrence of at least one successful transmission event , however , the maximum prevalence reached for those outbreaks is likely to be a more important parameter ( Figure 4 ) . Consequently , we explored the peak infection rate with our model and found that it varies significantly between locations and also with the ratio of vectors to hosts and human feeding patterns ( see Tables S2 and S3 for results with other vector to host ratios and meal preferences ) . When we set the ratio of mosquitoes to one human at 0 . 5 and human feeding rates at 100% , peak infection in New York was very small ( 0 . 0002% ) , peaking at the beginning of the high probability outbreak period . A similar pattern , but with higher prevalence values for CHIK in humans ( 0 . 1381% ) was observed for Atlanta . In Miami , however a high mean prevalence for CHIK in humans ( 25 . 0187% ) was observed throughout the year . Additionally , we calculated the number of days from pathogen introduction until the peak prevalence ( Table 1 ) . These calculations reveal that ( in general ) , when mosquito feeding preference is set to be only from humans , the time to peak prevalence is longer than when feeding preferences are broad . Thus , when human blood feeding is low , the epidemic peak usually occurs shortly after introduction because the chain of transmission could be easily interrupted . However , when human blood feeding is set to 100% , the epidemic peaks occurs within 20 days for cities in cooler climates ( i . e . New York ) and approximately one to three months for warmer locations . In these warmer locations several secondary cases are expected to follow the index case , and the stochastic interruptions of the chain of infection can only slow the development of the epidemic instead of stopping it .
Since the occurrence of the CHIKF outbreak in Italy in 2007 , the risk of similar outbreaks in the United States and other temperate countries has become a public health concern [42] . Our models , based on the introduction of one exposed individual , show that the probability of an outbreak in any of the three chosen locations varies by geographic location . As expected , for areas like the New York metropolitan region and Atlanta , that display a strong seasonality in temperature and therefore in mosquito abundance , this risk is bounded to the summer time and low prevalence levels are predicted . In locations with temperature patterns similar to those in Miami , which allow for year-round mosquito activity , the risk for outbreak is not bounded seasonally . In addition , predicted prevalence levels are higher for Miami because usually the chain of infection is not completely interrupted . These models also show that the proportion of people affected in an outbreak is reduced dramatically with increasing latitude ( Figure 4 ) . Furthermore , the models suggest the replenishment of susceptible individuals is not enough to create endemic foci , however this result could change when using more realistic models . Our model outputs display higher sensitivity to parameters controlling the proportion of blood meals from humans than the vector: host ratio . Nevertheless , the increased ratio of mosquitoes to humans led to a two-fold increase in the probability of outbreak at all locations . This result highlights the importance of vector control to reduce both the risk of outbreaks and the proportion of infected individuals . It is important to note that , in this modeling approach , both parameters may be interpreted as proxies for a reduction of human exposure to mosquitoes . Hence , our results confirm the relevance of public campaigns advising residents to control mosquitoes at home and take precautions to avoid mosquito exposure to reduce disease outbreaks . The time between CHIKV introduction and peak outbreaks revealed that , for locations with temperature patterns similar to those of Miami where mosquito populations may not undergo diapause , CHIKV infections might circulate at low levels for several months until reaching dramatic proportions . Early detection of cases in these regions will be important to reduce the magnitude of an outbreak . However , in locations such as New York and Atlanta , a critical temporal window for interventions could be identified and intervention during such periods may be enough to significantly reduce the probability of an outbreak . It is clear that these predictive models are highly sensitive to temperature patterns that govern mosquito population dynamics , and could be improved by using non-averaged temperature data ( i . e . sampling from the distribution of temperatures ) , and including other environmental factors such as rainfall and photoperiod that can have a significant influence on vector populations . In addition , reduction of individual exposure ( only modeled as a reduction in human feeding patterns here ) should be considered in order to have more accurate predictions . This modeling approach highlights the fact that a better understanding of epidemiological dynamics will require further studies on both biological and non-biological processes . Especially important will be: ( 1 ) further studies on diapause , abundance and feeding biology of Ae . albopictus , ( 2 ) the inclusion of multiple disease introduction events , either simultaneously of temporally spread , and ( 3 ) a better understanding of the evolution and plasticity of both pathogen and vector . Our results strongly suggest that , in the event of an introduction and establishment of CHIKV in the United States , endemic and epidemic regions would emerge initially , mainly defined by environmental factors controlling annual mosquito population cycles . These regions should be identified in order to plan different intervention measures . In addition , reducing mosquito population sizes ( and , consequently , reducing vector: human ratios ) can lower the probability and magnitude of outbreaks mainly for regions with strongly marked seasonal temperature patterns . Typical control strategies for vector borne diseases are: ( 1 ) reduction of vector population , ( 2 ) reduction of host exposure to infectious mosquito bites , and ( 3 ) isolation of infective hosts . This model also allows for evaluation of the effects of changes in the mosquito feeding patterns . Simulation results suggest that a reduction of vector population and human exposure could be very effective for a reduction of both the risk of an outbreak and the population at risk . The results presented here simulating significant CHIK outbreaks in the US were based on a conservative approach of one exposed individual introduced to a region [45] . Given CHIKV infections in returning US travelers [45] , [46] and the low numbers of infected individuals needed to spark an outbreak , we conclude that US health systems should be vigilant . | Chikungunya fever is a mosquito-borne viral infection showing a surprising potential for geographic expansion . Similar to other tropical infectious diseases having no vaccine and no specific treatment , the main control strategy for Chikungunya remains reduction of mosquito population size . We developed a model for disease introduction that combines a climate based mosquito population dynamics stochastic model with an epidemiological model in order to identify temporal windows during which disease introduction through one exposed individual might compromise the health status of the entire human population . We ran this model with temperature data from different locations showing the geographic sensitivity of this risk . The identification of temporal windows with epidemic risk at different spatial locations is key to guiding mosquito population control campaigns . Locations with marked seasonal variation also have a season with high epidemic risk matching the period in which mosquito populations survive and grow , therefore controlling mosquito population sizes might be an optimal strategy in those areas . However , locations with other temperature patterns may need additional control strategies to avoid epidemics . To our knowledge , this is the first model to explore Chikungunya introduction in the USA . Our modeling approach can be used for other vector borne diseases and can be expanded to compare the outcome with different control strategies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"public",
"health",
"and",
"epidemiology",
"infectious",
"disease",
"epidemiology",
"theoretical",
"ecology",
"population",
"biology",
"infectious",
"disease",
"control",
"infectious",
"diseases",
"environmental",
"epidemiology",
"epidemiology",
"biology",
"vectors",
"and",
"hosts",
"infectious",
"disease",
"modeling",
"mosquitoes",
"ecology",
"viral",
"diseases"
]
| 2012 | Modeling Dynamic Introduction of Chikungunya Virus in the United States |
Individual susceptibility to HIV is heterogeneous , but the biological mechanisms explaining differences are incompletely understood . We hypothesized that penile inflammation may increase HIV susceptibility in men by recruiting permissive CD4 T cells , and that male circumcision may decrease HIV susceptibility in part by reducing genital inflammation . We used multi-array technology to measure levels of seven cytokines in coronal sulcus ( penile ) swabs collected longitudinally from initially uncircumcised men enrolled in a randomized trial of circumcision in Rakai , Uganda . Coronal sulcus cytokine levels were compared between men who acquired HIV and controls who remained seronegative . Cytokines were also compared within men before and after circumcision , and correlated with CD4 T cells subsets in foreskin tissue . HIV acquisition was associated with detectable coronal sulcus Interleukin-8 ( IL-8 aOR 2 . 26 , 95%CI 1 . 04–6 . 40 ) and Monokine Induced by γ-interferon ( MIG aOR 2 . 72 , 95%CI 1 . 15–8 . 06 ) at the visit prior to seroconversion , and the odds of seroconversion increased with detection of multiple cytokines . Coronal sulcus chemokine levels were not correlated with those in the vagina of a man’s female sex partner . The detection of IL-8 in swabs was significantly reduced 6 months after circumcision ( PRR 0 . 59 , 95%CI 0 . 44–0 . 87 ) , and continued to decline for at least two years ( PRR 0 . 29 , 95%CI 0 . 16–0 . 54 ) . Finally , prepuce IL-8 correlated with increased HIV target cell density in foreskin tissues , including highly susceptible CD4 T cells subsets , as well as with tissue neutrophil density . Together , these data suggest that penile inflammation increases HIV susceptibility and is reduced by circumcision .
Two million individuals acquired HIV-1 ( HIV ) in 2014 , contributing to the nearly 37 million living with this still incurable infection [1] . While most individuals acquired the virus through heterosexual sex [2] , the per act risk of female-to-male transmission is generally low ( less than 1/250 per coital act in low income countries [3] ) . This risk is also highly variable , and is dependent on factors in both the infected and uninfected partner [4] . Susceptibility of an uninfected male partner has been epidemiologically linked to younger age [5 , 6] , race [7] , genital co-infections [8] , and lack of male circumcision [9 , 10] . However , the biological mechanisms by which these parameters alter HIV susceptibility remain incompletely understood . Mucosal inflammation and immune activation are hypothesized to enhance HIV susceptibility . In the genital mucosa , CD4 T cells expressing CCR5 are the primary targets of HIV [11–14] , with potential transport and amplification by local dendritic cell subsets [15] . Thus , if inflammation leads to the recruitment of CCR5+ CD4 T cells , it will provide additional target cells for HIV . HIV also preferentially infects and replicates in activated CD4 T cells [16–21] , and so augmented immune activation may also facilitate the establishment of productive mucosal infection . Contribution of mucosal inflammation to genital HIV susceptibility is consistent with data from female rhesus macaques , where pro-inflammatory cytokines promote the recruitment and activation of CD4 T cells in the vaginal mucosa [13] , and the number of CCR5+ CD4 T cells at the site of mucosal challenge dictates the likelihood of subsequent productive SIV infection [22] . Furthermore , observational studies in South African women have linked pro-inflammatory genital cytokines to HIV acquisition [23] and increased CD4 T cells in the cervical mucosa [24] . While there are no similar data from men , asymptomatic herpes simplex virus type-2 ( HSV-2 ) infection is associated with a 3-fold increased risk of HIV acquisition in heterosexual uncircumcised men [25] , perhaps due to increased CCR5+ CD4 T cells in foreskin tissue [26 , 27] . Randomized clinical trials have conclusively shown that male circumcision reduces HIV susceptibility in heterosexual men [28–30] , but the biological mechanisms underlying this protection remain incompletely understood . One hypothesis is that circumcision reduces genital inflammation and immune activation , either through the prevention of viral STIs [31] , the reduction of inflammatory anaerobic bacteria [32] , or through other mechanisms yet to be defined , which in turn this reduces the density of potential target cells for HIV . This hypothesis is supported by ex vivo experiments demonstrating that the inner aspect of the foreskin has an increased density of HIV target cells [33–35] and more efficient virus transfer from Langerhans cells to local CD4 T cells [15] than the outer aspect , which is contiguous with the shaft skin that remains after circumcision . These observations suggest that the intact foreskin constitutes an immunologically activated tissue milieu that promotes target cell recruitment and dendritic cell maturation [36–38] . We hypothesized that elevated levels of pro-inflammatory penile cytokines would be associated with HIV acquisition in uncircumcised men and with an increased density of HIV target cells in foreskin tissue , and that cytokine levels would be reduced by circumcision . To test these hypotheses , we performed a case-control study of coronal sulcus cytokines and HIV acquisition among men who participated in a randomized controlled trial ( RCT ) of male circumcision in Rakai , Uganda [29] . We then examined whether these inflammatory cytokines declined after circumcision in a subset of men who were enrolled in the trial but who did not acquire HIV . Finally , we used samples from a cross-sectional study of men undergoing elective circumcision [39] to assess the correlation between prepuce cytokine levels and foreskin HIV target cell density .
To assess the relationship of coronal sulcus cytokines with seroconversion , we performed a nested case-control study comparing men who acquired HIV during the Rakai RCT of circumcision ( n = 60 , cases ) to men who remained persistently seronegative ( n = 120 , controls ) . All men in this analysis were randomized to receive delayed circumcision and remained uncircumcised throughout the trial . Participant demographics are presented in Table 1 . HIV seroconversion was associated with occupation , marital status , number of sex partners , condom use , alcohol consumption , and self-reported genital STI symptoms ( genital ulcer , genital warts , urethral discharge ) , as previously reported [40] . All cytokines examined were detected in coronal sulcus swabs , although many were detected infrequently . IL-8 was most common , detected in 60% of coronal sulcus swabs ( concentration range >1 . 5–7405 . 7pg/ml in swabs suspended in 1ml transport medium ) , followed by MIG ( range >0 . 3–6 . 9pg/ml ) , which was detected in 25% of swabs . Other cytokines , ( GM-CSF , MCP-1 , MIP3α , IL-1a and RANTES ) were detected infrequently ( <10% of participants , Table 2 ) . Cytokine detection was not associated with sexual behavior or demographic factors ( S1 and S2 Tables ) , but was associated with self-reported STI symptoms ( genital ulcer , genital warts , urethral discharge ) . Men who acquired HIV were more likely to have detectable levels of the chemoattractant cytokines IL-8 ( aOR 2 . 58 , 95% CI: 1 . 40–6 . 40 ) and MIG ( aOR 3 . 05 , 95% CI: 1 . 15–8 . 06 ) at the visit prior to seroconversion ( Table 2 ) . The increased odds of HIV acquisition did not change after adjusting for covariates associated with either the detection of cytokines ( S1 and S2 Tables ) or seroconversion ( Table 1 ) , including self-reported STI symptoms . HIV seroconversion was not associated with the detection of other cytokines ( GM-CSF , MCP-1 , MIP3α , IL-1a and RANTES ) , but power was limited due to the low prevalence of these cytokines . However , when the total number of detectable cytokines was considered as the primary exposure , the odds of seroconversion was found to increase significantly with the presence of two or more cytokines ( aOR 3 . 88 , 95% CI 1 . 21–12 . 50; Fig 1 and Table 2 ) . Since coronal sulcus cytokines were associated with increased HIV susceptibility , we examined circumcised and uncircumcised men who remained persistently seronegative to determine how circumcision impacts coronal sulcus cytokines levels . Enrolment demographics of men randomized to receive either immediate circumcision ( “circumcised” , n = 80 ) or delayed circumcision ( “uncircumcised” , n = 80 ) were similar ( S3 Table ) . Detectable IL-8 declined significantly after circumcision ( Fig 2 ) , even though the prevalence of detectable coronal sulcus cytokines was similar between the two groups at enrollment ( S4 Table ) . Among men who received circumcision , the prevalence of detectable coronal sulcus IL-8 declined significantly by month 6 post-circumcision ( PRR month 6 compared to enrollment was 0 . 59 , 95% CI: 0 . 44–0 . 87; Fig 2 ) and continued to decline throughout the 24 month follow-up period ( PRR 0 . 29 , 95% OR 0 . 16–0 . 54 ) ; the decline between months 6 and 24 was significant ( PRR 0 . 49 , 95% CI 0 . 25–0 . 96 ) . There were no significant changes in IL-8 detection among uncircumcised men . Even though MIG was associated with seroconversion , it did not change significantly after circumcision . Likewise , the prevalence of other coronal sulcus cytokines ( MIG , MCP-1 , MIP3α , IL-1a and RANTES ) showed no significant change after circumcision ( S4 Table ) . We found that coronal sulcus IL-8 and MIG were associated with increased HIV susceptibility , and that circumcision significantly reduced IL-8 . Given that IL-8 and MIG are both chemoattractant cytokines associated with recruitment of immune cells to sites of inflammation [41] , we therefore examined the link between levels of prepuce cytokines and the density of pro-inflammatory and HIV-susceptible immune cell populations in foreskin tissues . We measured IL-8 and MIG levels in coronal sulcus swabs collected from 89 men who underwent elective adult circumcision at the Rakai Health Sciences Program ( RHSP ) Circumcision Service Program , in whom we previously characterized foreskin T cell populations [39] . Participant demographics are provided in S5 Table; no behavioral characteristics recorded correlated with levels of IL-8 or MIG . IL-8 levels were above the LLOQ in 94 . 4% of participants ( 84/89 ) , and MIG was detectable in 51 . 7% ( 46/89 ) . We examined the correlation of each cytokine with the density of total CD4 and CD8 T cells , and also with the following HIV target cell populations: ( 1 ) CD4 T cells expressing the HIV co-receptor CCR5 ( CD3+/CD4+/CCR5+ ) ; ( 2 ) Th17 cells ( CD3+/CD4+/IL-17A+ ) ; ( 3 ) Th1 cells ( CD3+/CD4+/ IFNγ+ ) ; and , ( 4 ) CD4 T cells producing TNFα ( CD3+/CD4+/ TNFα+ ) . IL-8 concentration correlated with the density of both CD4 and CD8 T cells ( Fig 3A and 3B; p<0 . 05 ) , and with the density of CD4 T cell subsets known to be preferential HIV target cells: CCR5+ CD4 T cells , Th17 cells , Th1 cells , and TNFα+ CD4 T cells ( Fig 3C , 3D , 3E and 3F; all p≤0 . 02 ) . Having detectable coronal sulcus MIG was only associated with a non-significant trend of increased total CD4 ( 44 . 0 vs . 33 . 5 cells/mm2 , p = 0 . 08; Fig 3G ) , but with a significant trend to increased CD8 T cell density ( 35 . 5 vs . 22 . 7 cells/mm2 , p = 0 . 04 ) . To investigate the relationship of other tissue immune cell populations with prepuce cytokine levels , we next assessed neutrophil ( CD15+ ) and dendritic cell ( CD207+ Langerhans and CD11c+ dermal dendritic cell ) density in a subset of men with high ( n = 5; median IL-8 = 3422 . 6pg/ml , all MIG detectable ) and low ( n = 5; median IL-8 = 1 . 8pg/ml , all MIG undetectable ) coronal sulcus cytokine levels ( Fig 4A–4C ) . Neutrophils were found in both the epidermis and dermis , often in concentrated foci; Langerhans cells were found almost exclusively in the epidermis , and CD11c+ cells were predominantly in the dermis , as previously reported [42] . Men with high prepuce cytokine levels had a 4-fold higher density of tissue neutrophils than men with low cytokine levels ( 22 . 6 vs . 5 . 6 cells/mm2 , p = 0 . 016; Fig 4D ) . However , densities of dendritic cell populations were similar in men with high and low cytokine levels ( CD11c: 10 . 0 vs . 8 . 6 cells/mm2 dermal tissue , ns; CD207: 80 . 6 vs . 49 . 8 cells/mm2 epidermal tissue , ns ) . To rule out potential confounding of coronal sulcus cytokine levels by the vaginal secretions of a female sexual partner , cytokines were also assessed in female partner vaginal swabs collected on the day of male circumcision [43] for all 89 men in this analysis . Both IL-8 ( median 699 . 5pg/ml , range 1 . 5–5693 . 9pg/ml ) and MIG ( median 5 . 9pg/ml , range 0 . 3–853 . 4pg/ml ) were detectable in vaginal swabs , but we found no correlation between vaginal and penile cytokines within couples ( Spearman’s rho: IL-8 = 0 . 17 , MIG = 0 . 14 , both not significant ) , suggesting that cytokines detected in coronal sulcus swabs did not originate from the female partner .
Our study demonstrates a significant link between pro-inflammatory coronal sulcus cytokines and HIV acquisition in heterosexual men . Specifically , the chemoattractant cytokine IL-8 was associated with both an increased odds of seroconversion and an increased density of highly-susceptible HIV target cells in the foreskin . In addition we found that male circumcision , a procedure that significantly reduces HIV acquisition , progressively reduced detection of coronal sulcus IL-8 during two years of follow-up ( PRR of 0 . 29 at 24 months post-circumcision ) . Overall , these results suggest that the protective effect of male circumcision against HIV may be mediated in part through reductions in genital inflammation and the subsequent inflammation-mediated recruitment of HIV-susceptible cells to the foreskin . Although the mechanism ( s ) underpinning the relationship between cytokines and HIV susceptibility could not be fully elucidated by this observational study , the observation that coronal sulcus IL-8 and MIG were associated with HIV seroconversion is in keeping with a recent report that pro-inflammatory vaginal cytokines in women predict HIV acquisition [23] , and with in vitro experiments demonstrating that IL-8 increases HIV susceptibility in cervical explants [20] . Both IL-8 and MIG belong to the chemokine family , a group of structurally similar small molecules that act as chemoattractants for immune cells expressing appropriate receptors . Since HIV predominantly infects CD4 T cells [11 , 13 , 44] , both the availability and the HIV-permissivity of local CD4 T cells may dictate whether or not infection is established , with a limited number of highly susceptible cells driving initial mucosal infection [45 , 46] . Therefore , recruitment or activation of specific subsets of CD4 T cells that are especially permissive to HIV may be important; Th17 and Th1 cells are highly HIV-permissive in vitro and are preferentially depleted in vivo during acute infection [47–50] , and Th17 cells have recently been shown to be the primary targets of SIV , representing 64% of infected cells 48 hours after vaginal challenge [44] . Furthermore , men who are regularly HIV exposed but remain seronegative ( HESN men ) have a decreased relative abundance of both Th17 and TNFα+ CD4 T cells in their foreskin tissue [43] . Our finding that detectable coronal sulcus IL-8 was significantly associated with an increased overall number of CD4 T cells , including an increased density of highly susceptible Th17 , Th1 and TNFα+ CD4 T cells , suggests that target cell availability may contribute to the association between coronal sulcus chemokines and HIV susceptibility . While the association between foreskin IL-8 and HIV target cells does not prove this cytokine recruits or is produced by HIV target cells , a causal relationship is plausible . Epithelial cells ( including keratinocytes ) and macrophages responding to early immune stimuli are the main source of IL-8 during skin inflammation [51–54] . While many cell types express the IL-8 receptors CXCR1 and CXCR2 [55] , IL-8 primarily recruits and activates neutrophils [51 , 56 , 57] , which is consistent with our data showing a 4-fold higher density of neutrophils in foreskin tissues of men with high prepuce IL-8 levels . Neutrophils recruited by IL-8 are activated by bacterial antigens in the presence of inflammatory cytokines ( IFNγ ) to produce Th17-recruiting chemokines ( MIP-3α and MCP-1 ) and Th1-recruiting chemokines ( MCP-1 and IP-10 ) [58–60] . In turn , Th17 and Th1 cells produce IL-17 and IFNγ , respectively [61] , both feeding back into the inflammatory cascade: IL-17 stimulates epithelial cells to produce more IL-8 [62] , and IFNγ contributes to neutrophil chemokine production [60 , 63] . Reciprocally , Th17 cells may directly contribute to IL-8 levels: Th17 cells have recently been shown to be the only subset of CD4 T cells to produce high levels of IL-8 [60 , 64 , 65] . In support of this , treatment with antibodies preventing the formation and action of Th17 cells ( secukinumab , specific for IL-17A/IL-23 ) prevents neutrophil recruitment to the skin and keratinocyte immune activation , and reduces local levels of IL-8 [66] . Thus , IL-8 may be part of a positive-feedback loop , whereby local innate immune cells recruit neutrophils through IL-8 , which in turn recruit HIV target cells through MIP-3α and MCP , and these HIV target cells then produce inflammatory cytokines , feeding back into local innate immune cell activation and IL-8 production . We hypothesize that this positive-feedback loop provides a causal basis for the association that we observed between prepuce IL-8 levels and HIV target cells , and that HIV target cells recruited by this mechanism drive the observed increased risk of seroconversion . HIV seroconversion in our cohort was also associated with the detection of MIG and was increased if multiple cytokines were present . MIG is produced by macrophages in response to IFNγ and directly recruits activated T cells through CXCR3 [67] , consistent with our observations that men with detectable MIG had increased numbers of foreskin CD8 T cells and a trend towards increased CD4 T cells ( p = 0 . 08 ) . The recruitment of IFNγ-producing T cells by MIG may also feed into to the relationship between IL-8 and HIV target cell recruitment , since IFNγ contributes to neutrophil activation and Th1/Th17 cell recruitment ( as above ) . Of note , MCP-1 and MIP-3α were the next most abundant cytokines detected in this study ( after IL-8 and MIG ) , and these two chemokines have been shown to specifically mediate the recruitment of Th17 cells by neutrophils [60] . We therefore hypothesize that increased risk of seroconversion with multiple cytokines is due to MIG , MCP-1 and MIP-3α feeding into the local inflammatory feedback loop described above . However , while this observational study provides important in vivo human data , further in vitro or animal studies will be necessary to completely define the causal nature of these associations . Other cell types , such as Langerhans cells and dermal dendritic cells have been shown to facilitate HIV infection in the foreskin [68] , and may contribute to increased HIV-susceptibility observed in this study . While tissue density of Langerhans cells and dermal dendritic cells was not associated with prepuce IL-8 or MIG , local tissue inflammation may contribute to dendritic cell maturation and function , thereby facilitating HIV transfer to susceptible T cell populations , as has been previously described [68 , 69] . The reduction in coronal sulcus chemokines after male circumcision sheds light on a potential biological mechanism by which circumcision protects against HIV acquisition: reduced penile inflammation . We observed a significant decrease in coronal sulcus IL-8 at the first follow-up visit after circumcision ( PRR = 0 . 59 at 6 months , 95% CI: 0 . 40–0 . 88 ) , and IL-8 continued to decline up to study conclusion ( PRR = 0 . 29 at 24 months , 95% CI: 0 . 16–0 . 54 ) , which was significantly lower than at month 6 . This may reflect a gradual decline in HIV target cells within remaining penile tissue , as effector T cells have been shown to be slow to clear a site of previous infection , and are enriched in tissue sites for months after antigen becomes undetectable in the skin [70] . In vitro studies also suggest that target cell availability contributes to the efficacy of male circumcision , as the inner aspect of the foreskin contains a comparatively high density of HIV target cells [15 , 33–35] . Male circumcision may reduce HIV susceptibility by reducing penile inflammation and HIV target cell availability . The factors causing penile inflammation could not be fully delineated in this study . Prepuce cytokine levels in uncircumcised men do not correlate with urethral cytokines ( Kaul R and Galiwango R; unpublished ) , and we found no association with female partner vaginal cytokines , suggesting that prepuce cytokines do not derive from these sources , and may be produced in foreskin tissue . No significant associations were observed between IL-8 or MIG and age , number of sex partners , condom use , genital washing or seroprevalent STIs ( HSV-2 and syphilis ) . While self-reported STI symptoms were associated with increased detection of both IL-8 and MIG , this did not fully account for the association between these cytokines and seroconversion . Additionally , even in symptom-free participants uninfected by HSV-2 and syphilis , we observed IL-8 concentrations ranging over 1000-fold ( >1 . 5 to 2626 . 9pg/ml ) , suggesting that “normal” penile immune parameters are highly variable , and may be affected by factors other than classical STIs . This is in agreement with findings in South African women , where vaginal chemokine levels were associated with increased risk of HIV acquisition , but were incompletely explained by the presence of STIs [23] . Determining the factors contributing to heterogeneity in genital inflammation warrants further research as prior simulation studies have shown that variability in HIV susceptibility can affect HIV epidemic dynamics significantly and may explain differences in HIV epidemic trajectories between populations [71] . One possible contributor to penile inflammation is the resident microbiome . Alterations in the vaginal bacterial community in women , such as bacterial vaginosis , are associated with increased risk of HIV [72–75] , possibly due to local inflammation [76 , 77] . Th17 cells are essential in the defense against bacterial infections [78 , 79] , and colonization by pathogenic bacteria may increase HIV susceptibility by increasing Th17 cell density [50] . We have previously found that uncircumcised men are more likely to have BV-associated anaerobic bacteria in their sub-preputial space [80] , and that circumcision gradually reduces both the total bacterial load and the abundance of these anaerobes [32] . Of note , while anaerobe abundance decreased rapidly within 6 months post-circumcision , it continued to decline for up to 24 months . This is similar to the gradual decline in IL-8 levels that we observed over the same period , and so the role of the penile microbiome as a driver of tissue inflammation and HIV susceptibility may be an interesting area for future study . A limitation of the current work was the low concentration of cytokines in coronal sulcus swabs , especially swabs collected during the circumcision RCT . Swabs from the RCT were stored at -80°C for up to 10 years between collection and cytokine analysis . Cytokines , including IL-8 , have been shown to degrade after 4 years , despite ideal storage conditions [81] . This likely explains the difference in detectability of IL-8 between swabs collected during the RCT and those collected from men attending the Circumcision Service Program , as swabs from the latter group were analyzed within one year of collection . However , it is unlikely that IL-8 degradation can account for differences observed between comparator groups in this study . In the case control study of HIV seroconversion , controls were matched to cases based on visit ( time ) and swab storage time did not vary between groups ( median 4 . 7 years for both groups ) . Additionally , we found no correlations between IL-8 concentration and date of swab collection , suggesting that variability in swab storage time due to the relatively short duration of the trial ( August 2003- November 2006 ) cannot account for the differences in IL-8 levels observed when examining the impact of circumcision on cytokine levels . It is possible that associations with other cytokines may have been missed due to low analyte concentration , explaining why cytokines observed to be released ex vivo from foreskin explants were not detectable in swabs in this study [33 , 35 , 69] . In conclusion , penile inflammation is an important risk factor for HIV acquisition in heterosexual men . HIV acquisition was associated with elevated levels of coronal sulcus IL-8 and MIG , which correlated with an increased density of T cells in the underlying foreskin tissue . In particular , prepuce concentrations of IL-8 were correlated with both an increased overall tissue density of CD4 T cells , as well as an increased density of specific highly HIV-susceptible CD4 T cell subsets . Finally , circumcision progressively reduced coronal sulcus IL-8 for up to 24 months after the procedure , which suggests a reduction in penile inflammation may be one mechanism by which circumcision is protective against HIV . Identifying causes of penile inflammation and immune activation in otherwise healthy men may lead to novel interventions to reduce the sexual transmission of HIV .
We examined samples and data collected from two study populations enrolled through the Rakai Health Sciences Program ( RHSP ) in Uganda: one enrolled in an RCT of male circumcision , conducted from 2003–2006 [29]; and the second enrolled in an observational cross-sectional study through the RHSP Circumcision Service Program between 2010–2011 [39] . Study design and sample population selection are described in detail in the Statistical Methods Section , below . Dacron swabs moistened with sterile saline and rotated twice around the full circumference of the penis at the coronal sulcus were collected from all men at enrollment and each follow up visit during the RCT , and once , immediately prior to circumcision , from the Circumcision Service participants . The same clinical officers collected swabs throughout both studies and care was taken to collect each swab in a consistent manner . Female partners of Circumcision Service Participants were asked to insert a Dacron swab into the vagina , rotate once , and remove it . Swabs were immediately placed in 1ml undiluted AMPLICOR STD Specimen Transport Kit medium ( Roche Diagnostics , Indianapolis , IN ) at 4°C for less than 4 hours , and then suspended , aliquoted and stored at −80°C . Foreskin tissue removed during circumcision was also collected from Circumcision Service Participants . Tissue was processed immediately upon surgical removal: two sections from distal locations on the foreskin ( one from the approximate center of the inner aspect , and one from the center of the outer aspect ) were snap frozen into cryomolds in Optimal Cutting Temperature ( OCT ) compound ( both Fisher Scientific , Toronto , Canada ) for immunohistochemistry; and one large section containing equal area of the inner and outer aspects reserved for T cell isolation . An electrochemiluminescent detection system using a custom Human Ultra-Sensitive kit from Meso Scale Discovery ( Rockville , MD ) was used to assay cytokines in coronal sulcus swabs from both RCT and Service Program participants . Cytokines assessed were: IL-1α ( interleukin-1α ) , IL-8 , MCP-1 ( monocyte chemotactic protein-1 ) , MIG ( monokine induced by γ-interferon ) , MIP-3α , RANTES ( Regulated on Activation , Normal T cell Expressed and Secreted ) , and GM-CSF ( granulocyte macrophage colony-stimulating factor ) . Samples from each of the three analysis sets ( Coronal sulcus cytokines and HIV acquisition , Impact of circumcision on coronal sulcus cytokines , and Prepuce cytokines and foreskin T cell density ) were run on plates from the same manufacturer’s lot , with samples from participant groups in each analysis set ( i . e . cases and controls , circumcised and uncircumcised ) distributed randomly and evenly proportioned across plates . Samples were run in duplicate , and results with a coefficient of variation ( CV ) above 20% for the two wells were re-run . An internal biological control was run in at two concentrations in duplicate on every plate to monitor plate-to-plate variability: internal control was made from pooled mucosal samples from 5 donors , with any low level analytes augmented by adding the relevant recombinant human protein . This sample was aliquoted for single use , and run both neat and diluted 1/20 on each plate ( for biological and low-concentration controls ) . Plates were re-run if the concentration of any analyte in the internal control was >3 standard deviations different from the average concentration from that analyte for the first 5 plates run . Plates were imaged using the Sector Imager 2400A platform ( Meso Scale Discovery ) . The study lower limit of quantification ( LLOQ ) for each analyte were as follows: IL-1α = 0 . 6pg/ml; IL-8 = 1 . 5pg/ml; MCP-1 = 0 . 6pg/ml; MIG = 0 . 3pg/ml; MIP-3α = 3 . 0pg/ml; RANTES = 0 . 6pg/ml; and GM-CSF = 0 . 3pg/ml . Cytokine concentrations reported are not normalized , and are that of swab resuspended in 1ml transport buffer . Levels are therefore significantly lower than true concentration on the penis surface . T cells were isolated from foreskin samples obtained from Service Program participants as previously described [39] . Mononuclear cell counts were determined by trypan blue exclusion and 10-20x106 cells ( depending on yield ) were stimulated with either 1ng/ml phorbol-12-myristate-13-acetate ( PMA ) and 1μg/ml ionomycin ( both from Sigma; St . Louis , MO , USA ) or vehicle ( 0 . 1% DMSO ) in the presence of 5μg/ml Brefeldin A ( GolgiPlug , BD Biosciences ) . After stimulation , samples were stained for CD3 ( UCHT1 ) , CD4 ( RPA-T4 ) , CD8 ( SK1 ) and CCR5 ( 2D7/CCR5; all BD Biosciences ) . Samples for intracellular staining were permeabilized using BD Cytofix/Cytoperm solution and stained with TNFα ( MAb11 ) , IFNγ ( B27; all BD Biosciences ) , or IL-17A ( eBio64DEC17; eBiosciences ) . Samples were acquired using a FACSCalibur flow cytometer ( BD Systems ) . Gating was performed as previously described [39] by investigators blinded to participant status and cytokine levels . Proportions of T cell subsets were converted to absolute numbers per mm2 foreskin tissue using CD3 IHC as previously described [83] . Briefly , OCT-cryopreserved tissues were sectioned , fixed in 2% formaldehyde , and frozen for batch staining . Sections were stained with anti-CD3 antibody , followed by biotin-labeled secondary , Alkaline Phosphatase Streptavidin Labeling Reagent and Substrate Kit Vector Red ( all Vector Labs , Burlingame , CA ) , and then counterstained with Mayer’s Hematoxylin ( Fisher Scientific ) . The number of CD3+ T cells per mm2 of tissue for each patient was derived from the average of two biopsies taken from distal locations on the foreskin ( median 6 . 10mm2 tissue/patient analyzed ) . Whole sections were scanned using the TissueScope 4000 ( Huron Technologies , Waterloo , Canada ) and image analysis software ( Definiens , München , Germany ) was used to delineate the apical edge of the epidermis to a depth of 300μm into the dermis ( excluding artifacts or folds ) . CD3+ cells within this area were manually counted by a single investigator blinded to cytokine levels and participant status . Neutrophil , Langerhans cell , and dermal dendritic cell density was assessed using immunofluorescence in a subset of men with high ( n = 5 ) and low ( n = 5 ) levels of coronal sulcus cytokines . Tissues cryopreserved in OCT were sectioned to 5μm using a Leica CM3050 cryostat ( Leica Microsystems , Wetzlar , Germany ) , mounted on glass microscope slides , fixed for 7 minutes in ice-cold acetone , air-dried , and frozen at -80°C for batch staining . For staining , slides were thawed , permeabilized in PBS-Tween 20 for 20 minutes , and blocked using a streptavidin/biotin blocking kit ( Vector Labs ) and 10% normal rabbit serum . Neutrophils were visualized using biotin-labeled mouse anti-human CD15 antibody ( eBiosciences ) followed by DyLight 488 Streptavidin ( Vector Labs ) secondary . Dendritic cells were visualized with goat anti-human CD207 antibody ( R&D Systems ) and biotin-labeled mouse anti-human CD11c ( eBiosciences ) . Slides were then washed and mounted using Vectashield HardSet Mounting Medium with DAPI Counterstain ( Vector Labs ) , according to manufacturer’s instructions . Whole sections were scanned using the Zeiss Axioscan ( Carl Zeiss Microscopy , Cambridge , UK ) and image analysis software ( Definiens ) was used to delineate and quantify the epidermal and dermal tissue ( excluding artifacts or folds ) . Definiens was then used to count cell populations using a threshold set by an investigator blinded to cytokine levels and participant status . CD15+ cells were counted within total foreskin area ( median area analyzed = 4 . 06 mm2 ) , CD207+ cells in the epidermal tissue ( 1 . 36 mm2 ) , and CD11c+ cells within dermal tissue ( 3 . 39 mm2 ) . We used Stata 13 . 1 for Mac ( College Station , TX , USA ) to conduct statistical analysis and Prism 5 . 0 ( GraphPad Software; La Jolla , CA , USA ) to construct graphs . Flow cytometry data was analyzed in FlowJo 9 . 8 . 2 ( Treestar; Ashland , OR , USA ) . All tests two-sided with α = 0 . 05 . | The per-contact risk of infection with HIV through sexual exposure is low and highly variable . Understanding the biological basis for this variability could help in the development of new methods to prevent infection . There is some evidence that penile inflammation , even in the absence of any clinical symptoms , may increase HIV-susceptibility by recruiting CD4 T cells , the immune cell type that is the principal target of HIV . We analyzed soluble inflammatory mediators in prepuce swabs collected longitudinally from initially HIV-negative men enrolled in a randomized controlled trial of adult circumcision . We found that these inflammatory mediators were elevated in men who went on to acquire HIV . We also found that higher levels of these mediators were associated with an increased density of HIV-susceptible target cells in the underlying foreskin tissue and that circumcision reduced their levels , which may help to explain why circumcision reduces HIV risk by 60% or more . Together , these data suggest that penile inflammation , in the absence of genital infections , increases HIV susceptibility and is reduced by adult male circumcision . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"blood",
"cells",
"t",
"helper",
"cells",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"pathology",
"and",
"laboratory",
"medicine",
"immune",
"cells",
"pathogens",
"antigen-presenting",
"cells",
"immunology",
"microbiology",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"developmental",
"biology",
"dendritic",
"cells",
"rna",
"viruses",
"signs",
"and",
"symptoms",
"molecular",
"development",
"neutrophils",
"reproductive",
"system",
"procedures",
"white",
"blood",
"cells",
"inflammation",
"animal",
"cells",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"t",
"cells",
"immune",
"response",
"immune",
"system",
"diagnostic",
"medicine",
"cell",
"biology",
"physiology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"lentivirus",
"circumcision",
"organisms"
]
| 2016 | Chemokine Levels in the Penile Coronal Sulcus Correlate with HIV-1 Acquisition and Are Reduced by Male Circumcision in Rakai, Uganda |
The immune system has developed a number of distinct complex mechanisms to shape and control the antibody repertoire . One of these mechanisms , the affinity maturation process , works in an evolutionary-like fashion: after binding to a foreign molecule , the antibody-producing B-cells exhibit a high-frequency mutation rate in the genome region that codes for the antibody active site . Eventually , cells that produce antibodies with higher affinity for their cognate antigen are selected and clonally expanded . Here , we propose a new statistical approach based on maximum entropy modeling in which a scoring function related to the binding affinity of antibodies against a specific antigen is inferred from a sample of sequences of the immune repertoire of an individual . We use our inference strategy to infer a statistical model on a data set obtained by sequencing a fairly large portion of the immune repertoire of an HIV-1 infected patient . The Pearson correlation coefficient between our scoring function and the IC50 neutralization titer measured on 30 different antibodies of known sequence is as high as 0 . 77 ( p-value 10−6 ) , outperforming other sequence- and structure-based models .
The prediction of antibody ( Abs , or immunoglobulins , Igs ) affinity for antigens is among the most interesting open challenges across bioinformatics and structural immunology . Most of the current methods rely on the structures ( either experimentally resolved or modeled ) of both antibodies and their cognate antigens to predict their binding affinity . Currently , available methods are time demanding and , more importantly , their predictions are hard to assess [2 , 3] . On the other hand , because of the scarcity of available data-sets for which both Abs sequences and their affinity for an antigen are known , there is still no method that can model the affinity as a function of the sequence of the antibody variable region . Also , it is still not clear if and how it would be possible to set up a coherent fitting procedure to estimate the ( possibly ) huge number of parameters of a generic mapping from the space of Abs sequences to the affinity for the antigen . Thanks to the recent developments of sequencing techniques ( e . g . Deep Sequencing , and Next Generation Sequencing ) , Repertoire Sequencing ( Rep-Seq ) experiments ( see [4] for a review of the argument ) start to be routinely performed . Recently , the complete Ig repertoires of several simple organisms such as the zebra-fish , whose immune system has only ∼300 . 000 Abs producing B cells , have been sequenced [5] . Higher organisms , such as humans , show a remarkably more complex immune system and it is widely accepted that the typical human Ab repertoire amounts to ∼109−10 different molecules . In this case , a large sample of the entire repertoire can be extracted ( see for example [6] for Rep-Seq experiment on Igs in human ) . Rep-Seq data allow for a detailed description of the sequences distribution based on Maximum Entropy ( MaxEnt ) modeling of repertoires , as it has been proven in the case of zebra-fish Abs [7] and human T cell receptors [8 , 9] . While these studies focus on a model-based description of the initial repertoire of the adaptive immune system arising mainly from the V ( D ) J genetic rearrangement , here we focus on the affinity maturation process . A number of statistical mechanics inspired methodologies have been recently successfully devised to analyze evolutionarily related proteins for inferring structural properties and , in particular , residue-residue contacts [10] . In particular , homologous proteins can be characterized in terms of multiple sequence alignments ( MSAs ) . In spite of the considerable sequence heterogeneity ( up to only 40% sequence identity ) in families of homologous proteins , their folded structures are often almost completely conserved [11] . A MaxEnt modeling technique developed more than a decade ago , could detect signals of the evolutionary pressure beyond the sequence variability in MSAs of homologous proteins [12] . Maintaining the same underlying idea that co-evolution of residue pairs is related to their spatial proximity in the folded protein structure , a large number of works successfully reconsidered MaxEnt in different flavors: ( i ) the application of mean-field approximations known as Direct-Coupling Analysis ( DCA ) [13–15] , ( ii ) pseudo-likelihood maximization ( PlmDCA ) , [16–18] , ( iii ) Multivariate Gaussian Modeling ( MGM ) , [19 , 20] . All these methods rely on the inference of a generative probabilistic model for sequences in the presence of selective pressure . This feature makes this kind of analytic techniques particularly suited for the study of Ab affinity maturation . In fact , this process closely resembles a Darwinian evolutionary framework where B-cell clones compete for the antigen in the germinal centers , and it is now widely accepted that the affinity for the target antigen represents the main contribution to the fitness in this evolutionary scenario . Thus , as qualitatively sketched in Fig 1 , for every antigen , the evolutionary dynamics explores the space of Ab sequences searching for the global optimum of the fitness function , i . e . the best affinity for the related antigen . Here we exploit the evolutionary nature of the affinity maturation process by applying a MaxEnt inference techniques originally developed for the analysis of homologous protein families . The above mentioned plethora of model inference methods aim at reconstructing a reliable contact map from the space of homologous protein sequences through an analysis of residues coevolution that disentangle indirect correlations , but in our context , they provide little information on Abs internal structure . However , the inference procedure provides a natural and reliable scoring function ( see Section “Inference Methods” ) from the space sequences to that of binding affinity for the target antibody related to the probability for a sequence to appear in the data set that we can use as a proxy to the binding affinity to the antigen , in the spirit of series of recent publications [21–23] where deep sequencing of the immune repertoire was used to predict binding vs . non-binding Abs with different therapeutic applications . Finally , we report that very recently maximum entropy modeling has been also used in [24] to predict the fitness landscape of the HIV-1 protein from the relative abundance of the virus strains , and in [25] to predict in silico the effect of mutations related to disease and antibiotic drug resistance .
Wu and coworkers [1] used 70 sequenced heavy chain variable regions , which originated mostly from immunoglobulins using the IGHV1-2 gene , for constructing chimeric antibodies by combining them with the light chain of VRC-PG04 . Among these , 45 have been tested for their neutralization power against 20 HIV-1 mutations . When included in the sequencing data set and used as input for the clustering procedure , 30 of these 45 tested Abs are found to belong to the hypermutated cluster . The remaining 15 ( none of which was found to be neutralizing ) belong to the germline cluster . Although in general the neutralization power depends on both the light and the heavy chain sequences ( cf . Fig . 4A in [1] ) , the light chain plays only a minor role in the interaction ( most notably steric contacts with its CDR1 and CDR3 regions ) here , as visible from the solved structure of VCR-PG04 ( PDB code 3SE9 ) . We therefore will make the simplifying assumption that the neutralization measurements on chimeric Abs depend on the heavy chain contribution alone . Under the assumption that the hypermutated cluster is a statistically representative sample of the Abs that underwent affinity maturation against gp120 , we can use the statistical properties of this set of sequences to construct a predictor for the Abs neutralization power . We thus inferred an MGM on the MSA of this cluster and used the MGM-score of the inferred model as a proxy for the neutralization power of the related Abs . Although the inference step is completely blind to the binding affinity of the Abs ( the binding affinities of sequences belonging to the hypermutated cluster were not measured in [1] ) , nonetheless the capability of predicting binding energies is not unexpected . Indeed , the aim of a maximum entropy model of the hypermutated set , is to provide an accurate statistical description of the set of Abs responding to gp120 , and so it is not completely surprising that , according to the model , sequences with low probability are more likely to have a low binding affinity for the antigen compared to sequences of high probability . To test the predictive power of the method , we used the panel of 30 sequences ( not included in the hypermutated cluster ) tested for HIV neutralization power and compared the IC50 neutralization titer with the MGM-score of the same sequence . Note that values of IC50 that are reported in [1] as greater than 50 μg/ml ( not-neutralizing ) are considered here to be equal to this value . The two quantities are compared by means of the Pearson correlation coefficient . We consider as measures of the neutralization power the average IC50 over the different neutralized viruses . A scheme of the model inference and testing procedure is shown in Fig 3 . The result of the model inference procedure depends on the choice of the regularization parameter π defined in the “Inference methods” section . We therefore repeated the test procedure for different values of π . In Fig 4 the Pearson correlation coefficient between the MGM-score and the average IC50 over the neutralized viral isolates is shown for different values of π . The two panels refer to the two score proposed: the original inferred MGM-score and the MGM-score with gap correction ( see Section “Score with gap correction” for details ) . We thus argue that the MGM-score inferred on a representative Rep-Seq data set provides a remarkably good proxy for the neutralization power of the analyzed sequence . We also display the details of our best performance on a per-virus base in Fig 5 . We also assessed the performance of the MGM-score to discriminate binding vs . non-binding sequences . The dataset in this somehow simpler task reduces in a set of 21 non-binding and 24 binding sequences . The performance of the MGM-score are displayed in terms of the ROC curve shown in Fig . F in S1 Text ( red curve ) : the ( normalized ) area under the ROC ( AUROC ) turns out to be 0 . 97 . We also compared this value against a much simpler scoring strategy defined in terms of the Hamming distance from the consensus sequence of the hypermutated cluster . As shown in Fig . F in S1 Text ( blue curve ) , the AUROC turns out to be 0 . 86 . We also inferred the model using PlmDCA [17] rather than MGM . The results are shown in Fig . G in S1 Text: The best Pearson correlation coefficient obtained with this method of inference is slightly worse than the one obtained with MGM . This result is non-trivial since PlmDCA is known to perform better than MGM in terms of protein contact prediction . We also note that in a recent publication [25] , a variant of DCA ( mean-field DCA ) that is essentially equivalent to MGM was used to successfully predict the ΔΔG between mutants and wild type sequences for the beta-lactamase TEM-1 . A natural question is whether simpler inference strategies might achieve equally good results , and in particular whether it is necessary to use the second order statistics ( i . e . multivariate vs univariate statistics ) to infer Abs neutralization power . To this end , we tested a simpler version of the model , factorized over the different residues of the MSA . In this model the non-diagonal J terms are set to zero so that the residues are statistically independent ( see Section “Multivariate Gaussian Modeling” and [20] ) . As shown in Fig 4 ( squares and dashed lines ) , the Pearson correlation coefficient is dramatically reduced , dropping from a maximum of 0 . 77 for the full MGM to a maximum of 0 . 49 for the factorized model . Our neutralization power predictor was compared with another sequence based method , the HMM-score ( see Section “Using Hidden Markov Models to predict binding affinities” ) . This score takes only correlations between nearest neighbors in the sequence into account . Interestingly , as displayed in Fig 4 , the prediction quality of this method is between the one obtained using the factorized MGM-score and the one obtained using the full MGM-score . This supports the observation that long range intragenic epistatic signals are crucial to reproduce neutralization power . An important step in the procedure is to correctly identify the set of sequences that underwent affinity maturation towards the same epitope . Indeed , MGM models trained on different sets ( for example the entire set of sequences coming from the germline of interest ) display no significant correlation with neutralization measurement . Some portions of the MSA are observed to be more important than others in reproducing the affinity function: The correlation between the inferred likelihood and the neutralization titers is essentially the same when only the ∼60 more variable residues of the hypermutated cluster MSA are used to construct the MGM , dismissing ∼3/4 of the columns of the MSA . Data of this MSA reduction analysis are reported in S1 Text ( see Section “Affinity predictions” ) . Our predictor was also compared with a structure-based method: we produced structural models for all the 45 antibody/antigen complexes for which the IC50 was measured and predicted their binding affinity using FoldX ( see Methods for details ) . The results of this structural method show no significant correlation ( r = −0 . 23 , p-value = 0 . 13 ) with the experimental data . Taken together , our findings indicate that: ( i ) MGM inferred on the proper set of clonally expanded sequences contains enough information to predict the neutralization power of Ab sequences . This suggests that the procedure can be used as a tool to generate new and highly neutralizing Abs; ( ii ) taking into account ( pairwise ) intragenic epistatic effects in the model improves remarkably the accuracy of the affinity prediction .
The identity/divergence analysis performed in [1] on the whole deep sequencing data set indicates that sequences with inferred IGHV1-2 germline gene ( the same of VRC-PG04 ) are characterized by: ( i ) the presence of a cluster of highly mutated sequences that is well separated from the cluster of typically mutated sequences; ( ii ) Abs with a different IGHV inferred germline gene display a more uniform ( i . e . less clustered ) structure . We performed an independent identity/divergence analysis on the data set resulting from our bioinformatics analysis in which we retain only productive sequences of IGHV1-2 origin . Our results are in complete agreement with [1] , as shown in Fig 7 . There we compare the identity to VRC-PG04 and the divergence from IGHV1-2*02 germline gene at a nucleotide level for each sequence in the data set . Identity/divergence analysis gives a glimpse of the structure of the sample in the space of sequences . Nevertheless , a less biased analysis is required in order to test the cluster structure . We thus performed a sequence-based clustering analysis . Among the different clustering algorithms available , we chose the shallow tree clustering algorithm [32] since it provides a criterion of robustness against noise ( see S1 Text Section “Sequence clustering analysis” ) . The clustering algorithm is based on the Hamming distance between sequences . The most robust solution ( see S1 Text for an explanation of what robust means in this context ) found by the algorithm is a partition of the sequences into two clusters: a germline cluster composed of 2878 sequences ( 1634 unique ) centered on the IGHV1-2*02 and IGHJ2*02 germline genes ( with an average sequence divergence of ∼5% from the germline ) , and a hypermutated cluster composed of 3896 sequences ( 1578 unique ) more similar to the broadly neutralizing antibody VRC-PG04 ( with an average sequence divergence of ∼35% from the germline , see Section “Clustering analysis” in S1 Text for details ) . These results are confirmed by a test with the k-means clustering algorithm ( run with k = 2 ) . Information about the two clusters and their MSA characteristics are resumed in Table 2 . In the present work , we assume the hypermutated cluster to be a representative sample of the Abs that underwent affinity maturation for neutralizing HIV-1 gp120 . The structure of VRC-PG04 in complex with gp120 ( PDB-id 3SE9 ) has been subjected to both visual inspection and quantitative predictions to assess the importance of each somatic mutation observed in the antibody to the binding affinity towards the antigen . Somatic mutations were retrieved using the IMGT database [34] . We used the FoldX software [35] to predict the difference in binding energy ( ΔΔG ) of the actual antibody with all the mutants obtained reverting each single somatic mutation to the original residue observed in the germline gene IGHV1-2 .
In the present study , we proposed a sequence based maximum entropy model to analyze Ab affinity for the antigen . The predictive validity of the model has been tested using Rep-Seq data and neutralization power measurements from an HIV-1 infected donor [1] . The interplay between the HIV-1 virus and the immune response provides an interesting framework for our purpose: the affinity maturation of the Abs of interest ( those whose epitope is the gp120 CD4 binding site ) causes a dramatic increase of their neutralization power and a pronounced mutation ratio in comparison with the germline genes . This high density of mutations allows us to easily select sequences in the immune repertoire that respond to the antigen . A maximum entropy model constructed on this set of hypermutated sequences has been successfully used as a predictor of the neutralization power of Abs . This predictor has been successfully assessed against experimental neutralization measurements of different viral isolates . These positive results suggest that the procedure could be used as a tool for generating new and highly neutralizing Abs . In analogy with the application to protein families [12–20] , the MaxEnt model has been used for predicting residue-residue contacts in the Rep-Seq sample without obtaining positive results . This is not surprising since the time-scale involved in the affinity maturation process ( years ) is not comparable to the typical evolutionary time-scale in protein families ( millions of years ) . The structure of the inferred statistical interactions is probably mostly driven by the interaction with the epitope and further investigations in this sense represent an interesting development of this work . Nevertheless , the joint analysis of the sequencing data statistics and neutralization measurements has been shown to provide some consistent structural information on antigen recognition mode . In conclusion , the use of maximum entropy models can unveil relevant features of the protein fitness function . These features are related to the affinity maturation process and in particular to the evolutionary dynamics of the B cell population . This could be of interest for a statistical population genetics analysis of the affinity maturation process ( for example in the spirit of [39] and [40] ) . The present case study shows how MaxEnt methods can be a useful tool for tackling immunological questions in a time when Rep-Seq data are becoming increasingly popular in immunology ( see for instance [41] , where T receptor repertoires are studied ) . | Affinity maturation is a very complex biological process which enables activated B-cells to produce antibodies with increased affinity for a given antigen . Once B-cells begin to proliferate , each of the progeny cells introduces mutations in the antigen binding region in order to explore different affinities for the antigen . Selection rounds occurring in the so-called germinal centers in lymph nodes and spleen prune out poorly binding receptors and clonally expand good binders . Thanks to high-throughput sequencing techniques it is now possible to have access to a fairly representative sample ( of the order of 105 to 106 sequences ) of the immune repertoire of a given individual . Our approach is to first exploit this large amount of sequence data to infer a statistical model for the sequenced portion of the immune repertoire , and then to use the inferred probability of this model as a score when predicting the neutralization power of a given antibody sequence for the antigen of interest . The results we obtained on a specific data set of sequences of an HIV-1 patient show that our score correlates very well with experimentally assessed neutralization power of specific antibodies of known sequence . The performance of the method crucially relies on the ability of our model to account for long-range intragenic epistatic interactions between residues along the whole antibody chain . | [
"Abstract",
"Introduction",
"Results",
"Methods",
"Discussion"
]
| [
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"markov",
"models",
"evolutionary",
"biology",
"split-decomposition",
"method",
"pathogens",
"immunology",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"multiple",
"alignment",
"calculation",
"mathematics",
"rna",
"viruses",
"molecular",
"biology",
"techniques",
"antibodies",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"immune",
"system",
"proteins",
"sequence",
"alignment",
"proteins",
"medical",
"microbiology",
"hiv",
"hidden",
"markov",
"models",
"microbial",
"pathogens",
"hiv-1",
"evolutionary",
"immunology",
"molecular",
"biology",
"probability",
"theory",
"biochemistry",
"dna",
"sequence",
"analysis",
"computational",
"techniques",
"physiology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"lentivirus",
"organisms"
]
| 2016 | Maximum-Entropy Models of Sequenced Immune Repertoires Predict Antigen-Antibody Affinity |
Nipah virus is a broadly tropic and highly pathogenic zoonotic paramyxovirus in the genus Henipavirus whose natural reservoirs are several species of Pteropus fruit bats . Nipah virus has repeatedly caused outbreaks over the past decade associated with a severe and often fatal disease in humans and animals . Here , a new ferret model of Nipah virus pathogenesis is described where both respiratory and neurological disease are present in infected animals . Severe disease occurs with viral doses as low as 500 TCID50 within 6 to 10 days following infection . The underlying pathology seen in the ferret closely resembles that seen in Nipah virus infected humans , characterized as a widespread multisystemic vasculitis , with virus replicating in highly vascular tissues including lung , spleen and brain , with recoverable virus from a variety of tissues . Using this ferret model a cross-reactive neutralizing human monoclonal antibody , m102 . 4 , targeting the henipavirus G glycoprotein was evaluated in vivo as a potential therapeutic agent . All ferrets that received m102 . 4 ten hours following a high dose oral-nasal Nipah virus challenge were protected from disease while all controls died . This study is the first successful post-exposure passive antibody therapy for Nipah virus using a human monoclonal antibody .
Nipah virus ( NiV ) together with Hendra virus ( HeV ) are closely related highly pathogenic zoonoses and are the type species within the paramyxovirus genus Henipavirus . Both viruses can cause significant morbidity and mortality in a variety of vertebrate species including humans . The henipaviruses are categorized as zoonotic biosafety level 4 ( BSL-4 ) agents which has limited an extensive examination of their in vivo pathogenic features and the development and evaluation of therapeutics or vaccines . NiV and HeV are select agents of biodefense concern that are classified as priority pathogens in category C by the National Institute of Allergy and Infectious Diseases and the Centers for Disease Control and Prevention , with the potential to cause significant morbidity and mortality in humans and major economic and public health impacts ( reviewed [1] ) . Pteropid bats ( family Pteropodidae ) , commonly known as flying foxes , are the predominate natural reservoirs for both HeV and NiV ( reviewed [2] ) although evidence of henipavirus infection has now been reported in a wider range of both frugivorous and insectivorous bats [3] , [4] . Since their initial recognition , both viruses have repeatedly re-emerged . In total , 13 HeV outbreaks have occurred in Australia in 1994 , 1999 , 2004 , and 2006–2009 , and have always involved horses as an intermediate host with some human infections including four fatalities , the most recent in September 2009 ( reviewed [2] ) [5] , [6] . NiV has also repeatedly caused spill-over events involving hundreds of human cases since 1998 with at least nine recognized occurrences primarily in Bangladesh and India since 2001 ( reviewed [2] ) with the most recent in March 2008 [7] . Several of the more recent NiV outbreaks have had higher rates of acute respiratory distress syndrome in conjunction with encephalitis , epidemiological findings consistent with multiple rounds of person-to-person transmission [8] , higher case fatality rates ( ∼75% ) , and direct transmission of virus from flying foxes to humans via contaminated food has been demonstrated [9] , [10] . In addition to their highly pathogenic nature , the henipaviruses are also distinguished from all other paramyxoviruses by their unusually broad host tropism . Host cell infection by NiV and HeV requires two membrane-anchored envelope glycoproteins; the attachment ( G ) glycoprotein which binds the viral receptor , and the fusion ( F ) glycoprotein which drives virus-host cell membrane merger [11] . The henipavirus G glycoprotein lacks hemagglutinin and neuraminidase activities and the F glycoprotein is a typical class I fusion glycoprotein ( reviewed in [12] ) . The host cell membrane anchored proteins , ephrin-B2 and ephrin-B3 ligands , have been shown to be the receptors employed by the henipaviruses [13] , [14] , [15] , [16] . There are presently no licensed therapeutics available to treat infection caused by the henipaviruses . Recently , we isolated and extensively characterized a neutralizing human monoclonal antibody ( hmAb ) , m102 . 4 , which recognizes the receptor binding domain of the HeV and NiV G glycoproteins . This hmAb potently neutralized both viruses in vitro and maintained its biological activity in vivo suggesting its possible utility as a passive therapeutic modality following henipavirus infection [17] . Here we report the development and characterization of a novel ferret model of acute NiV infection and associated disease as well as conduct the first henipavirus therapeutic antibody trial using the hmAb m102 . 4 . Together , our data demonstrate that NiV-mediated disease in the ferret closely resembles that seen in humans with the presence of both respiratory and neurological disease . We further demonstrate that m102 . 4 is an effective post-exposure therapeutic representing the first antiviral drug candidate showing in vivo efficacy in treating lethal NiV-mediated disease , and it is the first human mAb therapeutic developed and tested for the treatment of henipavirus infection .
In humans , disease resulting from NiV infection can vary in intensity from an acute febrile illness or one progressing to severe central nervous and respiratory disease . Pathological findings show systemic vasculitis , necrotizing alveolitis and meningoencephalitis [18] , [19] . The disease in experimentally infected cats and hamsters is similar [20] , [21]; but in hamsters meninoencephalitis is more prominent , while cats develop an acute respiratory disease [22] . Here , we sought to assess a new ferret model of NiV pathogenesis where our preliminary observations had confirmed susceptibility to NiV infection , with development of systemic vasculitis and involvement of the central nervous and respiratory systems . Ferrets have emerged as a model for several viral respiratory diseases including avian influenza [23] , severe acute respiratory syndrome [24] ) , and morbilliviruses [25] , close relatives of henipaviruses [26] . They offer the combined advantages over either of the aforementioned laboratory animal species of being relatively small mammals , while displaying complex behaviors especially in relation to their handlers that may be used to advantage in clinical assessments . They are however also sufficiently large to enable repeated collection of a wide range of clinical samples throughout the course of an experimental infection , as well as administration of potential therapies in a manner similar and consistent with human medicine . We initiated a NiV minimal infectious dose study ( MID50 ) for the purpose of determining an appropriate challenge dose for subsequent work that would reliably productively infect naïve ferrets . Doses of 50 , 500 , 5 , 000 or 50 , 000 TCID50 were each administered to groups of two ferrets oral-nasally; the most likely route of natural infection . Based on prior experience with NiV infection in cats [22] , [27] , [28] and using similar parameters , we defined infection endpoints in ferrets for the purpose of humane euthanasia; these were used as surrogates for lethality . Endpoints were fever plus signs of rapidly progressing clinical illness including both generalized ( e . g . inappetance , depression ) and localizing ( e . g . dyspnea , neurological signs ) disease signs . Ferrets that received 50 TCID50 ( 1–50 and 2–50 ) , and one animal that received 500 TCID50 ( 3–500 ) remained well throughout the period of observation , did not shed detectable virus or viral RNA , did not seroconvert , and their tissues were normal at post mortem and histological examination . The remaining ferrets developed fever ( 4–7 days post-infection ( dpi ) ) and rapidly progressive clinical illness ( 6–8 dpi ) . Ferrets inoculated with 50 , 000 TCID50 ( 7–50000 and 8–50000 ) were euthanized 6 and 7 dpi , respectively; those inoculated with 5 , 000 TCID50 ( 5–5000 and 6–5000 ) were euthanized 8 and 10 dpi , respectively , and one ferret inoculated with 500 TCID50 ( 4–500 ) was euthanized 9 dpi . Clinical signs in affected ferrets included severe depression , cough , serous nasal discharge , dyspnea and subcutaneous edema of the head ( 8–50000 ) ; severe depression , orthopnea , and expiratory dyspnea ( 7–50000 ) ; severe depression , orthopnea and cutaneous ecchymoses ( 6–5000 ) ; severe depression , vomiting and hypothermia ( 5–5000 ) ; and obtunded with tremor and hind limb paresis ( 4–500 ) . Gross pathological findings comprised varying degrees of subcutaneous hemorrhagic edema of the head and neck , centered upon local lymph nodes in all diseased ferrets; focal raised pin-head hemorrhagic nodules scattered throughout the pulmonary parenchyma; and scattered intra-abdominal petechial hemorrhages . Histopathology of diseased ferrets revealed acute focal necrotizing alveolitis and pulmonary vasculitis , acute glomerular necrosis , focal necrosis of the spleen , and severe diffuse subacute inflammation of the organs and connective tissues of the head and neck . Less commonly , there was mild focal nonsuppurative meningitis ( 4–500 , 5–5000 and 8–50000 ) , focal cystitis ( 4–500 ) , severe acute necrotizing salpingitis ( 4–500 ) , acute focal coagulative necrosis of the adrenal cortex ( 4–500 and 5–5000 ) , and severe acute thyroiditis ( 6–5000 ) . Two ferrets demonstrated neurological signs ( 4–500 , 5–5000 ) and had nonsuppurative meningitis . Syncytia were usually present in lesions and often contained abundant viral antigen ( Figure 1 ) . In mildly affected lymph nodes there was focal mononuclear cell and neutrophilic inflammation of the capsule , accompanied by a zone of subcapsular lymphocyte depletion . In more severely affected lymph nodes , there was severe extensive hemorrhagic and coagulative necrosis , often resulting in the destruction of the entire node with antigen staining mainly in syncytia at the living margin of the necrotic focus ( Figure 1 ) . Clinically healthy ferrets had no gross pathological changes , and no viral antigen or significant histological lesions were found in any tissues . To examine virus dissemination and tissue tropism , NiV N gene RNA was measured in pharyngeal and rectal swabs , blood sampled over the course of infection and in a variety of tissues recovered at necropsy . Among ferrets that developed clinical disease , NiV RNA was detected in the blood ( Figure 2A ) at the time of euthanasia and , in two of these ( 5–5000 , 6–5000 ) , also in the sample collected two days earlier . One ferret ( 3–500 ) that had remained healthy during the period of observation demonstrated a low level of viral RNA in blood 21 dpi . Interestingly , this ferret shared a cage with ferret 4–500 , from which viral RNA was recovered in pharyngeal and rectal swabs 9 dpi suggesting viral shedding . Viral RNA ( albeit some at low level ) was detected in pharyngeal swabs from all ferrets demonstrating clinical disease ( Figure 2A ) and in rectal swabs from two of these ( Figure 2A ) . Notably , NiV RNA was not detected in any tissues from clinically healthy ferrets . For all other animals with clinical signs , significant levels of viral RNA were detected in adrenal , kidney , lung , bronchial lymph node and spleen tissues ( Figure 2B ) . Although at lower relative levels , NiV RNA was also detected in the bladder , liver , ovary or testes , uterine horn and in the brain ( Figure 2B ) . All infected ferrets had NiV RNA in the olfactory lobe of the brain and ferret 4–500 , which showed tremors and hind limb weakness , had the highest level in the occipital lobe . Although real-time PCR can help discern virus distribution in an infected animal , it does not discriminate between infectious and non-infectious genetic material . For these reasons , virus isolation was attempted from samples and tissues with NiV RNA relative expression levels above 10 . NiV was recovered from almost all tissues that had high levels of NiV RNA ( Figure 2B , top panel ) as indicated by the ( + ) above individual data bars . Lungs and lymphoid tissues were sites of extensive virus replication . Virus was also isolated from tissues with lower NiV RNA levels , including the occipital lobe , uterine horn and ovary from ferret 4–500 and the olfactory lobe from ferret 8–50000 ( Figure 2B , bottom panel ) . For ferrets where NiV was isolated from brain , histological lesions included nonsuppurative meningitis . Recently , we described a hmAb , m102 . 4 which engages the receptor-binding site of the viral attachment G glycoprotein , potently neutralizes HeV and NiV in vitro , and retained its biological activity in vivo [17] . For the evaluation of m102 . 4 efficacy in the ferret model two time points were selected , 24 hrs before ( pre- ) or 10 hrs after ( post- ) NiV challenge . Each treatment group contained three ferrets except for the control group ( n = 2 ) . At the indicated time , 50 mg of m102 . 4 was administered via intravenous catheter . One ferret in the post-challenge group ( 28-post ) was given m102 . 4 intraperitoneally due to difficulty in venous catheterization . Control ferrets ( 22-con and 25-con ) were given PBS . Ferrets were challenged oral-nasally with a 5 , 000 TCID50 dose of NiV ( ten-fold the MID50 as determined by the dose-ranging study above ) . By 6 dpi , ferrets 22-con , 25-con and 23-pre were febrile ( >40°C ) and control animals started to demonstrate signs of clinical illness similar to those seen in the initial MID50 study . By 8 dpi the controls were severely depressed , had subcutaneous edema of the head and cutaneous hemorrhages and they were euthanized . By 8 dpi , 24-pre and all ferrets in the post-group were also febrile with variable levels of depression and suppression of play activity; a notable sign of abnormal ferret behavior . At 10 dpi , the temperatures of 23-pre and 27-post had started to fall , and each demonstrated moderate ( 23-pre ) or mild ( 27-post ) edema of the throat . 24-pre and 26-post remained febrile with no localizing signs . By 13 dpi , ferrets 23-pre and 24-pre were depressed and inappetant with cutaneous ecchymoses; ferret 23-pre had marked hind limb paresis and generalized tremor and both animals were euthanized . All other ferrets ( three in the post- and one in the pre-group ) were well and free of any disease signs and remained so until the end of the study ( Table 1 ) . Upon study completion ( 20 dpi ) all surviving animals were euthanized . For the controls and two ferrets in the pre-challenge group that succumbed to infection , gross and microscopic pathology revealed findings similar to those described above in the MID50 study . However , in ferrets 23-pre and 24-pre the frequency of pinpoint hemorrhagic lesions observed in the pulmonary parenchyma was reduced and lesions were much smaller suggesting the disease progression in the respiratory tract had been dampened ( Figure 3 ) , consistent with their survival to 13 dpi . Histopathological findings in controls and in the two diseased pre-challenge group ferrets were similar to those described in the MID50 study . However , for ferret 23-pre , NiV antigen was present in ependymal tissue ( Figure 4A ) and for ferret 24-pre NiV antigen was present within neurons and neuropil of the cerebellum ( Figure 4B ) , further supporting the suggestion that ferrets could be an appropriate model for NiV-mediated meningoencephalitic disease . No significant pathological abnormalities were found in any of the surviving ferrets . Blood , swabs and tissues were tested for the presence of NiV RNA and results are shown in Figure 5 . Ferret 21-pre , the only animal to survive in the pre-challenge group , had detectable levels of viral RNA in the blood 3 dpi ( Figure 5A ) whereas all other samples were negative . All other ferrets had significant viremia by 8 dpi with the highest levels of NiV RNA in the controls . By 13 dpi , viral RNA levels in ferrets 23-pre and 24-pre were similar to those found in controls whereas NiV RNA levels had decreased in ferret 26-post and were undetectable in the other post-challenge group ferrets . Viral RNA levels in oral swabs were highest among the controls and pre-challenge group; however , post-challenge group ferrets shed virus over the course of infection ( Figure 5A ) . NiV RNA levels in rectal swabs were low except in 23-pre and 24-pre ( Figure 5A ) . Together these data demonstrated that all treated ferrets ( except 21-pre ) had significant amounts of systemic and mucosal NiV for at least 10 dpi . Controls and the two diseased pre-group ferrets had similar high levels of viral RNA in all tissues ( Figure 5B ) . Conversely , 21-pre and all post-challenge group ferrets had significantly reduced levels of viral RNA ( Figure 5B ) . Virus isolation was attempted on all samples with detectable NiV RNA ( Figure 5 ) . On day 8 post-infection , NiV was isolated from the blood of one control animal ( 22-con ) , from a pharyngeal swab from the other control animal ( 25-con ) and from a rectal swab from ferret 23-pre . All other blood and swab samples were negative for virus isolation despite the high levels of viral RNA detected . Virus was isolated from the majority of control animal tissues and the two diseased pre-group ferret tissues; whereas , for 21-pre and all post-group ferrets , tissues were negative for NiV . Together , these data demonstrate that treatment with m102 . 4 reduced viral replication and spread , and provided a significant therapeutic benefit leading to the survival of all infected animals in the post-challenge group . To evaluate whether a correlation between m102 . 4 concentration and protection existed , levels of m102 . 4 in serum were measured ( Figure 6 ) . Coincidently , ferret 21-pre had the highest levels of serum m102 . 4 pre-challenge along with the longest-lasting serum levels ( Figure 6A ) and was the only animal from this group that survived . By 3 dpi , all pre-challenge group ferrets had significantly lower m102 . 4 concentrations compared to the post-challenge group ( Figure 6B ) . Importantly , m102 . 4 concentrations on day 3 correlated with survival and may indicate the existence of a critical therapeutic window following virus exposure . By 6 dpi , m102 . 4 levels in all animals had dropped and were the lowest among the post-challenge group . Interestingly , on 6 and 8 dpi , 23-pre had higher serum m102 . 4 levels than those in the post-group and succumbed to disease ( 13 dpi ) whereas 27-post had low levels of m102 . 4 on these days and survived infection . Antibody specific for NiV G was measured in serum post-challenge . NiV-G specific antibodies were detected in controls by 6 dpi ( Figure 7 ) . Ferret 21-pre had high levels of G-specific antibodies 10 dpi which likely reflected m102 . 4 . For all other ferrets , NiV G-specific antibodies were low on 10 dpi with higher amounts of antibody present in the post-challenge group . By 13 dpi , NiV-G specific antibody responses had increased further in all ferrets with the highest levels found in the 27-post and 28-post ferrets . In the present study , we have described m102 . 4 , a potent neutralizing human monoclonal antibody . Its mechanism of action involves binding a single epitope within the receptor binding domain of the henipavirus G glycoprotein and affectively blocking receptor engagement and virus infection . HeV and NiV cause acute and often fatal disease in animals and humans and although virus neutralization escape mutants are unlikely in 7–10 days; we evaluated whether resistant virus was present in ferrets 23-pre and 24-pre . Virus isolation was repeated on all tissues homogenates in the presence and absence of m102 . 4 . Virus was isolated from all samples in the absence of m102 . 4 as before; however , no virus was isolated from tissues when m102 . 4 was present ( data not shown ) . Together , these data demonstrate that the virus circulating in ferrets 23-pre and 24-pre at the time of euthanasia was still highly sensitive to m102 . 4 and no evidence of escape mutants was found . In addition , we also examined m102 . 4 neutralizing activity against different isolates of NiV and HeV , including , NiV-Malaysia , NiV-Bangladesh , HeV-1994 and HeV-Redlands . Importantly , m102 . 4 neutralized all viruses potently: HeV-Redlands was completely neutralized at 1 . 25 µg/ml m102 . 4 and all other isolates were even more sensitive with complete neutralization occurring at 0 . 63 µg/ml m102 . 4 . These data support the potential broad applicability of m102 . 4 as a post-exposure therapeutic option for henipavirus-infected individuals .
A new ferret model of NiV infection and pathogenesis has been developed in the present study . Unlike the cat and hamster models , the ferret model is unique and exhibits both severe respiratory and neurological disease , and generalized vasculitis . NiV-mediated disease in humans has been described as a systemic endothelial infection accompanied by vasculitis , thrombosis , ischaemia and necrosis [18] . Histopathological changes in infected individuals were especially noted in the central nervous system ( CNS ) with widespread presence of viral antigen in neurons and other parenchymal cells in necrotic foci in the CNS and also in endothelial cells of affected blood vessels . Both vasculitis and endothelial infection was also seen in most organs examined . In ferrets , clinical disease included vascular fibrinoid necrosis in multiple organs , necrotizing alveolitis , syncytia of endothelium and alveolar epithelium , and necrotizing lymphadenitis . Histopathological lesions included severe focal necrotizing alveolitis , vasculitis , degeneration of glomerular tufts , and focal necrosis in a wide-range of other tissues . Significant quantities of viral antigen in blood vessel walls and syncytial cells were frequently present . Viral antigen was present in neurons and infectious NiV was isolated from multiple organs including the brain . Several animals also demonstrated significant neurological disease . A widespread presence of NiV antigens in neurons and other parenchymal cells in the CNS was not as extensive as in NiV-infected ferrets as compared to infected humans , and this could relate to the length of disease course between these experimental versus natural infections . We speculate that a more widespread infection in the CNS would have developed if animals were allowed to succumb to disease without euthanasia . Overall , NiV-mediated disease in ferrets has all the hallmarks of NiV-infected humans and represents a novel relevant animal model . Recent vaccine studies have demonstrated that henipavirus G-specific antibodies are critical for protection from disease [22] , [28] . Evidence of passive protection against NiV , and more recently HeV , challenge using hamster polyclonal antiserum or murine mAbs reactive to NiV glycoproteins was provided in the hamster [29] , [30] , [31] . However , in those prior studies , both the challenge virus and antibodies were administered by intraperitoneal injection either simultaneously or immediately before or following challenge . Here we examined a neutralizing fully-human mAb , with the potential for human use , in a consistently susceptible animal model where challenge and drug delivery mimicked a potential real-life scenario . The hmAb m102 . 4 is an ideal therapeutic , targeting a critical functional domain of the G glycoprotein , and is a fully human IgG . We achieved full protection from NiV-mediated disease in ferrets who received a single infusion of antibody 10 hrs post-challenge and protection in one animal treated prior to challenge . Ferret 21-pre was the only survivor in the pre-group and it had the highest pre-challenge level of m102 . 4 and the best antibody longevity . Uniquely , fever , clinical signs and virus were not detected at any time in this animal , with the exception of an extremely low level viremia 3 dpi . These findings suggest that the combination of an initial high level of antibody and subsequent longevity may provide protection . For the remaining ferrets , higher levels of m102 . 4 on day 3 post-challenge correlated with protection from disease whereas levels thereafter did not , potentially highlighting the first 3 days following infection as an important therapeutic window . On day 3 , lower m102 . 4 concentrations in the pre-group may not have been sufficient to neutralize all virus , allowing the initiation of infection . In the post-group the higher concentrations 3 dpi may have sufficiently neutralized enough virus and/or prevented widespread dissemination of virus , providing a window for the animal to mount an effective immune response . In the post-challenge group , the host immune system may have had an opportunity ( 10 hrs ) to initiate a primary response before antibody-mediated depletion of virus . In the pre-challenge group , the antibody-virus complexes most likely formed almost immediately following challenge and the primary or innate response may have been dampened . As m102 . 4 levels decreased over time and virus replication initiated , the post-challenge group may have had a more potent secondary response allowing survival . Indeed , the animals in the pre-challenge group developed severe disease 7 days after m102 . 4 levels decreased , a timeline similar to the disease progression in naive animals . Whether primary or secondary , the antibody responses in the post-challenge group appeared earlier ( 10 dpi ) and were more robust . The survival of all ferrets in the post-challenge group represents a major step forward towards a viable passive antiviral therapeutic for treating henipavirus infection . Since m102 . 4 targets the receptor-binding domain of G [32] and is a conformation-dependent antibody , we believe it will maintain efficacy in vivo during the course of its application in an infected host . Given the acute nature of henipavirus infection and its receptor specificity , mutation and escape from m102 . 4 sensitivity seems unlikely . Mutations generating m102 . 4 neutralization escape variant viruses would likely also acquire defects in receptor binding and fitness . Our data support this notion as no escape mutants were found in treated animals that succumbed to disease . Further examination of m102 . 4 has confirmed its potent cross-reactive neutralization activity against the NiV-Malaysia [33] , the original HeV-1994 [34] , the recent HeV-Redlands [5] and NiV-Bangladesh isolates [35] . Examination of m102 . 4 effectiveness in treating animals exposed to these henipavirus isolates will be important . In future studies , it will also be critical to increase the number of animals studied and further explore the therapeutic window post-challenge . Although 10 hrs is relevant to a known exposure , such as a laboratory accident , it will be important to define the limits of the therapeutic window . These data represent the first human antibody therapy successfully evaluated in vivo for prevention of lethal henipavirus infection and strongly suggests m102 . 4 may prove to be a useful therapeutic used to treat disease in people caused by these important emerging pathogens .
Eight adult ( 1–2 year old ) ferrets were used for the NiV minimal infectious dose study and eight adult ferrets were used for the in vivo m102 . 4 efficacy study . All animal studies were endorsed by the CSIRO , AAHL , Animal Ethics Committee . Ferrets were housed in pairs in cages in a BSL4 room , fed twice daily with a complete premium dry food and provided with water ad lib . For implantation of temperature transponders , NiV challenge , administration of m102 . 4 and specimen collection , ferrets were anaesthetised by intramuscular injection of ketamine ( Ketamil , Troy Laboratories , New South Wales ( NSW ) , AU ) and medetomidine ( Domitor , Pfizer Animal Health , NSW , AU ) at dose rates of approximately 3 mg/kg and 0 . 03 mg/kg , respectively . For anesthetic reversal , atimepazole ( Antisedan; Novartis , Pendle Hill , NSW , AU ) was given intramuscularly at 50% the dose of medetomidine . Following challenge , ferrets were assessed daily and scored out of 10 for a range of clinical observations , including alertness , playfulness , curiosity , depression , food consumption , feces production and respiration rate . Real-time monitoring of body temperature for all animals was done using transponders implanted subcutaneously [28] . Rectal temperatures and body weights were recorded on sampling days . Once ferrets were clinically assessed to be exhibiting signs of severe disease from which natural recovery was considered unlikely , animals were euthanized by intravenous injection of sodium pentabarbitone . Ferrets were inoculated oronasally with a low passage NiV isolate ( NiV-Malaysia; EUKK 19817; stock virus titer 4 . 3×106 TCID50/ml ) [36] . For the MID50 study NiV doses were: 5×104 , 5×103 , 5×102 and 50 TCID50 ( 2 ferrets per dose ) . Based on the outcome of the MID50 study , for the m102 . 4 efficacy study all ferrets were challenged with 5×103 TCID50 . The m102 . 4 hmAb was prepared as previously described [17] . Ferrets were anaesthetised and a 20-g intravenous catheter placed in the left jugular vein . m102 . 4 was administered via catheter by slow infusion over 4 minutes to 6 ferrets ( 50 mg per ferret ) ; 3 ferrets received antibody 24 hours before NiV challenge and 3 ferrets received antibody 10 hours after challenge . Two ferrets received PBS , one at 24 hours before challenge and the other at 10 hours after challenge . Catheters were withdrawn and animals were allowed to recover from the anaesthesia . Blood , oral swabs and rectal swabs were collected 6 , 8 , 10 , 13 and 20 or 21 days post-infection ( dpi ) for the purpose of assisting in establishing the infection status of each animal . For the m102 . 4 study an additional sample was collected during incubation period ( 3 dpi ) . Two aliquots of whole blood were removed from each sample and serum was collected and aliquoted . One aliquot of whole blood was added to RiboPure lysis buffer ( Ambion Inc . Austin , TX , USA ) containing sodium acetate . Duplicate swabs were placed in AVL viral lysis buffer ( Qiagen Pty Ltd , Clifton Hill , Victoria , AU ) or PBS . After euthanasia , tissue samples were collected aseptically from lung ( apical and diaphragmatic lobes ) , brain ( olfactory and occipital lobes ) , heart , bronchial lymph nodes , spleen , liver , kidney , bladder , and adrenal gland . In females , the uterine horn and ovaries were collected and in males the testes were collected . Tissues were either fixed in 10% neutral buffered formalin , submerged in RLT lysis buffer ( Qiagen Pty Ltd ) containing β-mercaptethanol and 1 mm stainless steel beads ( BioSpec Products Inc . , Bartlesville , OK , USA ) or submerged in PBS containing 1 mm stainless steel beads . For the MID50 study , all specimens were placed at −80°C . For the m102 . 4 efficacy study , all blood , swab and tissue samples were processed immediately for RNA extraction and duplicate samples in PBS were stored at −80°C for virus isolation . RNA was purified from blood cells using the RiboPure-Blood kit ( Ambion Inc . ) , from swabs using the QIAamp viral RNA kit ( Qiagen Pty Ltd ) and from tissues using the RNeasy Mini kit ( Qiagen Pty Ltd ) . Prior to processing , tissues were homogenized for 2 cycles of 30 sec using a Mini-Bead Beater ( Biospec Products Inc . ) and centrifuged to pellet debris . Taqman PCR assays were preformed as previously described [28] . Samples were amplified in a GeneAmp 7500 Sequence Detection System ( Applied Biosystems , Foster City , CA ) . Ct values representing NiV nucleocapsid ( N ) gene expression were analyzed and data were recorded as relative NiV N gene expression levels . Virus isolation was performed as previously described [28] and only attempted from specimens positive for NiV RNA . Tissues were processed by routine histological methods and sections of tissue were stained with hematoxylin and eosin to examine histopathological changes . Separate sections were stained by immunohistochemical techniques as previously described [20] using a rabbit polyclonal antiserum against NiV nucleoprotein . Multiplexed microsphere assays were performed as previously described [37] . A Luminex® 100 IS™ machine and MiraiBio software ( MiraiBio Group , South San Francisco , CA ) were utilized for all assays: Master Plex CT v1 . 0 for data acquisition and MasterPlex QT v 2 . 0 for data analysis . All samples were assayed simultaneously and concentrations were extrapolated from a standard curve using non-linear regression analysis ( GraphPad Software , San Diego , CA ) . | Nipah virus and Hendra virus are closely related and highly pathogenic zoonoses whose primary natural reservoirs are several species of Pteropus fruit bats . Both Nipah and Hendra viruses can cause severe and often fatal disease in a variety of mammalian hosts , including humans . The henipaviruses are categorized as biosafety level 4 ( BSL-4 ) agents , which has limited the development of animal models and the testing of potential therapeutics and vaccine countermeasures . We show here a new ferret model of Nipah virus pathogenesis in which the underlying pathology closely mirrors the illness seen in Nipah virus-infected humans , including both respiratory and neurological disease . We also show that m102 . 4 , a cross-reactive neutralizing human monoclonal antibody that targets the viral attachment glycoprotein , completely protected ferrets from disease when given ten hours after a lethal Nipah virus challenge . This study is the first successful and viable post-exposure passive antibody therapy for Nipah virus using a human monoclonal antibody . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"virology/new",
"therapies,",
"including",
"antivirals",
"and",
"immunotherapy",
"virology/animal",
"models",
"of",
"infection",
"virology/emerging",
"viral",
"diseases"
]
| 2009 | A Neutralizing Human Monoclonal Antibody Protects against Lethal Disease in a New Ferret Model of Acute Nipah Virus Infection |
The human herpes virus Epstein-Barr virus ( EBV ) latently infects and drives the proliferation of B lymphocytes in vitro and is associated with several forms of lymphoma and carcinoma in vivo . The virus encodes ~30 miRNAs in the BART region , the function of most of which remains elusive . Here we have used a new mouse xenograft model of EBV driven carcinomagenesis to demonstrate that the BART miRNAs potentiate tumor growth and development in vivo . No effect was seen on invasion or metastasis , and the growth promoting activity was not seen in vitro . In vivo tumor growth was not associated with the expression of specific BART miRNAs but with up regulation of all the BART miRNAs , consistent with previous observations that all the BART miRNAs are highly expressed in all of the EBV associated cancers . Based on these observations , we suggest that deregulated expression of the BART miRNAs potentiates tumor growth and represents a general mechanism behind EBV associated oncogenesis .
Epstein-Barr virus ( EBV ) is a ubiquitous human herpes virus . It infects virtually every human being for life and in the overwhelming number of people this persistent infection is benign [1] . The virus achieves this by entering into a quiescent latent state within circulating , resting , memory B lymphocytes where no viral proteins are expressed [2 , 3] . However , in order to achieve this state the virus needs to latently infect naïve B lymphocytes and drive them to become activated proliferating lymphoblasts ( growth transcription program\Latency III ) so that they can then differentiate through the germinal center to become latently infected , resting memory B cells . This ability to drive B cell proliferation has long been believed to explain why EBV is associated with several forms of cancer , including lymphomas ( Burkitt’s ( BL ) and Hodgkin’s ( HD ) ) and carcinomas ( nasopharyngeal ( NPC ) and gastric ( GaCa ) ) [1 , 4] . However , a major inconsistency with this idea was uncovered when detailed analysis of viral latent protein expression was performed on infected cells in vivo and in tumor cells . These studies revealed that the tumors do not express the full panoply of potentially oncogenic , growth promoting , EBV latent proteins found in the growth program . Rather , the EBV gene expression profiles seen in the tumors reflects those found at specific stages of B cell infection in vivo [4] . This finding has led to the proposal that the tumors arise directly from these infected B cell types . In the most extreme case ( BL ) , only one viral protein is expressed , EBNA1 [5 , 6] which is the gene expression pattern characteristic of latently infected memory B cells [7] when they divide . These counterintuitive observations strongly imply that the virus has evolved to minimize the oncogenic risk posed by the growth program/Latency III latent proteins , which otherwise would threaten the host within which the virus persists . On the other hand , it is well known that the EBV episome , or plasmids derived from it , will be rapidly lost if there is no selective advantage to their retention [8–11] . The persistence of the viral episome within the tumors therefore implies that the virus must be contributing to the growth and/or survival of the tumors [8 , 12–17] . Since most of the growth promoting EBV latent proteins are absent , the question arises as to what the virus is contributing to the tumors that causes the viral episome to be retained . Recently , attention has become focused on abundant non-coding EBV RNAs [18–23] . These include ~40 miRNAs , four of which are encoded from the BHRF region and the remainder from the BART region . We have shown that all of the BART miRNAs are highly expressed in all of the tumor types , especially the carcinomas [22 , 24] . This includes a subset of the BART miRNAs whose expression in vivo is specifically associated only with the Latency III growth program ( Latency III associated BART miRNAs ) and thus , would not be expected to be associated with the viral latency programs found in the tumors ( Latency I or II ) . To date , this represents the only example of inappropriate expression of Latency III growth program genes in the tumors and suggests that these miRNAs may play a role in tumor development . Although the functions of most of the BART miRNAs remains unresolved several in vitro studies have suggested that they may contribute to oncogenesis . BL-derived cell lines with constitutive expression of BART miRNAs are able to inhibit apoptosis induced by the loss of EBV in vitro , implicating a survival role for BART miRNAs in BLs [25] . BART 5 represses the p53 up-regulated modulator of apoptosis ( PUMA ) , thereby protecting EBV-infected cells from virally-induced apoptosis and in addition high expression of BART 5 correlates with the low abundance of PUMA in NPC tissues [26] . BART 13* likely plays an anti-apoptotic role by targeting the Wnt-signalling enhancer CAPRIN2 [27] , Bart 3* suppresses the DICE1 tumor suppressor to promote cellular growth [28] and BART Cluster 1 miRNAs are able to suppress the Bcl-2 interacting mediator of cell death ( Bim ) , thus protecting cells from apoptosis [29] . However , until now no data in support of a role for the EBV BART miRNAs in tumorigenesis in vivo has been presented . Since it is not possible to study the tumors in situ in humans , we developed a mouse model of one ( NPC ) [30] . In this paper we use this in vivo model to demonstrate that the miRNAs confer a significant growth advantage to EBV associated tumors in vivo .
We have previously reported an orthotopic mouse model which faithfully recapitulates locally invasive and metastatic EBV-positive nasopharyngeal carcinoma ( NPC ) [30] . In this model , luciferase-tagged C666–1 cells were injected into the nasopharyngeal epithelium of the highly immune deficient mouse strain NOD . Cg-PrkdcscidIl2rgtm1Wjl/SzJ ( NSG ) . To gain insight into the possible role of EBV BART miRNAs in tumorigenesis we wished to ascertain if tumor growth and/or metastasis in this model was associated with preferential expression of specific BART miRNAs . Therefore , we examined the BART miRNA expression profiles of C666–1 cells before and after inoculation and growth as tumors in the NSG mouse model and then again after the tumors had been explanted and grown in vitro . A complete description of the parental lines , tumors , and tumor explant cultures developed and used in this study is given in S1 and S2 Tables . To profile the BART miRNAs we employed a PCR based assay that we have characterized in detail previously [24] . This assay will detect ≤10 copies of each mRNA with a linear response up to ≥108 copies . It has the advantage that it allows the quantitative profiling of the expression of a large number of miRNAs from small tissue samples that otherwise would be impossible to study , such as human tissue biopsies and the moue tumors studied here . Fig . 1A and B show examples of C666–1 derived tumors grown in the NSG mouse model . The miRNA profiles of the cell lines before , during and after growth as tumors are shown in Fig . 2A . We did not observe specific changes in individual miRNAs rather all of the BART miRNAs were significantly up regulated when the cells grew as tumors in vivo compared to the parental lines ( p<0 . 001 ) . When the tumors were explanted for culture in vitro to become cell lines , the miRNA expression declined again to that of the parental line . However , this decline was fully reversible since upon reinjection back into the mice , the miRNA level again increased . In all , this process was reiterated three times with the same result , up regulation in vivo and down regulation in vitro ( not shown ) . This suggests that increased expression of the BART miRNAs may confer a growth advantage to the tumor cells in vivo . Two of the three BHRF1 miRNAs also showed elevated copy numbers in the tumors although their expression levels were modest when compared to the BART miRNAs ( not shown ) . Metastasis is generally characterized by expression of the signature protein Snail a key transcriptional factor in epithelial mesenchymal transition that is known to promote metastasis [31] . As expected , Snail expression was greatly increased in the C666–1 metastases compared to the primary tumors ( Fig . 3A ) . However , when we compared the BART miRNA expression profiles of primary tumors with metastasis they were indistinguishable ( Fig . 3B ) . Indeed we have developed a highly metastatic derivative of the C666–1 tumor with a 100% metastasis rate in mice and even these cells showed no variation in BART miRNA expression ( not shown ) . Thus , this analysis did not detect a correlation between EBV BART miRNA expression and metastasis . We wished to confirm these findings in a second carcinoma model therefore we chose to study the gastric carcinoma ( GaCa ) cell line AGS-BX1 . Successful engraftment of immunosuppressed mice with this cell line has not been reported previously . Consistent with this , mice injected subcutaneously with 107 AGS-BX1 cells failed to develop tumors ( not shown ) . Interestingly though , 4 out of 10 mice injected in the nasopharyngeal epithelium ( I . N . ) grew tumors , with metastatic dissemination being observed in 3 of the mice ( Fig . 1 C and D ) suggesting that this is also a good model to study GaCa in vivo . We profiled the BART miRNA expression levels in the parental line , tumors and explants and observed similar results to those seen with the NPC model ( Fig . 2B ) . Although not as large as the NPC line , the increase in BART miRNA expression in the tumors was nevertheless substantial when compared to the parental line ( p<0 . 001 ) . However , again no significant difference was seen between the primaries and metastases ( Fig . 3 C and D ) and the expression level decreased when the tumors were grown in vitro . The extent of reduction was not as dramatic as that seen with NPC and had not returned to the levels seen in the parental line at the time the cells were tested . This likely reflects the fact that the AGS-BX1 explant lines were grown for a much shorter time in vitro then the NPC lines and may not have had sufficient time to fully reduce their expression level . To discover if the up regulation of the BART miRNAs was specific to the viral miRNAs or represented a more general phenomenon we first examined the expression of three cellular miRNAs mir-9 , mir-34a and mir-26a . As seen in Figs . 4 A and C all three cellular miRNAs tested were expressed at similar levels in the parental lines , tumors and in the explanted tumor cells grown in vitro . A trend was observed as with the BART miRNAs in that levels were slightly higher in the tumors and tended to decrease again in the explants . However , these changes were not consistent ( mir-9 was mostly highly expressed in the NPC explants ) and did not always achieve statistical significance . In particular changes in the GaCa cells never achieved significance . Thus , unlike the BART miRNAs , the cellular miRNA increases were small and not reproducible between different cell types . This study confirms that the up regulation we have observed in BART miRNA expression in vivo was not a consequence of a global increase in the production of all miRNAs . We also tested mRNA levels for two cellular genes that are associated with epithelial mesenchymal transition , CDH1 and Snail [32 , 33] ( Fig . 4 B and D ) . While the levels of CDH1 remained unchanged , Snail levels actually decreased in the tumors compared to the parental or explanted tumor cells grown in vitro for both NPC and GaCa . These results demonstrate that the increased expression of the BART miRNAs was not a consequence of a general up regulation in gene expression . Lastly , we checked the mRNA levels of EB viral latent genes including the small viral RNA EBER1 ( Fig . 4 B and D ) . The levels of all three were higher in the tumors for both the NPC and the GaCa cells when compared to the parental lines . However , the levels of the EBNA1 and LMP1 transcripts were strikingly up regulated for the NPC tumors suggesting that these latent proteins might also play a role in NPC tumor growth . We conclude that there is an increase in the expression of the BART miRNAs when EBV positive carcinoma cells are grown as tumors in vivo and this is reversed upon re-culture in vitro . However , we saw no consistent changes in BART expression when tumors underwent metastasis . This suggests that the BART miRNAs may confer a significant growth advantage to EBV positive tumor cells in vivo that is not reiterated in vitro . To test whether the up regulation of BART miRNAs is tissue specific , i . e . restricted to the carcinomas , we evaluated their expression in a B cell lymphoma . Single cell suspensions of the luciferase tagged EBV-positive BL ( BL36 ) cells could not be successfully injected into the nasopharyngeal epithelial tissue of NSG mice , therefore we injected them intravenously into the tail vein . Lymphoma developed in all injected animals ( Fig . 5A and B ) . For a detailed description and account of the tumors used in this study see S1 and S2 Tables . Upon profiling the BART miRNA expression pattern we saw a similar result to that obtained with the carcinomas . The expression level of all the BART miRNAs was up regulated ( p<0 . 001 ) , when compared with the parental line ( Fig . 5C ) . As with the carcinomas this effect was reversible when the tumors were re-cultured in vitro . We have reported here an increased expression of the BART miNAs in EBV associated tumors in vivo that is not due to a general increase in cellular transcription but may be associated with increases in expression levels of other viral latent genes . To assess the contribution of the BART miRNAs alone to tumor development we took advantage of an EBV negative derivative ( AGS ) of the AGS-BX1 GaCa cell line . Luciferase tagged AGS cells were transfected with either an empty oriP/EBNA1 vector ( AGS-EBNA1-EMPTY ) or one that expresses all of the BART miRNAs ( AGS-EBNA1-BART ) [25] . These cells were then inoculated into the nasopharynx of NSG mice and tumor growth and metastasis in vivo were monitored ( Fig . 6 ) . After 60 days , 2 of 5 control mice had developed small but detectable tumors . Strikingly however , all 5 mice with AGS-EBNA1-BART ( 100% incidence ) had malignancies , indicating that the BART miRNAs promote tumor formation . Furthermore , the BART-expressing tumors appeared to be more aggressive , as these mice deteriorated rapidly , developing large tumors and requiring sacrifice beginning at day 74 . In contrast , by this time , still only two control mice showed detectable tumors that remained small and the mice appeared healthy . Thus , BART miRNAs enhanced both the rate of tumor formation ( p = 0 . 03 by Fishers exact test at day 74 ) and tumor progression/fatality ( p = 0 . 03 by Fishers exact test at day 88 ) . Kaplan Meier analysis of the complete data set combined from two such experiments confirmed that the AGS-EBNA1-BART mice exhibited a significantly higher overall level and rate of mortality relative to the AGS-EBNA1-EMPTY mice ( Fig . 7A , p = 0 . 017 ) . The more aggressive nature of the BART+ tumors was confirmed by measurement of the tumor burden in the two populations of mice . The combined data from two experiments is shown in Fig . 7B . No differences in tumor burden ( measured as luciferase radiance ) were evident between the two groups on day 60 , but subsequently differences became apparent as the tumors progressed . By day 74 , when most of the AGS-EBNA1-BART mice needed to be sacrificed , the tumor burden was approximately 10-fold higher than that of the AGS-EBNA1-EMPTY ( ~1 . 4×106 p/s/cm2/sr in BART+ vs 1 . 5×105 p/s/cm2/sr in EMPTY ) , suggesting that the BART miRNAs promote tumor growth in vivo . By taking measurements of luciferase emission from the same tumors over time it was possible to estimate their relative growth rates . This estimate suggested that the AGS-EBNA1-BART tumors were growing approximately twice as fast as the AGS-EBNA1-EMPTY tumors . We conclude that the BART miRNAs strongly promote tumor growth in vivo . When we measured the BART miRNA expression profiles in the parental AGS-EBNA1-BART line and the tumors we observed the same effect as with the EBV positive tumor lines . We did not observe differential expression of the BART miRNAs rather all of the BART miRNAs were elevated in the in vivo tumors ( Fig . 8A ) . For a list of all tumors and cells used for profiling see S1 and S2 Tables . This result confirms that BART miRNA up regulation in vivo is specific and not simply an indirect consequence of a general increase in viral latent gene expression . Insight into the mechanism by which the BART miRNAs potentiated tumor growth was provided by in vitro analysis of AGS-EBNA1-BART and AGS-EBNA1-EMPTY tumors immediately after explant . Both AGS-EBNA1-EMPTY and AGS-EBNA1-BART parental lines grew at identical rates in culture prior to inoculation ( Fig . 8B ) . However , after explant the AGS-EBNA1-BART tumor cells grew significantly faster ( ~60% faster ) than the AGS-EBNA1-EMPTY tumor cells ( Fig . 8C ) . AGS-EBNA1-BART tumors cells also showed a slightly higher rate of colony formation after explantation ( S1A and B Fig . ) . Surprisingly , in light of recent studies on the effects of BART miRNAs on AGS cells in vitro [29] we observed no difference in the levels of apoptosis between the BART positive and BART negative parental lines when treated with the DNA damaging agent etoposide ( S1C Fig . ) . We did observe a small decrease in sensitivity for the AGS-EBNA1-BART tumor explants when compared to AGS-EBNA1-EMPTY tumors . ( 8% specific reduction in apoptotic cells ) ( S1D Fig . ) Although the differences in colony formation and apoptosis were statistically significant they were small suggesting that the main effect of the BART miRNAs is to provide a growth advantage to the tumors that is seen in vivo and detected upon immediate culture of the explants but lost after long term in vitro culture These data suggests that the up regulated expression of the BART miRNAs confers a selective advantage to tumor growth in vivo that is not seen in vitro . Furthermore , they demonstrate a direct link between the up regulation of the BART miRNAs , that we have consistently observed with tumors in vivo , and enhanced tumor growth . Plasmids constructed from an EBV oriP/EBNA1 vector do not integrate into the host genome when transfected into cells , but instead persist episomally [11] . However , these episomes are rapidly lost in the absence of selection pressure for their retention [9 , 11 , 34] . In the presence of drug selection the parental lines maintained ~450–500 copies of the EBNA1-BART and EBNA1-EMPTY plasmids per cell and as expected , both plasmids were lost when the lines were cultured in the absence of drug selection ( not shown ) . This is consistent with our conclusion that the BART miRNAs do not convey a growth advantage to cells in vitro . Analysis of the plasmid copy numbers from a collection of the tumors ( Table 1 ) revealed that the plasmid had been lost from the AGS EBNA1-EMPTY tumors ( n = 4 ) , which had an average copy number of less than 1 per cell ( 0 . 1±0 . 1 episomes per cell ) . However , the AGS EBNA1-BART tumors ( n = 6 ) , retained an average of 8 . 9±3 . 5 episome copies per cell ( p = <0 . 001 ) which were only lost when the tumors were explanted and grown again in culture . Thus , the lack of drug selection in vivo , caused a precipitous drop in plasmid copy number for both cell types resulting in the loss of all the EBNA1-EMPTY plasmids whereas the EBNA1-BART plasmids were stably retained at around 5–10 copies per cell . In parallel , the level of BART miRNA expression increased . Therefore , amplification of the episome copy number cannot explain the increased expression of the miRNAs in the tumors . This result demonstrates that expression of the BART miRNAs confers a selective advantage to the tumor cells in vivo that alone is sufficient to ensure retention of the plasmid that expresses them and confirms that this advantage applies to in vivo but not in vitro growth . We conclude that the EBV BART miRNAs do not provide a detectable growth or survival advantage when the cells expressing them are grown in vitro . However , in vivo they provide a pronounced advantage to the tumors specifically causing them to be seeded more efficiently and grow faster and more aggressively . Next we asked if expression of the EBV miRNAs in AGS cells would provide for a higher rate of metastasis . For mice receiving AGS-EBNA1-EMPTY cells , 4 out of 6 mice with tumors ( 66 . 7% ) developed metastases ( Fig . 9A ) , whereas only 3 out of 8 mice with AGS-EBNA1-BART tumors ( 37 . 5% ) developed metastases . Although these numbers do not achieve statistical significance , they are consistent with a trend that the BART miRNAs do not exacerbate metastasis and might actually impede it . Lastly , we asked if the BART miRNAs could confer a more invasive phenotype to the tumor cells , as cell invasion is a critical step in metastasis . To test this we performed trans-well Matrigel invasion assays comparing AGS-EBNA1-EMPTY cells with AGS-EBNA1-BART cells . In this assay , the ability of cells to invade through an extracellular matrix of a Matrigel-coated porous membrane in response to chemoattractants is assessed . After a 24 hour incubation , the AGS-EBNA1-BART cells demonstrated a slightly less invasive phenotype than did the AGS-EBNA1-EMPTY cells , as the absolute cell numbers of AGS-EBNA1-EMPTY on the trans-side ( migrated cells ) were around 2-fold higher than that of AGS-EBNA1-BART ( Fig . 9B and C ) , suggesting that EBV BART miRNAs may also have a negative effect on invasion . We conclude that our experiments provide no evidence for BART miRNAs contributing positively to invasion and/or metastasis .
In this paper we have demonstrated that the EBV encoded BART miRNAs confer a selective growth advantage to EBV positive tumor cells in vivo . This growth promotion occurs in parallel with an up regulation in the expression of all the BARTs . Furthermore , elevated expression of the BART miRNAs in vivo was observed for every EBV positive tumor type we tested . This suggests that the BART miRNAs confer a growth advantage to all EBV positive tumors . The lack of a readily accessible and manipulable animal model of EBV infection and tumorigenesis means that most studies on the virus’s biology must be conducted in vitro . Thus to date , all studies on BART miRNA functions have been performed in vitro and have yet to be verified or shown to be biologically meaningful in an in vivo setting . Lacking evidence from humans , mouse models can provide an alternate approach to investigate the role of EBV encoded genes in vivo . Inoculation of immunocompromised mice ( e . g . severe combined immune deficiency or SCID mice ) with EBV-positive tumor-derived cell lines has been extensively used to study the role of the virus in tumorigenesis [30 , 35–41] . However , these models rarely , if ever , recapitulate key aspects of the tumors behavior , such as invasion and metastasis . We have previously described a mouse model that accurately reproduces locally invasive and metastatic EBV-positive carcinoma [30] and have applied this model here to study the role of the EBV BART miRNAs in tumor development . We have shown that expression of the BART miRNAs resulted in more efficient tumor seeding , larger tumors and higher and more rapid mortality . Indeed the tumor burden in the BART+ mouse group was ~10 times greater after 74 days of growth in vivo than with the BART- group . Furthermore , the newly explanted tumors grew 60% faster in culture if they expressed the BART miRNAs , an increase in growth rate sufficient to account for the differences seen in vivo . However , this effect was not sustained in long term cultures . Although we cannot definitely rule out other activities for the miRNAs in tumor development we saw only minimal changes in other functions we assayed including colony formation and resistance to apoptosis . For example , we confirmed the work of others that the BART miRNAs had an anti-apoptotic effect [25–29] , However , this effect was very modest in our system and only detectable after in vivo growth . Contrary to previous findings [29] we did not observe any effect of the BART miRNAs on apoptosis when expressed in AGS carcinoma cells in vitro . Independent proof that expression of the BART miRNAs conferred a selective growth advantage in vivo was provided by the observations on plasmid retention . OriP based plasmids do not integrate , but persist as episomes that are lost unless selective pressure is applied for their retention [11] . The retention of the EBV episome has been interpreted as compelling evidence that the virus is involved in the development in EBV positive tumors [12–15 , 17] . Thus our observation that the BART expressing oriP plasmids ( EBNA1-BART ) were similarly retained in the tumor cells in vivo , whereas the empty vector ( EBNA1-EMPTY ) was lost , supports our conclusion that the BART miRNAs contribute to tumor growth . It follows that expression of the BART miRNAs is responsible in part for the retention of EBV in human tumors . Confirmation that our observation was an in vivo phenomenon came from the finding that both plasmids were lost from the cells in culture . These plasmids were not only selectively retained in vivo , but the BART miRNAs were expressed in the tumors at a much higher level than in the parental line grown in vitro , demonstrating a direct link between high BART copy number and rapid tumor growth . It is interesting to note that the EBNA1-EMPTY plasmid was lost in vivo , even though it expresses the viral tethering protein EBNA1 and we did not see an increase in EBNA1 expression when the AGS-EBNA1-BART in vitro lines grew as tumors in the mice ( not shown ) Together these observations suggest that EBNA1 does not provide a detectable advantage to tumor growth in our system . This raises the provocative question of whether it will be possible to develop a therapy that can be applied to all EBV tumors , based on ridding the tumors of the viral episome and therefore the virus , by silencing the BART miRNAs . It is known that the BART miRNAs are not essential for the in vitro transformation and growth of B cells [42 , 43] . This is consistent with our conclusion that the growth enhancement function of the BARTs only applies in vivo . However , deciphering the BART miRNAs responsible for the effect we have observed and the target genes of those miRNAs will be extremely challenging . Our observation that the copy number of all the BART miRNAs increases in vivo for all the tumor types we have studied here and our previous finding that all of the BART miRNAs are highly expressed in biopsies from all types of EBV associated tumors raises the possibility that all of the BART miRNAs may be contributing to tumor growth . Despite the extensive effort of many laboratories there remain only a few well characterized targets of the BART miRNAs [44] . Of these anti-apoptotic targets predominate however , as stated above we have been unable to confirm previous studies suggesting a role in apoptosis resistance in AGS carcinoma cells in vitro [29] . It is likely that the gene targets of the BART miRNAs are linked to specific in vivo growth functions of both normal and malignant cells . One possible candidate is the difference in geometry between in vivo ( 3-dimensional ) and in vitro ( 2-dimensional ) growth . For example , it is possible that the BART miRNAs regulate signaling pathways that control tumor hypoxia in vivo , which could result in activation of a broad array of mitogenic , pro-invasive and pro-angiogenic genes [45–49] . It is also possible that the growth advantage provided by the miRNAs is a function of the interaction of the tumor cells with the surrounding environment in vivo . For example , the miRNAs may induce the tumor cells to express factors or cell surface changes that elicit growth promoting signals from the surrounding murine milieu in the form of soluble factors such as cytokines and/or recruited cells that potentiate tumor growth . We have also seen no evidence that the BART miRNAs contribute to invasion or metastasis . Indeed we have observed a trend whereby the miRNAs may actually impede these processes since the rate of metastasis was lower for the BART+ tumors . This result could have arisen because the BART+ tumors grow more rapidly and may kill the mice before the tumors have time to metastasize . However , we also observed that the BART miRNAs have a negative impact on invasion , an important corollary of metastasis . It may simply be that the increased proliferation rate driven by the BART miRNAs marginally diverts the cellular metabolism away from the processes required for invasion and metastasis . One possible concern with our studies is the high levels of expression we see for the BART miRNAs in tumors . We have previously estimated the BART miRNA copy numbers in biopsy material for the three tumor types studied here [50] . Even taking into account the presence of non-tumor cells in the biopsies and imprecision caused by technical difficulties in recovering the miRNAs from small biopsy samples it is clear that the mouse tumors are expressing the miRNAs at levels at least 10 fold higher than in the biopsies . However , the process driving this up regulation in the mice is physiologic , not for example an artifact of ectopic expression . This would suggest that there may be constraints on the miRNA copy level in the human host that are not imposed in the highly immune incompetent NSG mouse . One question that arises is with respect to the in vivo specificity of the up regulation of the BART miRNAs . Is it conceivable that if we could artificially drive the copy numbers high enough in vitro , we would see an effect on growth ? However , in this scenario it is difficult to explain why the copy numbers are reduced again upon culturing the tumors in vitro since any growth advantage due to higher BART copy number should ensure their maintenance . This strongly argues that the selective growth advantage provided by the elevation of the BART miRNAs in vivo is not sustained in vitro . In summary , therefore , we conclude that the BART miRNAs provide a significant growth advantage to infected tumor cells growing in vivo . This effect may be a combination of growth promoting and survival functions provided by the miRNAs .
The cell lines used in this study were: NPC line C666–1; BL cell line BL36 ( gift of Dr . Jeff Sample ) ; GaCa cell lines AGS and AGS-BX1 ( gift of Dr . Lindsey Hutt-Fletcher ) ; AGS cells transfected with oriP//EBNA1 vectors; and mouse tumor-derived explant lines . The GaCa cell lines and GaCa tumor-derived explant lines were cultured in Ham’s F-12 medium containing 10% fetal bovine serum ( FBS ) , 2 mM sodium pyruvate , 2 mM glutamine , and 100 IU of penicillin-streptomycin . All other cells were maintained in RPMI 1640 medium with the same supplements . All adherent lines were passaged after trypsinization . BL36 , C666–1 , AGS and AGS-BX1 cells were infected with the pGreenFire1-CMV , TR011VA-1 lentivirus ( System Biosciences , Mountain View , CA ) , which expresses green fluorescent protein ( GFP ) . GFP positive cells were sorted and collected by Fluorescence Activated Cell Sorting ( FACS ) , and were subsequently cultured with appropriate medium . 2×106 luciferase expressing AGS cells were transfected with 5 ug of plasmid DNA from either the oriP/EBNA1-EMPTY ( p220 ) or oriP/EBNA1-BART vector ( p3829 ) ( gift of Dr . Bill Sugden ) using an Amaxa nucleofector ( Lonza ) , the V kit solution , and Program B-023 . After 24 hr , transfected AGS ( AGS-EBNA1-EMPTY and AGS-EBNA1-BART ) were selected in F12 medium supplemented with 150ug/ml of hygromycin . Female NOD . Cg-Prkdcscid Il2rgtm1Wjl/SzJ ( NSG ) mice ( the Jackson Lab , ME , USA ) ages 6–8 weeks were housed and maintained under sterile conditions with free access to food and water . For carcinoma models , 2 . 5–7 . 5×105 cells ( C666–1 , AGS , AGS-BX1 , AGS-EBNA1-EMPTY and AGS-EBNA1-BART ) resuspended in 50 ul of phosphate-buffered saline ( PBS ) , were injected with a 25-gauge needle into the nasopharyngeal compartment of NSG mice under anesthesia [30] . For the lymphoma model , 2 . 5xl05 BL36 cells in 100 ul PBS were injected i . v . via the tail vein into NSG mice . Disease progression was monitored based on overall health and bioluminescent imaging . Mice were intraperitoneally injected with luciferin followed by anesthesia with 3% isoflurane and subsequent measurement of bioluminescence using an IVIS 200 imaging system ( Xenogen ) . Tumor burden ( or volume ) is presented as the radiance ( photons per second per centimeter squared per steradian or p/s/cm2/sr ) for each tumor by determining the photon emission/second of a given tumor within a radius encompassing 5% or greater of maximal signal intensity . The Kaplan-Meier survival curve analysis was conducted with the Prism program . For histology analysis , tissue samples were fixed in 10% formalin buffer and stored in 75% ethanol prior to paraffin wax embedding , sectioning , hematoxylin and eosin ( H&E ) staining by the Animal Histology Core at Tufts Medical Center . A small proportion of tumor tissue was excised at surgical operation and rinsed with DPBS . The tissue was then transferred to a 100mm Petri dish with appropriate medium . For the carcinomas tissue was minced with sterile scalpels into smaller fragment ( <2 mm ) and pressed under a 3 . 0 µm PET membrane cell culture insert ( BD Biosciences ) followed by incubation at 37°C in a 5% CO2-humidified incubator . Carcinoma cells started to grow as a monolayer and attach to the dish in a few days . Occasionally , the dish also contained a few clumps of floating epithelial cells . For lymphoma explant cultures , the tissue was cut and minced into very small pieces in medium . The finely minced tissue in suspension was filtered with a 70 μm cell strainer to remove debris . The cells were pelleted by centrifugation at 1 , 500 rpm for 5 min and then resuspended in RPMI 1640 medium and grown as usual for suspension cell cultures . Tumor tissue was ground to a fine powder in a liquid nitrogen cooled mortar and pestle . For RNA , tissue powder was then extracted using Trizol ( Invitrogen ) according to the manufacturer’s instructions . For genomic DNA ( gDNA ) , tissue powder was extracted with DNAzol ( Invitrogen ) . Briefly , 1ml of DNAzol was added to the powder and the cell lysates were gently passed through pipettes several times for complete extraction . Insoluble material was removed by centrifugation at 10 000 g for 10 min at 25°C . The viscous supernatant was transferred into another tube and DNA was precipitated using 0 . 5 ml of 100% ethanol followed by inversion of the tubes several times . The tube was left at room temperature for 20 min and the DNA was pelleted by centrifugation at 14 , 000 g for 15 min at 4°C . The DNA was washed twice with 1ml 85% ethanol and then air dried . The genomic DNA was finally dissolved in 200 ul 8 mM NaOH and the pH adjusted to 7 by adding 20 uL 1 M HEPES buffer . RNA and DNA were directly purified from cell lines without using a mortar and pestle . EBV miRNA expression was measured using real-time multiplex reverse transcript ( RT ) -PCR [24] and is expressed relative to the ubiquitous small cellular RNA U1 . Briefly , a panel of stem-loop RT primers specific for the 3′ end of each mature miRNA was pooled and used for RT of purified RNA with the TaqMan MicroRNA RT kit ( Applied Biosystems ) , followed by TaqMan PCR using specific primers and probes to each miRNA . Calibration curves for estimating the EBV miRNA copy number were generated using serial dilutions of synthetic oligonucleotides for each miRNA . This assay has been used previously for profiling BART miRNA expression in infected human tissue , tumor biopsies and cell lines [22 , 24] . Details of the assay including the lower limit of detection ( ≤10 copies of each miRNA ) and linearity of the miRNA qPCR assay ( up to ≥108 copies ) is described in [24] . For the human miRNAs mir-9 , 26a and 34a , the same RT kit and human specific TaqMan MicroRNA assays were used ( Life Technologies ) . The expression of all miRNAs in all experiments was normalized to the ubiquitous , small cell RNA U1 . mRNAs were measured by real time RT-PCR . A list of the primers used is provided in the supplemental material . RNA was reverse transcribed with an iScript cDNA synthesis kit ( Biorad ) , followed by real-time PCR using IQ SYBR Green supermix ( BioRad ) and specific primers . EBER1 expression was assessed by TaqMan PCR as previously described . [51] The housekeeping genes GAPDH , actin and tubulin were used as internal control for normalization . Values relative to GAPDH are shown . The relative expression is expressed as 2ΔCt , where ΔCt = mean value Ct ( gene of interest ) − mean value Ct ( GAPDH ) . For gDNA real time PCR , DNA was diluted 100-fold prior to PCR amplification with the IQ SYBR Green supermix and specific primers . All SYBR Green real time PCRs were performed on a Bio-Rad iCycler . The protocol was as follows: step 1 , one cycle of 5 min at 95°C; step 2 , 40 cycles of 15 s at 95°C and 1 min at 60°C; step 3 , one cycle of 1 min at 95°C . Fluorescence was monitored at the end of each extension phase . After amplification , melting curves were generated to verify the specificity of amplification . Primers for amplification are listed in S3 Table . To measure the oriP/EBNA1 plasmid copy number per cell in each AGS-EBNA1-EMPTY and AGS-EBNA1-BART tumor sample we divided the plasmid copy number by the number of cells in each sample . To determine how many cells there were , real time qRT-PCR was performed on each sample for the GAPDH mRNA . The Ct values were then converted to cell number from calibration curves ( cell number versus RT-PCR signal ( Ct ) ) of serial dilutions of the corresponding cell lines ( AGS-EBNA1-EMPTY and AGS-EBNA1-BART ) . The determine the total plasmid copy number in the tumor samples we performed gDNA real time PCR for the EBNA1 gene and the Ct values were read from standard curves ( Ct versus copy number ) generated by using serial 10-fold dilutions of a standard EBV ( B95–8 ) containing suspension with 1 . 7 × 104 DNA copies/μl ( Advanced Biotechnologies ) . For the growth curves , three separate cultures of 3×104 cells were seeded in100 mm Petri dishes . The cell number was then counted in duplicate for each replicate using the Scepter 2 . 0 handheld automated cell counter ( Millipore ) at indicated time points . For the soft agar colony formation assays , 1×104 cells were suspended in medium containing 0 . 4% agarose and overlaid onto a solidified layer of medium-containing 0 . 8% agarose in 6-well plates . Three separate cultures were set up for each cell line tested . After two-three weeks , colonies were counted and photographed in five fields for each replicate . For each assay 3 independently established BART positive cell lines and 2 independently established BART- lines were tested . The results were expressed as the means ± SEM of the combined data for measurements on all replicates of all the cell lines in a given group . For apoptosis sensitivity 4×105 cells were plated onto 6 well plates . Three independent cultures were set up for each cell line . After overnight incubation , cells were treated with 80 uM of etoposide for 72 hr prior to harvest . Harvested cells were washed with Annexin V binding buffer and subsequently stained with Annexin V-APC ( BD Bioscience ) for 15 min , followed by fixation in 4% buffered paraformaldehyde at room temperature for 5 min . Apoptotic cells were analyzed in duplicate by flow cytometry and defined as positive for Annexin V-APC staining . The percentage of cells that had undergone apoptosis in response to etoposide was assessed by subtracting that of apoptotic cells in the untreated from the treated population ( %Delta Annexin-V+ ) . In vitro invasion assays were carried out as described previously [52] using complete Matrigel ( BD Biosciences ) . A total of 18 μg Matrigel at a concentration of 0 . 3 μg/μl was coated onto a FluoroBlok insert ( BD Biosciences ) with an 8 μm pore size membrane . The transwell insert was allowed to dry overnight at room temperature and rehydrated with 60 ul of serum-free Ham’s F-12 medium for 2 hr . Cells were seeded in triplicate at 5×104 per transwell insert . 500 μl of media containing 5% FBS were added to the lower well of the assay chamber to act as a chemoattractant . After 24 hr , the transwell inserts were placed onto wells containing 4 μg/ml calcein AM in Hanks’ balanced salt solution and incubated for 30 min at 37°C in 5% CO2 . Cells that had passed through the pores and reached the trans-side of the membrane were counted with the imaging software MetaXpress . The number of invading cells was averaged over triplicate wells and presented as the means ± SD . All animal experiments described in this study were performed according to the Guidelines of the Tufts University Division of Laboratory Animal Medicine and were in accordance with the National Institutes of Health guide for the Care and Use of Laboratory Animals . Animal protocols were approved by the Tufts University Institutional Animal Care and Use Committee ( protocol B2011–108 ) . | Epstein-Barr virus is a herpes virus that persistently infects essentially every human being for life . It also has the ability to latently infect B lymphocytes and cause them to proliferate indefinitely in culture , and is associated with several forms of carcinoma and lymphoma . The virus contains genes for ~30 miRNAs in its BART region . The functions of these miRNAs are mostly unknown , but it is clear that they are not required to drive the growth of infected cells in vitro . We have shown previously , however , that these miRNAs are all highly expressed in the EBV associated cancers and that their expression is deregulated suggesting they may play a role in vivo . Until now , the significance of BART miRNAs to tumor development in vivo was unknown . Here we have used a mouse xenograft model to show that these miRNAs , while having little or no discernible effect on the growth of infected cells in vitro , potentiate the seeding and growth of EBV associated tumors in vivo . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| []
| 2015 | The Epstein-Barr Virus Encoded BART miRNAs Potentiate Tumor Growth In Vivo |
The role of mechanical force in cellular processes is increasingly revealed by single molecule experiments and simulations of force-induced transitions in proteins . How the applied force propagates within proteins determines their mechanical behavior yet remains largely unknown . We present a new method based on molecular dynamics simulations to disclose the distribution of strain in protein structures , here for the newly determined high-resolution crystal structure of I27 , a titin immunoglobulin ( IG ) domain . We obtain a sparse , spatially connected , and highly anisotropic mechanical network . This allows us to detect load-bearing motifs composed of interstrand hydrogen bonds and hydrophobic core interactions , including parts distal to the site to which force was applied . The role of the force distribution pattern for mechanical stability is tested by in silico unfolding of I27 mutants . We then compare the observed force pattern to the sparse network of coevolved residues found in this family . We find a remarkable overlap , suggesting the force distribution to reflect constraints for the evolutionary design of mechanical resistance in the IG family . The force distribution analysis provides a molecular interpretation of coevolution and opens the road to the study of the mechanism of signal propagation in proteins in general .
Cellular functions such as growth , motility , and signaling are tightly coupled with mechanical forces [1]–[3] . Proteins play a pivotal role in such mechanically guided processes , as robust elements bearing cellular stress , and as mechanosensors transducing the mechanical signal into a biochemical response [4] . A fundamental question is how a protein of mechanical function has been evolutionarily designed to withstand and transmit high levels of stress . Analysis and design of macroscopic structures such as cells , organs , or implants is routinely guided by the calculation of force propagation to predict and improve mechanical response [5] , [6] . However , how force propagates through the microscopic structure of proteins upon external stress is currently unknown . Mechanical response of proteins can be directly revealed by measuring forces for protein unfolding [7]–[9] , activation [2] , [10] , and enzymatic action [1] , [11] . For different titin immunoglobulin ( IG ) domains and engineered variants thereof , unfolding forces have been measured and rupture mechanisms inferred [12]–[16] . These data provide important insight into the load-bearing structural motifs of IG . A more fundamental question is how mechanical load distributes through a protein . It is an obvious assumption that force propagates through the structure to parts which , being distant from the application site of the perturbation , cannot be straightforwardly inferred from unfolding forces . Currently , there is no direct way at hand to explain the frequent experimental observation of site specific changes in dynamics ( see , e . g . , [17] ) . In this study we present a new method to detect the mechanical network sustaining load within a protein from molecular dynamics simulations . We apply the force distribution analysis to I27 , an IG domain of the muscle protein titin and one of the most robust protein domains known . The atomic resolution force distribution analysis relies on an accurate three dimensional picture of atomic interactions . We have determined the first high-resolution crystal structure of I27 , a cornerstone for the interpretation of force spectroscopy experiments and crucial for our analysis . The sparse mechanical network spanning the I27 structure is reminiscent of the molecular networks revealed by statistical coupling analysis on multiple sequence alignments [18] , [19] . Such networks of evolutionarily coupled residues have been proposed to reflect dynamic or energetic couplings along signaling pathways in proteins [20]–[22] . A direct relation of coevolution with molecular behavior , however , remains to be found . We compare the obtained pattern of force propagation with a network of coevolved residues found in the IG domain family , and find strongly coevolved residues to play a dominant role in force distribution . We therefore suggest internal strain propagation to present a first microscopic interpretation of a coevolved network in a protein with mechanical function .
The crystal structure of wild-type human cardiac muscle I27 ( residues 5253–5341 , renumbered to 1–89 for simplicity ) has been solved to 1 . 8Å by molecular replacement , Figure 1A . There are six I27 molecules in the asymmetric unit arranged around the local non-crystallographic 6-fold axis , with the N-termini of I27 closest to the local axis and the C-termini pointing outward . The entire amino acid sequence from residue 1–89 is visible in electron density ( except L89 in chain E ) , as well as an additional 3–4 residues left from the TEV cleavage site after the N-terminal H6-tag ( -AMA- in chain A; none in chain B; and -GAMA- in chains C–F , see Methods ) . The crystal structure reveals the well-known IG-like fold of the titin I-band IG domains and closely resembles the structure of the average NMR ensemble published earlier ( PDB code 1TIT [23] ) ; with 1 . 23Å root mean square distance ( RMSD ) between Cα carbons from all superposed residues ( Figure S1 ) . The largest distance ( 3 . 66Å for G53 ) is explained by the presence of a drastically different turn conformation between strands D and E , removing this turn ( residues 52–55 ) reduces global RMSD to 1 . 05Å . As a comparison , the RMSD between non-crystallographic symmetry-related chains in the I27 crystal structure ranges from 0 . 16–0 . 35Å . Residues involved in the H-bond pairing across the A'G strand have RMSD between 0 . 6–1 . 4Å ( res . 11–16 ) and 0 . 5–1 . 5Å ( res . 78–87 ) . After globally fitting to the wild-type backbone , the RMSD at the position of the two mutations encountered in the NMR structure ( A42T , T78A ) is 0 . 4 and 0 . 6Å , respectively . Crystallographic refinement statistics are given in Tables S3 and S4 . The present crystal structure of I27 represents an improved model for further studies because it does not contain potential structural perturbations introduced by unwanted mutations , and due to the high quality diffraction data , with overall coordinate error of 0 . 15Å . Overall , there are more and shorter H-bonds in the crystal structure comparing to the previously described NMR structure [23] . An additional H-bond between E3 and S26 and a stronger bond between K6 and E24 are found between strands A and B . While the hydrogen bonding between the A' and G strand largely overlaps , the interaction between Y9-N83 is absent but E12-K87 are within interaction distance , the latter supposedly adding resistance . Of the two residues mutated in the NMR structure , T78 ( first residue of strand G ) makes a side chain H-bond interaction with the carbonyl oxygen of L2 ( preceding strand A ) which is absent in the NMR structure and which could potentially add further mechanical resistance to I27 . The implications of the differences observed between the interaction networks of both I27 structures for atomic force microscopy ( AFM ) and molecular dynamics studies are obvious , since site-directed mutagenesis of the residues involved in those interactions has been used as a tool to discern the mechanical properties of I27 . To elucidate distribution of mechanical stress in the titin I27 domain , we directly calculate forces between each pair of atoms i and j during MD simulations of the high-resolution crystal structure of I27 described above . Forces are calculated individually for bonded and non-bonded ( electrostatic and van der Waals ) interactions below the cutoff distance using the interaction potentials defined by the OPLS [24] force field . By considering pairwise instead of atomic forces , forces do not average to zero over time . The propagation of the externally applied mechanical perturbation is measured as the change in pairwise forces upon applying external stress , , defined as the difference between the average force in the strained state and the relaxed state as obtained from equilibrium and force clamp ( FCMD ) [25] molecular dynamics simulations , respectively . In the strained state , as in vivo , force is applied to the termini of I27 in opposing direction . It remains controversial if IG domains in muscle titin ever fully unfold under physiological forces [26]–[28] . We here are interested in the force propagation within the fully folded protein , the physiologically relevant force-bearing structure . To this end , we apply a constant force of 300 pN , low enough to keep the protein structure intact . Importantly , no break in the AB strand , which is known as the first rupture event during I27 unfolding [29] , was seen during 20 ns of simulation time . The heavy atom RMSD between average equilibrium and FCMD structures was ∼1 . 6Å for all simulations . For convergence , forces were averaged over ten equilibrium and eight FCMD trajectories , each 20 ns in length , respectively . To reduce noise further , mainly resulting from slow side chain fluctuations , data were normalized as described in Methods . Dimensionless normalized changes in force are denoted Δf . Regarding the previously established importance of of inter-sheet hydrogen bonding for mechanical robustness [29] , the highly resolved non-bonded interactions in the crystal structure are important for the atomic detail force propagation network we are analyzing here . Figure 1B shows the normalized force distribution along the protein sequence , , obtained from summing over all changes of pairwise forces of atom j upon stress application . The high signal-to-noise ratio indicates statistical significance of the data , with a remaining average error <35 of normalized force estimated from equilibrium data . Importantly , the force distribution pattern forms a spatially connected network of residues ( Figure 2A and Video S1 ) . The overwhelming majority of force signals are part of a network spanning the protein between the stretched termini , suggesting that the network indeed reflects propagation of the external stress into the structure . Remarkably , the mechanical network is sparse in the sense that large parts of the protein including strands C , D , E and F are not part of the network and thus apparently play no direct role in mechanical stability ( Figure 2B ) . The pattern is found to be highly anisotropic , with the terminal strands passing the tension along the strands adjacent to the force application sites and partially into the protein core through a connected network of mainly hydrophobic interactions spanning between I2 , V4 , N77 , I23 , L25 , V30 , W34 , H56 , F73 and T78 from the N-terminal side and V11 , A19 , F21 , L60 , M67 , and L84 from the C-terminal side . We observe very different mechanisms of force distribution for the A'G and AB strands ( Figure 2C ) . The A'G strand , known to be crucial for mechanical resistance [29] , forms a mechanical clamp . Under load it shows a strong increase in interstrand H-bond ( blue ) and side chain forces ( red ) . This is accompanied by a stiffening of the strand and neighboring residues as reflected by decreased root mean square fluctuations ( Figure S2 ) . In contrast , the AB strand , even though it has been shown to be the first point of rupture of the IG fold , does not show any major contribution to mechanical stability ( Figure 2C ) , in agreement with experimental data [14] . Instead , force applied at the N-terminus is directly deflected into the protein core via mainly hydrophobic side chain interactions ( red edges ) between strands A and G , bypassing the AB interstrand hydrogen bonds , what again involves stiffening ( Figure S2 ) . This illustrates that rupture points are not necessarily involved in taking up large conformational load . Interestingly , signal propagation via side chain interactions involving stiffening of connected residues was previously observed experimentally in another globular protein [17] . A number of structural features , such as the A'G interstrand hydrogen bonds or the interactions involving T78 , which was absent in the NMR structure due to the T78A mutation , are part of the force-bearing network . The determination of a refined crystal structure thus served as an important basis for our force distribution analysis . Sparse networks which span protein molecular structures in a spatially connected and anisotropic way have been previously observed for evolutionary couplings [18] , [19] . An obvious assumption is that coevolved and therefore presumedly functionally important residues are involved in distributing and sustaining mechanical stress in IG domains . We tested this by comparing the force distribution pattern with evolutionary data from the IG domain family . Statistical coupling analysis ( SCA ) [18] , [19] was performed to identify coevolved residues . As a measure of coevolution , the impact of a perturbation in the amino acid frequency at one site in a multiple sequence alignment ( MSA ) on the frequency at another site is used , here termed statistical coupling energy ΔΔE . We constructed a MSA denoted ( Dataset S1 ) containing sequences from the 152 IG domains found in human muscle titin . We thereby restricted the sequences to those evolved for the specific function of bearing mechanical load . Mapping of the highly coupled residues on the I27 structure shows many of the couplings to span a spatially connected network between physically close residues residing in the protein core . In contrast , a subset of coevolved residues was found to couple distantly . It apparently belongs to a conserved IG-IG interaction interface , which becomes obvious when mapping coevolved residues onto two adjacent IG domains ( PDB-entry 2RIK [30] , Figure 3B ) . Indeed principal component analysis on the perturbation matrix ( see Methods ) clearly separates a subset from the bulk that coincides with the interaction interface , namely residues G16 , E27 , P28 , M67 , G69 , N77 , and S80 ( Figure 3A ) . For direct comparison with the force distribution analysis , which was restricted to interactions within one domain , we exclude these interface residues from further analysis . We compared the evolutionary network with the force distribution pattern in I27; hereto we restricted our analysis to inter side chain forces , ( see Methods ) , as evolution mainly optimizes side chains . A distinct group of residues with highest evolutionary couplings , residues I23 , V4 , F73 , I2 , V30 , and L84 , were found to mainly couple among each other and to clearly separate from the bulk , as indicated by hierarchical clustering analysis on the coevolution data ( Figure 4A ) Remarkably , these evolutionarily strongly connected residues show a very strong response to the applied mechanical perturbation , being among those showing highest changes in force distribution values ( Figure 4B ) . This suggests that evolutionary and force distribution analysis show an overlapping set of residues which are crucial for mechanical robustness ( Figure 4C ) . The overall comparison of evolutionary couplings with inter-residue forces of all IG residues indicates a connection of the evolutionary with the mechanical network as well . ΔΔE and show a significant correlation , shown in Figure S3A , with a correlation coefficient of R = 0 . 52 ( t = 5 . 56 and for 86 data points as calculated using student's t-test ) . This correlation is remarkable , regarding that the two data sets , from molecular simulations and sequence analysis , are completely independent . Furthermore , constraints acting on the evolution of proteins can be expected to be of manifold nature and thus to blur the correlation . One of these additional constraints is the optimization of the IG-IG interaction interface . Indeed , excluding the interfacial residues ( see above ) increases the correlation coefficient to R = 0 . 60 ( with t = 6 . 62 and for 79 data points ) . For the same reason , the correlation is expected to weaken when including sequences into the alignment that are not necessarily designed to bear mechanical load . To test this , we constructed a second more diverse alignment denoted ( Dataset S2 ) , containing 282 sequences with high similarity to the I27 structure . Results from coupling analysis for overlap ( Figure S3B ) , suggesting that the observed conservation pattern is robust with regard to the MSA . We observe a lower correlation for than for ( 0 . 37 vs . 0 . 52 ) corroborating our conclusion that the overlap found between evolutionary couplings and force distribution reflects an optimization for mechanical robustness of IG domains . Similarly , overlap of with overall residue conservation is low ( R = 0 . 18 ) , suggesting that it is the network rather than individual residues that are important for mechanical stability . An alternative explanation for the observed correlation could be packing interactions of core residues that give rise to both , high evolutionary dependencies and high inter-residue forces . In particular , coevolution has been suggested to primarily reflect packing interactions or constraints from structural integrity or thermodynamic stability . We however find the correlation between ΔΔE and packing density , measured as the number of close atomic contacts in I27 ( within a 6Å cutoff ) , to be low , with R = 0 . 23 . The correlation of ΔΔE with of R = 0 . 52 is significantly higher ( p<0 . 05 ) , and thus can barely , if at all , be explained by packing density . Similarly , changes in thermodynamic fold stability of I27 upon point mutation , as measured previously for 29 residues [31] , do not correlate with ΔΔE ( R = 0 . 19 ) . Instead , we find the fold stability to correlate significantly with the number of close contacts ( R = 0 . 71 ) . Consequently , while thermodynamic stability can be largely explained by core packing constraints , the evolutionary couplings can not be considered to reflect a simple spatial relationship or thermodynamic constraints . While mechanical and evolutionary couplings are mainly found between residues vicinal in the structure , vicinity in turn is not an indicator for strong coupling . By force distribution and coevolution analysis , we have identified the force-bearing scaffold of IG domains . The analysis predicts these residues to be crucial for mechanical function . We therefore expected a loss in mechanical stability upon their mutation , and directly tested this by forced unfolding of I27 in silico mutants in force-probe MD ( FPMD ) simulations [25] . We here considered the force distribution of the native state as the physiologically most relevant state of IG . We thus monitored force changes for the rupture of the AB strand , the primary unfolding event [29] , and of the A'G strand , whose rupture is visible as the highest force peak in AFM experiments , upon in silico point mutations . We selected the nine residues with highest values , which include the cluster of highly coevolved residues , and performed at least 10 independent FPMD simulations for each candidate after mutation to alanine . For most of the mutants we observed statistically significantly lower forces ( p<0 . 01 in student's t-test , Table S1 ) required to unfold into the intermediate state , Figure 5A and 5B . Interestingly , three of the residues with high instead show little change in mechanical resistance upon mutation . In addition , none of the mutations gave rupture forces smaller than 500 pN . This may reflect the robustness of the mechanical network to local perturbations , allowing it to re-balance load . We performed six control simulations of residues that featured low values despite their location in the protein core . As expected , none or only minor decreases in rupture forces are found , with an overall destabilization significantly lower than for the nine selected high candidates ( p<0 . 03 , Figure 5B ) . Several residues located in the A'G strand show a high degree of force distribution into the protein core , including F21 and L84 , as indicated by their high signal , in contrast to the common notion of the primary importance of interstrand hydrogen bonding . Indeed , removing the hydrophobic contacts of F21 and L84 results in significantly decreased rupture force of the A'G strand ( Figure 6A ) . This renders F21 and L84 interesting targets for further experimental studies and stresses our observation that specific core interactions are important force-bearing interactions . Force spectroscopy experiments mainly probe the later step of unfolding of the intermediate via the C-terminal A'G strand rupture , since it involves the highest rupture forces [29] . For validating our FPMD simulations with experimental data , we monitored rupture forces of this unfolding step . We included additional in silico mutations for comprehensive comparison with a series of AFM experiments done by Clarke and co-workers [14] . Our results are in good agreement with their data , with a significantly decreased rupture force only found for the V13A and V86A mutants ( Figure 6B ) , but not for the others .
We present here the first force distribution network in a protein at atomic detail , thereby revealing the force-bearing scaffold rendering I27 mechanically robust . The network is in good agreement with previous experimental results [12] , [14] and provides a basis for the choice of mutants to rationally alter the mechanical response , even at sites distant from force application . Our force distribution analysis can be directly tested by force probe experiments of such mutants , and can in future allow the design of proteins having specific mechanical stability and response . It can also help in explaining and engineering stability in biological as well as synthetic materials , such as silk or polymers . Previous attempts to detect force distribution in proteins have been restricted to elastic network models [32]–[34] , which assume linear elasticity of the interactions . Importantly , we find the non-linear nature of non-bonded interactions to give rise to stiffening of structural motifs . Examining the extent of non-linearity in the mechanical response of IG and other proteins by determining the force pattern as a function of external stress will be the subject of future studies , and will allow the development of coarse-grain models without the need for assuming harmonic interaction potentials . Furthermore , while in elastic network models sequence specificity is either neglected [32] or taken into account implicitly by modifying inter-residue force constants [33] , [34] , our pairwise atomic analysis is directly sensitive towards residue types . An interesting question is how long an externally applied perturbation needs to propagate through a protein , which is not directly accessible by force spectroscopy . We find a sub-nanosecond time scale to be sufficient for convergence of forces in I27 ( Figure S4 ) , a rigid protein not undergoing major conformational changes upon stretching . Remarkably , this time scale is comparable to that measured experimentally for heat transport in helices [35] and to previous theoretical studies [36] . However , a more detailed force distribution analysis that is both time and spatially resolved is required to determine the timescale for force propagation and mechanical equilibration , which is the aim of future studies . The mechanical response of I27 appears to be remarkably robust with regard to point mutations within the force distribution network , which lower the rupture force never by more than ∼30% . In particular , a small subset of highly force bearing residues ( V30 , F73 , L78 ) do not cause any loss of overall stability upon mutation . This suggests a certain redundancy or plasticity of the mechanical network . Biological networks ranging from interaction networks to gene regulatory networks are increasedly well characterized and understood [37] , [38] . The force network described in this study represents a new type of biological network , which asks for graph theoretical analysis to further clarify its function , including splits , redundancy , hubs , and other properties . We hypothesize that strain propagation as revealed here acts in an allosteric protein as a mechanism to transduce an external signal through the protein core to distant sites . The force distribution analysis was here applied to the propagation of mechanical force as a perturbation acting on a protein , but can easily be extended to other types of perturbation , in particular to allosteric signals . Since forces can monitor allosteric signal propagation more sensitively than coordinates , our method is particularly suited for allosteric proteins not undergoing an obvious conformational change , i . e . , rigid proteins [39] , and dynamic allostery [40] for which changes in fluctuations in pairwise forces can be expected . Networks of coevolved residues and their relevance for protein stability and function have been exhaustively analyzed for many proteins [18] , [19] , [21] , [41]–[43] . It has been suggested that the molecular mechanism by which coevolved residues couple is of dynamic or energetic nature [20] . Couplings in binding free energies [44] have not yet been unambiguously correlated with evolutionary couplings . Recently , attempts have been made to compare couplings in dynamic fluctuations with evolutionary couplings [45] . In contrast to the notion of functional protein fluctuations propagating force , we here propose the stiffness , i . e . , the static nature , of the force-bearing scaffold to be functionally crucial ( Figure S2 ) . The relevance of this first molecular interpretation of evolutionary design for other proteins , with mechanical , allosteric , or other functions , remains to be investigated .
We use a modified version of Gromacs 3 . 3 . 1 to write out forces between each atom pair i and j . Forces include contributions of individual bonded ( bond , angle , dihedral ) and non-bonded ( electrostatic and van der Waals ) terms below the cutoff distance , which are stored separately for further analysis . The force between each atom pair is represented as the norm of the force vector and thus is a scalar . Attractive and repulsive forces are assigned opposite signs , forces are averaged over simulation time and converge to an equilibrium value . As we consider the direct force between each atom pair , this equilibrium value can be different from zero , even for the theoretical case of a system without any motion . We hereby obtain the advantage to be able to observe signal propagation even through stiff materials , where forces propagate without causing major atomic displacement . Atomic forces , i . e . , the sum over the force vectors acting on an atom , instead average out to zero over time and are not of interest here . A real world example for such force propagation is Newton's cradle . Due to the nature of the non-bonded potentials pairwise forces are most significant for atom pairs in relative close proximity , resulting in a force-propagation network comprised of a series of short to medium ranged connections . An approximation is used to represent angle and dihedral forces ( Figure S5 ) . Multi body forces such as hydrophobic effects and PME forces are not included and thus cannot be accounted for in the analysis . Average forces were written out every 10 ps . To obtain converged averages , forces for each atom pair were averaged for each trajectory and afterwards over all pulling and equilibrium trajectories , respectively . The normalized change in force is defined as the difference between forces in the strained state , , and equilibrium forces , , for each atom pair i , j . Normalization by the standard error of the mean ε accounts for differences caused by insufficient sampling , i . e . , slow side chain or backbone fluctuations that cannot equilibrate in simulation time . ( 1 ) The overall distribution , however , remains largely unchanged upon normalization , Figure 1C . The mechanical coupling of an atom j with respect to all other atoms is then defined as the absolute sum . ( 2 ) In analogy , for all pairs of side chains u and v , we sum up forces , where atom i and atom j must not be part of the same residue . Forces are averaged for each trajectory and afterwards over equilibrium and pulling trajectories . The normalized change in inter side chain forces is defined as the difference between force clamp and equilibrium forces normalized by the standard errors observed between equilibrium and force probe trajectories , respectively . ( 3 ) We then define the mechanical coupling for residue v as the absolute sum . ( 4 ) were carried out using Gromacs 3 . 3 . 1 [46] . The OPLS all atom force field [24] for the protein and the TIP4P [47] water model were employed . The crystal structure of I27 ( PDB-entry 1WAA ) was used as starting structure for all simulations . Simulation times were 20 ns for equilibrium and FCMD simulations . A constant force of 300 pN was applied to both terminal residues in opposing direction . The applied force was low enough to keep the protein structure intact and no partial rupture events were seen during the simulation time . Simulations were run in the NpT ensemble , temperature was kept constant at 300 K by coupling to the Berendsen thermostat [48] . The pressure was kept constant at p = 1 bar using anisotropic coupling to a Berendsen barostat [48] with and a compressibility of 4 . 5 10−5 bar−1 in the x , y , and z directions . In FPMD simulations of the I27 mutants all bonds were constrained using the LINCS [49] algorithm; an integration timestep of 2fs was used . No constraints and an integration timestep of 1fs was used for equilibrium and FCMD simulations . Lennard-Jones interactions were calculated using a cutoff of 10Å . At a distance smaller than 10Å , electrostatic interactions were calculated explicitly , whereas long-range electrostatic interactions were calculated by Particle-Mesh Ewald summation [50] . System coordinates were saved every 10 ps . The X-ray structure of I27 ( PDB-entry 1WAA ) was used as starting structure for all simulations . Protonation states of histidines were determined by optimizing the hydrogen bond network using Whatif [51] . All mutations were done using Whatif , starting from the equilibrated structure . Structures were solvated in a triclinic box with dimensions 55×55×100Å , containing ∼40 , 000 atoms . Sodium and chloride ions corresponding to a physiological ion strength of 100 mM were added . An energy minimization of 1000 steps using the steepest descent algorithm was followed by a 400 ps MD simulation with harmonical restraints on the protein heavy atoms with a force constant of k = 1000 kJ mol−1 nm2 to equilibrate water and ions . For mutants , this simulation was followed by a 1 ns MD simulation with the same harmonical constraints on backbone atoms only . A subsequent free MD simulation of 5 ns length was performed to equilibrate the whole system , during which the protein backbone root mean-square deviation ( RMSD ) was monitored . All mutants remained stable during free MD , with a heavy atom RMSD to the starting structure <2Å for all structures ( Table S2 ) . For each run , new velocities were chosen form a random distribution using Gromacs , followed by a 400 ps MD simulation with restraints on protein heavy atoms as described above . During unfolding simulations partial rupture at the N-terminus that leads to the intermediate state was measured by means of the distance between the of S26 and the backbone N of E3 . Rupture of the A'G strand was measured by monitoring the length of the V13-K85 hydrogen bond . Mutants were partially unfolded during a 12 ns FPMD simulation with a harmonic spring potential applied on both terminal residues , with a spring constant of k = 500 kJ mol−1 nm−1 . The springs were moved with a constant velocity of 0 . 02Å ps−1 each . Coupling analysis was carried out as described by Ranganathan and co-workers [18] , [19] . SCA assigns a “statistical coupling energy” ΔΔE to each position in a multiple sequence alignment . ΔΔE is determined upon perturbation by removal of sequences of the alignment that changes the amino acid distribution at a specific position . If the observed amino acid distribution at position b changes with respect to perturbation at position a , these positions are “coupled” . Intuitively , if evolution changed residue a it is also likely to change residue b to maintain the protein's functionality , i . e . , a and b are statistically dependent throughout evolution . We perform a set of small perturbations to the MSA by removing one row at a time . Each perturbation results in a small fluctuation of ΔΔE values for each position . The matrix containing perturbations versus ΔΔE fluctuations is referred to as perturbation matrix , and PCA was done on this matrix . We then define the coupling score between two positions as the cross product of the two ΔΔE trajectories , which is the norm of the two position vectors in perturbation space . We perform SCA on two sets of sequences . The first , , contains the sequences of all 152 IG domains in human muscle titin ( UniProtKB/Swiss-Prot Q8WZ42 ) . The second more general alignment was chosen from the IG I-set ( Pfam id: PF07679 ) family by similarity to I27 and contains 282 sequences . Hereto a Blast [52] search against all sequences from the I-set family was done and sequences with e-score <1 . 00 were selected . All sequences with similarity >90% to any other sequence were removed . SCA requires the set of sequences to be sufficiently diverse . Further the set of sequences should be well balanced , in the sense that the average similarity to all other sequences should be roughly the same for each sequence . We checked the diversity of sequences within the alignments and found it to be comparable to results published earlier . I . e . , the average sequence similarity within IG domains of human titin of 0 . 46 compares well with the average similarity of 0 . 48 for the alignment of PDZ-domain sequences used by Ranganathan and co-workers [18] . Figure S3C shows that the variance in average sequence similarity is low , indicating that both alignments are well balanced . To assess the alignment quality , we calculated the sequence entropy [53] , a measure for conservation at each position , and again find it to be very comparable with data published earlier ( PDZ = 1 . 95 , ) . Sequences were aligned using Dialign [54] that is reported to perform well on local sequence alignments [55] . In the final alignments , no position aligned to the I27 sequence contained more than 50% gaps . For the calculation of correlations with inter-side chain forces the terminal residues which were directly subjected to force and the ultra-conserved W34 were excluded . Hierarchical cluster analysis was done using Ward's algorithm [56] as implemented in R ( R Development Core Team ) . At each step in the analysis , the method considers the union of every possible cluster pair , and the two clusters with minimal square sum of error are combined . The city-block metric was used as distance measure . For the comparison of with residue conservation , conservation was calculated using Shannon entropy . The I27 structure was determined by molecular replacement with AMORE [57] using as search model a representative structure from a previously determined NMR ensemble ( PDB-entry 1TIT [23] ) . Of the six copies of I27 , four copies were orientated and placed successively with AMORE . Correctness of this incomplete model was assessed by the increase in correlation coefficient ( from 24 . 8% to 40 . 4% ) and the concomitant decrease in R-factor ( from 50 . 2% to 45 . 2% ) attained between the first and the fourth I27 molecule . The last two copies of I27 could not be automatically located with AMORE using the same protocol . Instead , this first model containing four I27 molecules was refined ( rigid body ) and then the missing two copies of I27 were placed manually . Since the I27 NMR structure contained two mutations not present in the wild type titin sequence ( A42T , T78A ) , the sequence in our I27 structure was modified accordingly . Refinement and modeling was performed iteratively using REFMAC5 [58] and Turbo-Frodo ( http://www . afmb . univ-mrs . fr/-TURBO- ) . The model has been refined to a final R-factor of 0 . 211 and R-free of 0 . 268 ( Table S3 ) . Upon convergence , the maximum-likelihood coordinate error estimation is 0 . 15Å . Residues 5253–5341 of human cardiac titin I27 ( renumbered 1–89 for simplicity ) were amplified from a cDNA clone ( accession code X90568 of the EMBL data library ) by polymerase chain reaction ( PCR ) and subcloned into the pETM11 expression vector for expression of I27 in E . coli fused to a TEV ( tobacco edged virus protease ) cleavable N-terminal His6-tag . An overnight preculture of BL21 ( DE3 ) cells transformed with the I27 expression plasmid was used as inoculum to 3 liters of Luria-Bertani medium plus 50 µM kanamycin , and I27 expression was induced at OD600 of 0 . 6 by the addition of 1 mM IPTG . Cells were harvested 3 h post induction , lysed by sonication in 20 mM Tris-HCl , 300 mM NaCl , 20 mM imidazole , pH 8 . 0 , and clarified by centrifugation at 20 , 000×g and filtration through a 0 . 22 µm membrane . The supernatant , containing soluble protein , was poured on a Ni2+-NTA agarose column and I27 was eluted with a linear imidazole gradient . Elution fractions containing I27 were pooled together and incubated with TEV protease for 3 h to remove the affinity tag . The cleaved tag and TEV were removed by passing the digestion over a second Ni2+-NTA column ( Qiagen ) . The flow-through , containing cleaved I27 , was dialysed overnight against 20 mM Tris-HCl , 2 mM DTT , pH 8 . 0 , loaded onto a MonoQ ( GE Healthcare ) ionic exchange column and eluted with a 0–1 M NaCl linear gradient . Final polishing of I27 was brought about by gel filtration chromatography on a Supedex 75 ( GE Healthcare ) column . Purified I27 was concentrated to 10 mg/ml in 20 mM Tris-HCl , 50 mM NaCl , pH 7 . 5 . Crystals of I27 were grown by hanging drop vapor diffusion in 20% PEG MME 550 , 75 mM MES , 7 . 5 mM ZnSO4 at pH 6 . 5 . The crystals were mounted in Hampton Research nylon loops , and cryoprotected in a cryosolution made of the mother liquor and 5% ( v/v ) glycerol . Crystals belonged to P212121 space group and X-ray data up to a maximum resolution of 1 . 8Å were collected on beam-line BW7B ( EMBL , Hamburg , DESY ) at a wavelength of 0 . 841Å at 100 K . One segment of 90° was sliced in 0 . 5° rotation steps to give complete and redundant data . Diffraction data were processed with MOSFLM [59] and scaled in SCALA [60] ( Table S4 ) . Cell content analysis and self-rotation function calculations indicated that the asymmetric unit contained 6 copies of I27 . | Many biological processes such as cell proliferation and signaling are guided by mechanical stress . Proteins as the molecular machinery behind these processes are reacting to or withstanding mechanical forces in specific ways . How mechanical stress propagates through proteins to induce a certain mechanical response is currently unknown . We here present a new method that detects force distribution in proteins , reminiscent of computational approaches used to engineer macroscopic structures . The method is based on molecular dynamics simulations during which we calculate changes in interatomic forces , here caused by pulling the protein . We apply this method to the extremely robust immunoglobulin domain of the muscle protein titin and obtain the mechanical network of stress-bearing elements spanning the protein scaffold . Mutations in this region were shown to result in a significant loss of mechanical stability . We then ask how the remarkable mechanical stability has been designed during evolution . To this end , we compare the force distribution network with a network of coevolved residues found in the immunoglobulin family . Both networks show a remarkable overlap , thereby suggesting that the observed stress propagation pattern reflects constraints used during evolution to render this protein mechanically robust . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"computational",
"biology/molecular",
"dynamics",
"computational",
"biology/comparative",
"sequence",
"analysis",
"biophysics/protein",
"chemistry",
"and",
"proteomics"
]
| 2009 | Mechanical Network in Titin Immunoglobulin from Force Distribution Analysis |
Bartonella quintana , the etiologic agent of trench fever and other human diseases , is transmitted by the feces of body lice . Recently , this bacterium has been detected in other arthropod families such as bed bugs , which begs the question of their involvement in B . quintana transmission . Although several infectious pathogens have been reported and are suggested to be transmitted by bed bugs , the evidence regarding their competence as vectors is unclear . Bed bugs at the adult and instar developmental stages were fed three successive human blood meals inoculated with B . quintana bacterium from day one ( D1 ) to D5; subsequently they were fed with pathogen-free human blood until the end of the experiment . Bed bugs and feces were collected in time series , to evaluate their capacities to acquire , multiply and expel viable B . quintana using molecular biology , immunohistochemistry and cultures assays . B . quintana was detected molecularly in 100% of randomly selected experimentally infected bed bug specimens ( D3 ) . The monitoring of B . quintana in bed bug feces showed that the bacterium was detectable starting on the 3rd day post-infection ( pi ) and persisted until day 18±1 pi . Although immunohistochemistry assays localized the bacteria to the gastrointestinal bed bug gut , the detection of B . quintana in the first and second instar larva stages suggested a vertical non-transovarial transmission of the bacterium . The present work demonstrated for the first time that bed bugs can acquire , maintain for more than 2 weeks and release viable B . quintana organisms following a stercorarial shedding . We also observed the vertical transmission of the bacterium to their progeny . Although the biological role of bed bugs in the transmission of B . quintana under natural conditions has yet to be confirmed , the present work highlights the need to reconsider monitoring of these arthropods for the transmission of human pathogens .
Bartonella quintana is a fastidious gram-negative bacterium that is regarded as a re-emerging human pathogen [1] . B . quintana DNA has been detected in the dental pulp of a 4000-year-old man [2] and in lice found in a mass grave of Napoleon’s soldiers in Lithuania , which suggests that many of the soldiers were affected by trench fever [3] . Trench fever was the first described clinical manifestation of B . quintana infection , and it affected thousands of soldiers during World Wars I and II [3] . Subsequently , B . quintana has been identified as an agent of bacillary angiomatosis in AIDS patients [4] , endocarditis [5 , 6] , chronic bacteremia [7 , 8] , and chronic lymphadenopathy [9] . The severity of Bartonella infection correlates with the immune status of the patient; the clinical manifestations can range from benign and self-limited to severe and life-threatening disease [10] . Although body lice are considered as the main vector of B . quintana [11] , this bacterium has also been found in other arthropods such as head lice [12 , 13] , ticks [14] and mites [15] . Recently , after the detection of B . quintana DNA in fleas [16] , it was experimentally demonstrated that the cat flea , Ctenocephalides felis , could acquire and excrete viable B . quintana in their feces [17] . These results supported the likely vector role of fleas in trench fever or other clinical manifestations caused by B . quintana [17] . The recent detection of B . quintana DNA in Cimex hemipterus ( tropical bed bugs ) collected from two prisons in Rwanda indicated that bed bugs could be involved in the transmission of B . quintana [18] . This raises the question of whether C . lectularius ( common bed bug ) could acquire and excrete viable B . quintana and thus constitute a potential competent vector . For this purpose , we used an experimental model infection of C . lectularius bed bugs using three different approaches: qPCR , culture and immunohistochemistry .
B . quintana strain Oklahoma ( ATCC 49793 ) [17] was used to infect the blood used to feed the bed bugs . The use , culturing and all procedures involving experimental infections of B . quintana were conducted in a Biosafety Level 2 room . B . quintana strain was grown as described previously [19] on 5% Columbia sheep blood agar plates ( BioMerieux , Marcy l’Etoile , France ) in a humidified atmosphere at 37°C supplemented with 5% carbon dioxide ( CO2 ) using the pouch of atmosphere generation system CO2 Gen ( Oxoid Ltd by Mitsubishi Gas chemical Company Inc , Japan ) . After 8 to 10 days of culture , the bacteria were harvested by adding four-hundred μL of phosphate buffered saline ( PBS ) , pH 7 . 2 ( BioMerieux , Craponne , France ) . Two-hundred microliters of the pure bacterial suspension were mixed with 2 mL of whole blood , and this was used as the blood meal to infect the bed bugs . The remaining 200 μL of the bacterial suspension were diluted up to 10–10 and cultured to estimate the number of colony-forming units ( CFU ) per microliter . Since 2012 , bed bugs ( Cimex lectularius ) have been maintained in a laboratory insectarium by our team at the WHO collaborative center for rickettsioses and other arthropod borne bacterial diseases in Marseille , France . This colony originated from bed bugs collected at the adult and the five instar stages from an infested apartment ( Aix-en-Provence , France ) using a modified Dyson DC34 hand vacuum system . They were maintained in containers kept in incubator at 60% humidity and 22°C . The bed bugs were fed once a week using citrated human blood obtained from the French Blood Establishment . Ethical approval for the use of in vitro human blood was obtained from the laboratory research ethics board of Molecular Hematology , French Blood Establishment . Two mL of blood was placed in a Hemotek artificial feeder machine ( Hemotek 5W1; Discovery Workshops , Accrington , UK ) covered by an artificial membrane of Parafilm M ( Sigma-Aldrich , Saint-Louis , Missouri , USA ) that was stretched to the twice of its length and width [20] . To prevent contamination during the experimental infection model , the Hemotek feeder and the recipient’s containers of bed bugs were introduced in a clear acrylic box . Two separate trials were conducted using C . lectularius drawn from the same colony at the same age . Prior to initiation of the infection , the bed bugs and their feces were shown to be free from B . quintana using qPCR . We formed 4 groups for each trial including 2 infected ( 1 adults and 1 larva group ) and 2 control groups ( 1 adults and 1 larva group ) ; each group consisted of 30 bed bugs . In the adult vials we used 10 males and 20 females , and also larval group was composed of 30 Larva 1 ( L1 ) bed bugs . The concentration of B . quintana in the infected suspension composed by the bacterial suspension and the blood meal was 6 x 108 CFU/mL bacteria in trial 1 and 8 x 105 CFU/ mL in trial 2 . Each group of bed bugs was fed 3 times in 5 days ( every other day ) with 200 μL of the bacterial suspension mixed with 2 mL of blood meal . The control groups were fed with 2 mL of uninfected blood mixed with 200 μL of PBS . Subsequently , all bed bug groups were fed with uninfected blood every other day starting on the 3rd day post-infection ( dpi ) until the end of the experiment . We tested 200 μL of the infected inoculum ( the infected blood suspension that the bed bugs fed on ) to ensure the presence of B . quintana in the infected blood meal using qPCR . We cultured 150 μL of the inoculum and plated dilutions up to 10–10 to ensure the viability and to determine the concentration of B . quintana in the infected inoculum . At the 3rd dpi , five viable bed bugs and approximately 20 mg of feces from each group ( from B . quintana exposed group of adults and instars and also from the control groups ) were recovered for analysis by qPCR . Feces were collected from a sheet of paper placed on the bottom of the bed bugs containers . Culture analysis of feces and two bed bugs were also performed; both tests were used to determine the acquisition and viability of B . quintana in bed bugs and in their feces . Four adult C . lectularius from the B . quintana-exposed group were immunohistochemically analyzed to determine the bacterial localization . Four bed bugs from the control group were also analyzed and served as controls . Starting on the 5th dpi , we recovered two adults and feces every 48 h to monitor the excretion of B . quintana through the end of the experiment ( 21st dpi ) . We screened five eggs from the container housing the infected adults by qPCR at the 3rd dpi to determine if the eggs were infected . Simultaneously , we recovered ten eggs to be reared in separate vials to obtain L1 and L2 larvae . The larvae were analyzed by qPCR to determine if any B . quintana acquisition occurred . The DNA of individual bed bugs and their feces were extracted using an automatic EZ1 robot ( QIAGEN-BioRobot_ EZ1 , Tokyo , Japan ) according to the manufacturer’s instructions ( EZ1 DNA Tissue Kit , QIAGEN , Hilden , Germany ) . First , we decontaminated the surface of the bed bugs by 5 min immersion in ethanol alcohol ( COOPER , Paris , France ) , followed by three 5 minutes immersions in sterile PBS as described previously [21] . Each bed bug was incubated overnight at 56°C in 180 μL of buffer G2 and 20 μL of proteinase K for pre-lysis followed by extraction using EZ1 robot . For all samples , the final elution volume was 100 μL . Template DNA was used in the qPCR assays targeting two specific B . quintana genes that encoded 3oxoacyl-[acyl-carrier-protein] synthase ( fabF3 ) and a hypothetical intracellular effector ( yopP ) [13] , which are both B . quintana-specific genes . The CFX96 ( Bio-Rad , France ) was used to perform each real time PCR . The qPCR was considered positive when the cycle threshold ( Ct ) was lower than 36 [17] . The number of B . quintana in each sample was calculated based on the DNA copy numbers . A qPCR standard curve was obtained by analyzing the fabF3 and yopP systems in serial dilutions of B . quintana culture , and the standard value was determined for duplicate trials [17] . The B . quintana infection density was quantified as the ratio of the log of the transformed fabF3 and yopP copy numbers per individual bed bug , feces , and blood meal . The cycle thresholds ( Ct ) values of [12 . 9; 14 . 5; 17 . 8; 22 . 0; 25 . 7; 28 . 9; 30 . 9; 34 . 3 and 36 . 0] correspond , respectively , to [4 x 109; 4 x 108; 4 x 107 4 x 105 4 x 104; 4 x 102; 4 x 101 and 4] CFU/mL . Regressions formula was realized as following: Y = -0 . 377X + 14 . 236 ( R² = 0 . 996 ) for fabF3 gene and Y = -0 . 372X + 14 . 158 ( R² = 0 . 996 ) for yopP gene . Approximately 500 μL of homogenized feces ( 20 mg in 500 μL of PBS ) from groups of infected and uninfected bed bugs with 5% sheep’s blood were filtered using a 0 . 8 μm filter ( Millex Ø 33 mm , Dominique Dutscher ) and were cultured on agar plates [17] . The bodies of the bed bugs were also cultured using the same method described for the culturing of feces . Immunohistochemistry was performed on 3 μm-thick , paraffin-embedded sections of formalin-fixed bed bugs using the Ventana Benchmark autostainer ( Ventana Medical Systems , Inc . ) [17] . Four infected bed bugs ( 2 from each trial ) and 4 uninfected bed bugs were analyzed ( 2 from each trial ) . After deparaffinization , each tissue section was incubated with polyclonal rabbit anti-B . quintana antibody , which was diluted 1:5000 as previously described [22] .
In the two trials , adults and L1 bed bugs were exposed to B . quintana three times in 5 days using B . quintana-infected blood meal . On the 3rd dpi , we individually analyzed five adults and five L1 C . lectularius by qPCR . The control groups ( fed on blood meal with 200 μL of PBS ) were negative by qPCR for the presence of the bacterium in both trials . In the B . quintana-exposed groups , we detected B . quintana in 100% ( 5/5 ) of the adult bed bugs and in 100% ( 5/5 ) of the L1 bed bugs in both trials . The quantities of B . quintana in each individual bed bug sample per trial as determined by qPCR of the fabF3 and yopP genes are given in Tables 1 and 2 . Bacterial quantities ranged between 5 . 8 x 107 CFU/ mL and 4 . 8 x 102 CFU/ mL in trial 1 and from 2 . 8 x 106 CFU/ mL to 6 x 10 CFU/ mL in trial 2 . Feces of adults and larva bed bugs were also tested by qPCR to evaluate the presence of B . quintana and to confirm the route of way of elimination . The results indicated the presence of the bacterium in the feces in both trials with 2 . 8 x 108 CFU/ mL in the adult feces and 5 . 5 x 107 CFU/ mL in the L1 feces in trial 1 and 9 . 1 x 103 CFU/ mL in the adult feces and 2 . 8 x 106 CFU/ mL in the L1 feces in trial 2 ( Table 3 ) . Immunohistochemical analysis of the 4 tested C . lectularius ( from the 3rd dpi ) from trial 1 and trial 2 demonstrated the presence of B . quintana as dense clusters of immunopositive microorganisms in the midgut and hindgut of the gut tract ( Fig 1 , Table 3 ) .
Here , we report two experimental trials to investigate potential acquisition and transmission of B . quintana ( the agent of trench fever and other diseases ) by bed bugs ( C . lectularius ) . The results show that bed bugs ( adults and larva ) exposed to B . quintana can acquire the bacterium and eliminate it in feces . The bed bugs maintain and shed stercorarially B . quintana for up to 17th or 19th dpi depending on the inoculum concentration . However , B . quintana was detected viable in feces and was shown to be alive inside the body of the bed bugs at the 3rd dpi . Using immunohistochemistry , the bacterium was localized in the midgut and hindgut of the bed bugs digestive tract . Surprisingly , B . quintana was detected in eggs , L1 and L2 larvae . In this study , we used three validated approaches . First , qPCRs was perfomed to study the acquisition and elimination of the bacterium by C . lectularius . This technique is reliable because we used a set of two qPCR systems targeting yopP and fabF3 , which are known to be specific for B . quintana DNA detection , and we used negative and positive controls . Second , we cultivated the bacteria from the samples to determine if the eliminated bacteria were viable . This approach was also a validated technique [17] containing a negative and positive control . The third method was immunohistochemistry , which was used to localize the bacterium inside the body of the bed bugs . The immunohistochemistry experiments were conducted in a blinded fashion by one of us ( HLi ) , and the results were concordant with the qPCRs results . Cimex lectularius and C . hemipterus ( Cimicidae: Hemipetra ) , commonly called bed bugs , continue to increase in scope [23 , 24] . In recent years , these hematophagous arthropods have undergone a major resurgence in frequency and in geographic distribution leading to clinical problems . An increasing number of infestations have been reported in Europe [25 , 26] [23 , 27] America [28] , Australia [23] , Asia [29 , 30] [31 , 32] and Africa [18 , 33] . A bite causing cutaneous lesions is the most common clinical consequence of bed bugs on public health . In addition , mental health can be affected by knowledge of a bed bug infestation in one's living environment [23] . Bed bugs are suspected of transmitting infectious agents , however there is little evidence that such transmission has ever occurred . More than 45 pathogens associated with human infection and disease have been suspected to be transmitted by bed bugs [34] . Older scientific literature cited by Goddard and de Shaso [35] suggested that bed bugs may be vectors of yellow fever , tuberculosis , relapsing fever , leprosy , filariasis [36] , kala azar ( leishmaniasis ) , smallpox and HIV [37 , 38] . Yersinia pestis has also noted to develop inside the body of bed bugs , C . lectularius [39 , 40] . Verjbitzki [40] found with animal model infection of bed bugs with high virulence strain of Y . pestis can induce death of the guinea-pigs . They found also that three bed bugs are able to convey infection [39 , 40] . Jordansky and Klodnitsky [41] found that the number of Y . pestis bacilli in the bed bug's stomach increased from the third to the sixth day after the infected meal [39 , 41] . Throughout these animal models , it may be appear that bed bugs can play an important role to convey infection of plague and perhaps other pathogens . Hepatitis B virus has also been postulated as likely candidate for possible transmission by bed bugs [42–45] . Blow et al . 2001 [45] , offered evidence for stercorarial transmission of Hepatitis B viral agents from bed bugs in a time series and with transtadial transmission . Recently Salazar et al [46] assessed the vector competence of C . lectularius against Trypanosoma cruzi and it has been confirmed that T . cruzi was viable in bed bug feces . Goddard et al [47] have experimentally infected bed bugs with Rickettsia parkeri and found using immunofluorescence that the bacterium was present in the salivary gland at 15 days post infection [47] . Moreover , our laboratory recently detected B . quintana DNA in C . hemipterus collected from two prisons in Rwanda [18] . The only confirmed and known vector of B . quintana is body lice ( spread through feces ) . However , several studies suggested that hematophagous arthropods , such as flies , lice , fleas , or ticks can acquire or transmit Bartonella spp . [14] . Few studies have described the kinetics of elimination and the details of transmission of these bacteria . The results of our experiments are in agreement with many experimental infection models , such as the experimental infection of fleas with B . quintana [48] , where they found that B . quintana was detected in the beginning of the 3rd dpi , in fleas , as in our bed bug experimental model , . We also found that B . quintana was viable in feces and decreased gradually after the 3rd dpi , which was similarly observed using the experimental cat flea B . quintana infection model [17] . Concerning the detection of B . quintana in eggs , L1 and L2 larvae , the vertical transmission of Bartonella species was suggested to occur , but the transmission routes were unknown [49] . Using IHC , in the four specimens we localized the bacterium to the digestive tract but not in the ovary . The presence of B . quintana in eggs , L1 and L2 larvae may be , due to vertical non-transovarial or horizontal transmission . In our context , the transmission may have occurred by external contact of the eggs , L1 and L2 larvae with the viable B . quintana released in adult’s feces which could be strongly considered as horizontal transmission . However Morick et al , demonstrate that Bartonella-positive flea feces and gut voids are proper infection sources for flea larvae and indicate that is considered as vertical non transovarial transmission [49] . In conclusion , we showed that the bed bug C . lectularius can acquire B . quintana by feeding and release viable organisms into their feces . Therefore , bed bugs may play a role as vectors of trench fever or other diseases caused by B . quintana . Knowing that stringent criteria exist in biomedical research for indicting the roles of living agents as biologically significant reservoirs and/or vectors of pathogens [50] , more studies are required to better understand B . quintana persistence in both bed bugs and their feces and to understand the potential vector role of bed bugs in B . quintana other bacterial infections . | Bartonella quintana , the etiologic agent of trench fever and other human diseases , is known to be transmitted by the feces of body lice . Recently , the DNA of this bacterium has been detected in bed bugs . Several pathogens have been associated and suggested to be transmitted by bed bugs , despite the insufficient evidence to support this vector role . The aim of the present study was to assess the competence of bed bugs in the transmission of B . quintana using an experimental artificial model of infection . To this end , bed bugs were fed with human infected blood meals . On the 3rd day post-infection ( dpi ) B . quintana was detected molecularly in 100% of experimentally infected bed bug . The bacterium was also detectable in bed bug feces starting on the 3rd dpi and persisted until 18±1 dpi . Although immunohistochemistry assays localized the bacteria to the gastrointestinal bed bug gut , B . quintana was also detected in the first and second instars larva . The present work highlights the need to reconsider monitoring of bed bugs for the transmission of pathogens . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | Competence of Cimex lectularius Bed Bugs for the Transmission of Bartonella quintana, the Agent of Trench Fever |
The intestinal microbiota plays important roles in digestion and resistance against entero-pathogens . As with other ecosystems , its species composition is resilient against small disturbances but strong perturbations such as antibiotics can affect the consortium dramatically . Antibiotic cessation does not necessarily restore pre-treatment conditions and disturbed microbiota are often susceptible to pathogen invasion . Here we propose a mathematical model to explain how antibiotic-mediated switches in the microbiota composition can result from simple social interactions between antibiotic-tolerant and antibiotic-sensitive bacterial groups . We build a two-species ( e . g . two functional-groups ) model and identify regions of domination by antibiotic-sensitive or antibiotic-tolerant bacteria , as well as a region of multistability where domination by either group is possible . Using a new framework that we derived from statistical physics , we calculate the duration of each microbiota composition state . This is shown to depend on the balance between random fluctuations in the bacterial densities and the strength of microbial interactions . The singular value decomposition of recent metagenomic data confirms our assumption of grouping microbes as antibiotic-tolerant or antibiotic-sensitive in response to a single antibiotic . Our methodology can be extended to multiple bacterial groups and thus it provides an ecological formalism to help interpret the present surge in microbiome data .
Recent advances in metagenomics provide an unprecedented opportunity to investigate the intestinal microbiota and its role in human health and disease [1] , [2] . The analysis of microflora composition has a great potential in diagnostics [3] and may lead to the rational design of new therapeutics that restore healthy microbial balance in patients [4]–[6] . Before the clinical translation of human microbiome biology is possible , we must seek to thoroughly understand the ecological processes governing microbiota composition dynamics and function . The gastro-intestinal microbiota is a highly diverse bacterial community that performs an important digestive function and , at the same time , provides resistance against colonization by entero-pathogenic bacteria [7]–[9] . Commensal bacteria resist pathogens thanks to resources competition [1] , [8] , growth inhibition due to short-chain fatty acid production [10] , killing with bacteriocins [11] , [12] and immune responses stimulation [13] , [14] . However , external challenges such as antibiotic therapies can harm the microbiota stability and make the host susceptible to pathogen colonization [15]–[20] . Despite its importance to human health , the basic ecology of the intestinal microbiota remains unclear . A recent large-scale cross-sectional study proposed that the intestinal microbiota variation in humans is stratified and fits into distinct enterotypes , which may determine how individuals respond to diet or drug intake [21] . Although there is an ongoing debate over the existence of discrete microbiome enterotypes [22] , they could be explained by ecological theory as different states of an ecosystem [23] . Ecological theory can also explain how external factors , such as antibiotics , may lead to strong shifts in the microbial composition . A recent study that analyzed healthy adults undergoing consecutive administrations of the antibiotic ciprofloxacin , showed that the gut microbiota changes dramatically by losing key species and can take weeks to recover [24] . Longitudinal studies , such as this one , suggest that many microbial groups can have large and seemingly random density variations in the time-scale of weeks [25] , [26] . The observation of multiple microbial states and the high temporal variability highlight the need for ecological frameworks that account for basic microbial interactions , as well as random fluctuations [27]–[29] . Here we propose a possible model to study how the intestinal microbiota responds to treatment with a single antibiotic . Our model expands on established ecological models and uses a minimal representation with two microbial groups [30] representing the antibiotic-sensitive and antibiotic-tolerant bacteria in the enteric consortium ( Fig . 1 ) . We propose a mechanism of direct interaction between the two bacterial groups that explains how domination by antibiotic-tolerants can persist even after antibiotic cessation . We then develop a new efficient framework that deals with non-conservative multi-stable field of forces and describes the role played by the noise in the process of recovery . We finally support our model by analyzing the temporal patterns of metagenomic data from the longitudinal study of Dethlefsen and Relman [24] . We show that the dynamics of microbiota can be qualitatively captured by our model and that the two-group representation is suitable for microbiota challenged by a single antibiotic . Our model can be extended to include multiple bacterial groups , which is necessary for a more general description of intestinal microbiota dynamics in response to multiple perturbations .
We model the microbiota as a homogeneous system where we neglect spatial variation of antibiotic-sensitive ( ) and antibiotic-tolerant ( ) bacterial densities . Their evolution is determined by growth on a substrate and death due to natural mortality , antibiotic killing and social pressure . With these assumptions , we introduce , as a mathematical model , two coupled stochastic differential equations for the density of sensitives and tolerants ( and ) normalized with respect to the maximum achievable microbial density: ( 1 ) ( 2 ) In the physics literature these types of equations represent stochastic motion in a non-conservative force field . The first terms in correspond to the saturation growth terms representing the indirect competition for substrate and depend on , which is the ratio of the maximum specific growth rates between the two groups . If tolerants grow better than sensitives on the available substrate and the reverse is true for . They effectively describe a microbial system with a growth substrate modeled as a Monod kinetic [31] in the limit of quasi-steady state approximation for substrate and complete consumption from the microbes ( see Methods for details ) . Both groups die with different susceptibility in response to the antibiotic treatment , which is assumed to be at steady-state . defines the ratio of the combined effect of antibiotic killing and natural mortality rates between the two groups ( see Methods for details ) . While the system can be studied in its full generality for different choices of , we consider the case of because it represents the more relevant case where sensitives are more susceptible to die than tolerants in the presence of the antibiotic . A possible that mimics the antibiotic treatments is a pulse function . With this , we are able to reproduce realistic patterns of relative raise ( fall ) and fall ( raise ) of sensitives ( tolerants ) due to antibiotic treatment as we show in Fig . S4 in the Supporting Information Text ( Text S1 ) . Additionally , we introduce the social interaction term between the two groups , , to implement competitive growth inhibition [13] , [32] . In particular , we are interested in the case where the sensitives can inhibit the growth of the tolerants ( ) , which typically occurs through bacteriocin production [33] . Finally we add a stochastic term that models the effect of random fluctuations ( noise ) , such as random microbial exposure , which we assume to be additive and Gaussian . The analysis can be generalized to other forms of noise such as multiplicative and coloured . We first analyzed the model in the limit of zero noise , . In this case , we were interested in studying the steady state solutions that correspond to the fixed-points of equations ( 1 , 2 ) and are obtained imposing . We found three qualitatively-distinct biologically meaningful states corresponding to sensitive domination , tolerant domination and sensitive-tolerant coexistence ( see Text S1 ) . We evaluated the stability of each fixed point ( see Text S1 ) and identified three regions within the parameter space ( Fig . 2A ) . In the first region the effect of antibiotics on sensitive bacteria is very low ( ) and domination by sensitives is the only stable state ( sensitives monostability ) . In the second region the effect of the antibiotic on sensitives is stronger than their inhibition over tolerants ( ) and the only stable state is domination by tolerants ( tolerants monostability ) . Finally , in the third region ( ) both sensitive and tolerant dominations are possible and stable , while the third coexistence fixed point is unstable ( bistability ) ( see Text S1 ) . This simple analysis shows that multistability can occur in a gut microbiota challenged by an antibiotic where one group directly inhibits the other ( i . e . through the term ) . Furthermore , it suggests that multistability is a general phenomenon since it requires only that antibiotic-sensitive and antibiotic-tolerant bacteria have similar affinities to nutrients . This is a realistic scenario because tolerants , such as vancomycin resistant Enterococcus [18] , are often closely related to other commensal but antibiotic-sensitive strains and therefore should have similar affinity to nutrients . Finally , the solution of equations ( 1 ) and ( 2 ) reveals that hysteresis is present for values of and leading to multistability ( Fig . 2B ) . Similarly to magnetic tapes , such as cassette or video tapes , which remain magnetized even after the external magnetic field is removed ( i . e . stopping the recording ) , a transient dose of antibiotics can cause a microbiota switch that persists for long time even after antibiotic cessation . The previous analysis shows the existence of multistability in the absence of noise . However , the influx of microbes from the environment and/or the intra-population heterogeneity are expected in realistic scenarios and affect the bacterial density evolution in a non-deterministic fashion . This raises the question of how the noise alters the deterministic stable states and their stability criteria . We assume that the noise is a fraction of bacteria that can be added ( or removed ) at each time step , but on average has no effect since . This assumption is justified by the fact that a persistent net flux of non-culturable bacteria from the environment is unrealizable . We also assume that this random event at time is not correlated to any previous time , which corresponds to , where characterizes the noise amplitude and is the Dirac delta function . We calculated the stationary probability of the microbiota being at a given state by solving the stationary Fokker-Planck Equation ( FPE ) [34] corresponding to the Langevin equations ( 1 , 2 ) : ( 3 ) By numerically solving equation ( 3 ) as described in [35] , for increasing , we find that for small values of the most probable states coincide with the deterministic stable states given by ( Fig . 3A ) . However , by increasing the distribution spreads and the locations of the most probable states change and approach each other . As a consequence , the probability of an unstable coexistence , characterized by and , increases thus avoiding extinction . This intuitively justifies how recovery to a sensitive-dominated state within a finite time after antibiotic cessation becomes possible with the addition of the noise . Without noise , the complete extinction of sensitive bacteria would have prevented any possible recolonization of the intestine . Beyond a critical noise level ( ) bistability is entirely lost and the probability distribution becomes single-peaked with both bacterial groups coexisting . The microbiota composition at the coexistence state can be numerically determined from the solution of , as shown in Fig . 3B and Video S1 . Further investigations based on analytical expansion of the Langevin equations ( see Methods ) show that for small random fluctuations , , the first noise-induced corrections to the deterministic density are linearly dependent on with a proportionality coefficient determined by the nature of the interactions ( insets in Fig . 3B ) . These linear correction terms can be obtained as a function of the model parameters and , after substituting a particular set of values in the bistable region ( and ) , they are for sensitives and for tolerants . These numbers are different from those reported in the insets of Fig . 3B . However the discrepancy is due to the propagation of the boundary conditions when numerically solving the solution of the FPE using finite elements ( see Text S1 ) . This has important biological implications since it suggests that extinction is prevented and , more importantly , that a minority of environmental microbes can settle in the gut at a rate that depends on the strength of their social interaction with the established microbiota . The introduction of random perturbation affects the stability criteria of the stable states . In particular , we observe that the bistability region decreases when the noise amplitude increases ( Fig . 2C–F ) . At the limit , when the bistability is entirely lost and the only stable state is the one where both groups coexist . This concept was previously hypothesized but not explicitly demonstrated in a model of microbial symbionts in corals [30] . Our model predicts that in absence of stochastic fluctuations the recovery time is larger than any observational time-scale so that it is impossible to revert to the conditions preceding antibiotic perturbation ( see Fig . S4 in Text S1 ) . In reality , data show that this time can be finite and depends on the microbiota composition and the degree of isolation of the individuals [18] , [24] , [36] . Thus , we aim to quantitatively characterize how the relative contribution of social interaction and noise level affects the computation of the mean residence time . In order to determine the relative time spent in each domination state , we compute the probability of residence in each stable state using master equations [34] . This method is more efficient than simulating the system time evolution by direct integration of the Langevin equations because it boils down to solving a deterministic second-order differential equation . Furthermore , this approach scales up well when the number of microbial groups increases , in contrast to the numerical solution of the FPE which can become prohibitive when . In our model , the master equations for the probability of residing in the tolerant or sensitive domination state are: ( 4 ) where is the transition rate from state to , which can be obtained in terms of the sum over all the state space trajectories connecting to . By solving this system of equations at steady-state , we obtain the residence probabilities and . After computing the transition rate as a function of the parameters , as reported in the Methods , we determine , which is our theoretical prediction for the mean relative residence time spent in the tolerant domination state ( Fig . 4 ) . The theoretical predictions are in good agreement with those obtained by simulating the dynamics multiple times and averaging over different realizations of the noise . A first consequence from this analysis is that the time needed to naturally revert from the altered state depends exponentially on the noise amplitude ( ) . As such , we predict that for the case of an isolated system ( ) the switching time is exponentially larger than any other microscopic scale and the return to a previous unperturbed state is very unlikely . On the contrary , as the level of random exposure is increased , the time to recover to the pre-treated configuration decreases ( see Fig . S4 in Text S1 ) . Additionally , this method can be considered as a way to indirectly determine the strength of the ecological interactions between microbes which can be achieved by measuring the amount of time that the microbial population spends in one of the particular microbiota states . Therefore , it can potentially be applied to validate proposed models of ecological interactions by comparing residence times measured experimentally with theoretical predictions . We now focus on the dynamics of bacteria detected in the human intestine and test the suitability of our two-group representation by re-analyzing the time behaviour in the recently published metagenomic data of Dethlefsen and Relman [24] . The data consisted of three individuals monitored over a 10 month period who were subjected to two courses of the antibiotic ciprofloxacin . Since the data are noisy and complex , and the individual subjects' responses to the antibiotic are distinct [24] , identifying a time behaviour by manual screening is not a trivial task . We do it by using singular value decomposition ( SVD ) to classify each subject phylotype-by-sample data matrix into its principal components . Because of inter-individual variability we obtain , for each subject , the right and left eigenvectors associated to each eigenvalue . By ranking the phylotypes based on their correlation with the first two components we recover characteristic temporal patterns for each volunteer [37] , [38] . In all three subjects , we observe that , in spite of the individualized antibiotic effect , the two dominant eigenvalues or principal components together capture about 70% of the variance observed in the data ( Fig . 5A–C ) . Invariably , the first component shows a decrease in correspondence to antibiotic treatment and reflects the behaviour of antibiotic-sensitive bacteria ( green line in Fig . 5D–F ) . Conversely , the second component increases with the antibiotic treatment and represents antibiotic-tolerants ( red line in Fig . 5D–F ) . The observation that each subject's microbiota can be decomposed into two groups of bacteria with opposite responses to antibiotics supports the validity of the two-group approach used in our model . Classification of each individual's phylotypes as sensitive or tolerant can be obtained by determining their correlation with the two principal components ( see Text S1 ) ( information in the right-eigenarrays matrix from SVD ) . Bacteria correlated with component 1 are usually highly abundant before antibiotic treatment and drop strongly during treatment , often below detection . Vice-versa , bacteria correlated with component 2 are typically in low abundance before the antibiotic and increase with antibiotic administration ( Fig . 5G–I ) . Interestingly , despite significant inter-individual differences in recovery time ( Fig . 5G–I ) and individualized response of each subject , the data show that in each individual the majority of bacteria are antibiotic-sensitive and only a small but significant fraction are tolerant to ciprofloxacin ( see Text S1 ) . The recognition of these time-patterns could be considered as a possible tool to indirectly determine the susceptibility of non-culturable commensal bacteria to FDA-approved antimicrobial compounds . However , the presence of strains in the same phylotypes that display both behaviors in response to the drug may constitute a significant challenge for the success of this method . The time evolution of the phylotypes ( Fig . 5G–I ) qualitatively agrees with our theoretical prediction that after the antibiotic administration the system moves fast , meaning in a time smaller than any other observable time-scale , into a new stable state with less sensitives and more tolerants . Further , the data also suggest that the return to sensitive domination happens after a recovery-time scale that depends on the microbial composition .
We present a model of inter-bacterial interactions that explains the effect of antibiotics and the counter-intuitive observation that an antibiotic-induced shift in microbiota composition can persist even after antibiotic cessation . Our analysis predicts a crucial dependence of the recovery time on the level of noise , as suggested by experiments with mice where the recovery depends on the exposure to mice with untreated microbiota [18] . The simple model here introduced is inspired by classical ecological modeling such as competitive Lotka-Volterra models [39] , [40] , but relies on mechanistic rather than phenomenological assumptions , such as the logistic growth . Although more sophisticated multi-species models include explicit spatial structure to describe microbial consortia [33] , [41]–[43] , our model is a first attempt to quantitatively analyze the interplay between microbial social interactions ( ) and stochastic fluctuations ( ) in the gut microbiota . We find that these two mechanisms are the key ingredients to reproduce the main features of the dynamics in response to antibiotic ( sudden shifts and recovery ) . Our model can be easily generalized to include spatial variability and more complicated types of noise . Therefore we provide a theoretical framework to quantify microbiota resilience against disturbances , which is an importance feature in all ecosystems [44] . By introducing a new stochastic formulation , we were able to characterize composition switches within the context of state transition theory [45] , [46] , an important development over similar ecological models of microbial populations [30] . We present a new method to calculate the rate of switching between states that identifies the most likely trajectory between two stable states and their relative residence time , which can be subjected to experimental validation . Finally , we apply SVD to previously published metagenomic data [24] , which allows us to classify the bacteria of each subject in two groups according to their temporal response to a single antibiotic . The SVD method has been used before to find patterns in temporal high-throughput data , including transcription microarrays [37] and metabolomics [47] . Although our approach seems to capture well the main temporal microbiota patterns , we should note that the use of the Euclidean distance as a metric for microbiome analysis presents limitations and recent studies have proposed alternative choices [48]–[50] . We also opt for an indirect gradient analysis method [51] because we are interested in emergent patterns from the data regardless of the measurements of the external environmental variable ( i . e . presence or absence of the antibiotic ) [50] . We propose a mechanism of interaction between two bacterial groups to explain the lack of recovery observed in the experiments that can be validated in the near future . Although training the model with the available data sets would be of great interest , this will not be useful in practice because we need more statistical power to be predictive . However , we anticipate that a properly validated mathematical model of the intestinal microbiota will be a valuable tool to assist in the rational design of antibiotic therapies . For example , we predict that the rate of antibiotic dosage will play a crucial role . In order to let the microbiota recover from antibiotic treatment , it is better to gradually decrease antibiotic dosage at the first sign of average microbiota composition change , which has to be larger than the threshold community change represented by the day-to-day variability [26] , rather than waiting for tolerant-domination and then stopping antibiotic treatment . We show here the application of our theory to a two-bacterial group scenario because we are interested in the microbiota response when challenged with a single antibiotic . However , in more realistic conditions the microbiota is subjected to different types of perturbations , which may drive it towards more alternative stable states . Our theory of the microbial-states switches characterization can be naturally extended to more than two states and consists of the solution of the linear system of equations , where is the array of probability of residing in each stable state and is the matrix of transition rates among the states . The ongoing efforts to characterize the microbial consortia of the human microbiome can yield tremendous benefits to human health [52]–[55] . Within the next few years , we are certain to witness important breakthroughs , including an increase in the number of microbiomes sequenced as well as in sequencing depth . Yet , without the proper ecological framework these complex ecosystems will remain poorly understood . Our study shows that , as in other complex microbial ecosystems , ecological models can be valuable tools to interpret the dynamics in the intestinal microbiota .
The model introduced in equations 1 and 2 is derived from the more detailed model described below . We model the bacterial competition in a well-mixed system in the presence of antibiotic treatment by means of the following stochastic differential equations: ( 5 ) where we account for two bacterial groups; the intestinal resident sensitive flora and an antibiotic tolerant one . Additionally , we also consider the substrate and the antibiotic densities . The antibiotic time evolution is simply a balance between inflow and outflow ( i . e . no decay due to microbial degradation ) where is the system's dilution rate , which sets the characteristic microscopic time-scale , and is the constant density of the incoming antibiotic , which can be time dependent . Similarly the substrate concentration , , results from a mass balance from influx and microbial consumption . As for the antibiotic , is the constant density of the incoming nutrient ( i . e . the concentration of resources coming from the small-intestine ) . The second and third terms in the right-hand side of the second equation in ( 5 ) describe the amount of substrate consumed by bacterial growth assuming Monod kinetics where ( ) is the maximum growth rate for sensitives ( tolerants ) , is the half-saturation constant for growth , which parametrizes the bacterial affinity to the nutrient , and ( ) is the yield for growth for sensitives ( tolerants ) . The last two equations describe how sensitives and tolerants grow on the substrate available and are diluted with the factor . We mimic the effect of the antibiotic on the sensitives adding a term proportional to the sensitive density where the constant of proportionality is the antibiotic-killing rate . We also introduce a direct inhibition term , which mimics the inhibition of sensitive bacteria on the tolerants ( social interaction ) . Finally the Gaussian random variables , are the additive random patterns of exposure and represent the random microbial inflows ( outflows ) from ( to ) the external environment . It is convenient to scale the variables and set the dilution rate to unity ( ) . Therefore , all the rates have to be compared with respect to the system characteristic dilution rate . Introducing , , , , , , , , , and and dropping the tilde symbols , we obtain the following dimensionless model: ( 6 ) If we assume that the antibiotic is a fast variable compared to the microbial densities ( ) ( i . e . the time-scale at which the antibiotic reaches stationary state is smaller than that of the bacteria ) , we can solve for and obtain . If we also assume that the incoming substrate is all consumed in microbial growth , therefore maintaining the population in a stationary state with respect to the available resources , and that , similarly to the antibiotic , the resources equilibrate much faster than the bacterial densities ( quasi-steady state assumption , ) , we obtain that: ( 7 ) If we now define a new parameter describing the relative ratio of the combination of antibiotic killing and natural mortality ( i . e . wash-out ) between sensitives and tolerants , the model reduces to the two variables model in reported in equations ( 1–2 ) . The introduction of random noise has the important consequence of changing the composition of the stable states ( Fig . 3A ) . In order to characterize this phenomenon , we expand the solution of the Langevin equations ( 1–2 ) around one of the stable states obtaining the following set of equations for the variable : ( 8 ) where to simplify the notation we drop the explicit time-dependence . We can easily recognize the first derivative of the force on the right-hand side as the Jacobian matrix computed in one of the minima . This equation can be solved order by order by defining the expansion and writing the equations for each order as: ( 9 ) ( 10 ) Assuming that the initial condition at time zero is , which can always be neglected for long-term behaviour , the solution of equation ( 9 ) is ( 11 ) This means that the average location of the minima at zero order is not modified by the noise since . By computing the solution of the equation ( 10 ) we similarly find that: ( 12 ) The long-time average value of the first order correction now reads: ( 13 ) The time integral can be easily computed assuming that the eigenvalues of are negative , or at least their real part is , as it should be for stable fixed points; therefore we obtain that: ( 14 ) Thus , we find that the effect of random fluctuations is to correct the value of the stable points as if an external field , proportional to strength of the fluctuations , was present . This field is equal to the mean square displacement at large time opportunely weighted by the inverse of the curvature of the bare potential around the stable points , . The correlation can be now computed using equation ( 11 ) and reads: ( 15 ) Since the previous equation simplifies to ( 16 ) which results in . The mean residence time in each state is proportional to the residence probability defined in equation ( 4 ) . To obtain it , we need to compute the transition rate as a function of the model parameters as: ( 17 ) where and are the initial and final time and is the functional integral over the trajectory . Each time trajectory , solution of equations ( 1–2 ) , has an associated weight , defined as: ( 18 ) By discretizing the time so that with and the microscopic time step , we obtain that the Langevin equations can be written using the Ito prescription [56] as: ( 19 ) where we use the short notation and the initial value is . The time discretization allows us to interpret the functional integral in equation ( 18 ) as: ( 20 ) Since the noise is Gaussian and white , its distribution now reads: ( 21 ) This can be justified using the property of the delta-function and its discrete time version so that follows and . Using the properties of the delta function , and integrating out all s , the continuous limit expression of equation ( 21 ) is ( 22 ) where has an intuitive interpretation in thermodynamics and it is related to the entropy production rate [57] . By using stationary-phase approximation , it turns out that in the computation of the rate defined in ( 17 ) only one path matters , , which is the most probable path . Higher order factors are proportional to the term [45] , [46] , and therefore simplify with the denominator in equation ( 21 ) . This comes from the fact that several almost optimal paths can be constructed starting from . In the optimal path , the system stays in a stable state for a very long time , then it rapidly switches to the other stable state where it persists until . By shifting the switching time one obtains sub-optimal paths that , at the leading order in , give the same contribution of the optimal one and their number is directly proportional to . This leads to ( 23 ) The functional Gaussian integral can be computed [45] , [46] and only provides a sub-leading correction to the saddle-point contribution resulting in the transition rate formula , which is reported in the Results section . We now need to determine the optimal path and its associated action . This path is defined as the one where the functional derivative of is set to zero such that the initial and final states are fixed . This produces a set of second-order differential equations ( 24 ) which can be solved imposing the initial conditions on and . It is easy to verify that the downhill solution is and it is associated with null action . Meanwhile , the ascending trajectory , which is the one leading to a non-zero action and hence gives the transition rate value , is not given by , as it would be for conservative field of forces . This means that in presence of a dissipative term the reverse optimal path from the minimum to the maximum is different with respect to the one connecting the maximum from the minimum of the landscape . As the last point , we want to show that the action associated to the optimal path can be further simplified by noticing that ( 25 ) We can easily prove this condition by showing that the time derivative vanishes when equation ( 24 ) is satisfied and remembering that the optimal path connects two stable states where and . This property allows us to rewrite the action as: ( 26 ) We solved numerically the equation ( 24 ) using a trial-and-error approach . We varied the first-derivative at initial time in order to arrive as close as possible to the final point within some numerical precision . In principle the ideal trajectory connecting two stable points should be computed in the limit of but this trajectory will take infinite time . We report three examples of most probable paths connecting the points to and reverse for a chosen set of in Fig . S6 of the Text S1 . We first rarefy the raw phylotypes counts matrix as in [24] . We then normalize the logarithm of the counts according to the following procedure: 1 ) we add one to all the phylotypes counts to take into account also for the non-detected phylotypes in each sample , 2 ) we log-transform the data and 3 ) we normalize the resulting matrix with respect to the samples averages . In formulae , the count associated to phylotype in sample for each subject iswhere is the average value of the counts in each sample and is the total number of phylotypes . Among all possible normalization schemes , we decide to subtract the column averages because we aim at identifying patterns within samples based on their correlation in bacterial composition . Indeed , the covariance matrix of the samples is proportional to , where is the transpose matrix . SVD on the matrix is thus equivalent to the principal component analysis ( PCA ) performed on the samples covariance matrix . | Recent applications of metagenomics have led to a flood of novel studies and a renewed interest in the role of the gut microbiota in human health . We can now envision a time in the near future where analysis of microbiota composition can be used for diagnostics and the rational design of new therapeutics . However , most studies to date are exploratory and heavily data-driven , and therefore lack mechanistic insights on the ecology governing these complex microbial ecosystems . In this study we propose a new model grounded on ecological and physical principles to explain intestinal microbiota dynamics in response to antibiotic treatment . Our model explains a hysteresis effect that results from the social interaction between two microbial groups , antibiotic-tolerant and antibiotic-sensitive bacteria , as well as the recovery allowed by stochastic fluctuations . We use singular value decomposition for the analysis of temporal metagenomic data , which supports the representation of the microbiota according to two main microbial groups . Our framework explains why microbiota composition can be difficult to recover after antibiotic treatment , thus solving a long-standing puzzle in microbiota biology with profound implications for human health . It therefore forms a conceptual bridge between experiments and theoretical works towards a mechanistic understanding of the gut microbiota . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"ecosystem",
"modeling",
"ecology",
"biology",
"computational",
"biology",
"microbiology",
"microbial",
"ecology"
]
| 2012 | Social Interaction, Noise and Antibiotic-Mediated Switches in the Intestinal Microbiota |
Wnt signaling provides a paradigm for cell-cell signals that regulate embryonic development and stem cell homeostasis and are inappropriately activated in cancers . The tumor suppressors APC and Axin form the core of the multiprotein destruction complex , which targets the Wnt-effector beta-catenin for phosphorylation , ubiquitination and destruction . Based on earlier work , we hypothesize that the destruction complex is a supramolecular entity that self-assembles by Axin and APC polymerization , and that regulating assembly and stability of the destruction complex underlie its function . We tested this hypothesis in Drosophila embryos , a premier model of Wnt signaling . Combining biochemistry , genetic tools to manipulate Axin and APC2 levels , advanced imaging and molecule counting , we defined destruction complex assembly , stoichiometry , and localization in vivo , and its downregulation in response to Wnt signaling . Our findings challenge and revise current models of destruction complex function . Endogenous Axin and APC2 proteins and their antagonist Dishevelled accumulate at roughly similar levels , suggesting competition for binding may be critical . By expressing Axin:GFP at near endogenous levels we found that in the absence of Wnt signals , Axin and APC2 co-assemble into large cytoplasmic complexes containing tens to hundreds of Axin proteins . Wnt signals trigger recruitment of these to the membrane , while cytoplasmic Axin levels increase , suggesting altered assembly/disassembly . Glycogen synthase kinase3 regulates destruction complex recruitment to the membrane and release of Armadillo/beta-catenin from the destruction complex . Manipulating Axin or APC2 levels had no effect on destruction complex activity when Wnt signals were absent , but , surprisingly , had opposite effects on the destruction complex when Wnt signals were present . Elevating Axin made the complex more resistant to inactivation , while elevating APC2 levels enhanced inactivation . Our data suggest both absolute levels and the ratio of these two core components affect destruction complex function , supporting models in which competition among Axin partners determines destruction complex activity .
Cell-cell signaling is critical for cell fate decisions during embryonic development and cell fate maintenance during adult homeostasis . Altered signaling by these same pathways underlies most solid tumors . The Wnt signaling pathway provides a paradigm—it regulates cell fate choice in tissues throughout the body , maintains stem cell identity in many adult tissues , and is inappropriately activated in colorectal and other cancers [1] . Thus , understanding the mechanisms by which signaling occurs and is regulated are key issues for cell , developmental , and cancer biology . Work in both animal models and cultured mammalian cells provided a broad outline of Wnt signaling and its regulation [2 , 3] . The key effector is the transcriptional co-activator β-catenin ( βcat; Drosophila Armadillo; Arm ) . In the absence of signaling , βcat is captured by a multiprotein complex called the destruction complex . The scaffold proteins Adenomatous polyposis coli ( APC ) and Axin bind βcat and present it to the kinases glycogen synthase kinase-3 ( GSK3 ) and casein kinase 1 ( CK1 ) . They phosphorylate βcat , creating a binding site for an E3 ubiquitin ligase , thus targeting βcat for proteasomal destruction . When Wnt ligands bind to receptors , the destruction complex is downregulated , allowing βcat to accumulate , enter the nucleus and act together with the DNA binding proteins in the TCF/LEF family to transcriptionally activate Wnt-regulated genes . Work in cultured mammalian cells has added important aspects to this model [2]—here we focus on the action of the destruction complex and its regulation by Wnt signaling . Several different mechanisms have emerged by which Wnt signaling downregulate βcat destruction and thus activate downstream signaling . Wnt binding to the Frizzled:LRP5/6 receptors triggers assembly of the receptors along with the Wnt effector Dishevelled ( Dvl; Drosophila Dsh ) into a higher order signalosome [4–7] . LRP5/6 becomes phosphorylated and recruits the destruction complex to the plasma membrane , at least in part by interactions between the phosphorylated tail of LRP5/6 and Axin[8] . The phosphorylated LRP5/6 tail can directly inhibit GSK3 [9] . Alternate mechanisms for destruction complex inhibition also exist . Dsh can co-polymerize with Axin via their shared DIX domains , antagonizing its function [10] . Careful kinetic analysis revealed that Wnt stimulation reduces the rate of ßcat phosphorylation by both CK1 and GSK3 , reducing but not eliminating destruction complex activity [11] . Wnt signaling can trigger Axin dephosphorylation , reducing its interaction with both ßcat and LRP5/6 , thus reducing ßcat destruction [12] . Finally , another study suggested that after Wnt signaling the destruction complex remains intact and capable of phosphorylating βcat , but its transfer to the E3 ligase is prevented [13] . These studies provide important insights into key regulatory mechanisms by which Wnt signaling can inactivate the destruction complex , but leave as an open question which mechanism ( s ) is most prominent during signaling in vivo . The Wnt pathway is part of an emerging theme in cell signaling , in which self-assembly of multiprotein supramolecular signaling hubs creates non-membrane bound cellular compartments [14] . Three key steps in Wnt signaling are catalyzed by distinct supramolecular machines—the signalasome , involved in Wnt reception and destruction complex downregulation , the destruction complex itself , and the enhancesome , which mediates Wnt-regulated gene expression [3] . Key questions remain about the mechanism by which the active destruction complex targets ßcat for destruction in the absence of Wnt signaling . APC was originally viewed as the scaffold around which the destruction complex assembled , but subsequent work revealed that Axin fulfills this function , leaving APC’s molecular role a mystery . Further , while the destruction complex is typically represented in models as a simple four-protein complex , considerable evidence supports the idea that it is a large supramolecular protein assembly , built by self-polymerization of Axin and APC ( e . g . , [15–18] ) . Recent work provided new mechanistic insights into the molecular mechanisms by which APC functions , helping begin to transform the static , low-resolution textbook model of Wnt signaling into a more dynamic , high resolution view . Super-resolution microscopy of Axin and APC complexes assembled after overexpression in colorectal cancer cells provided the first look inside the active destruction complex . Axin and APC containing “puncta” were resolved into intertwined strands of each protein , presumably assembled by polymerization [17] . Combining this with assessment of APC and Axin dynamics and genetic and biochemical dissection of the two proteins provided novel mechanistic insights and a new model . First , they suggest APC promotes/stabilizes Axin multimerization , thus increasing destruction complex efficiency [17] . Second , they revealed a key role for two peptide motifs in APC , 20 amino acid repeat 2 and sequence B/the CID , both essential for destruction complex function [19 , 20] . These motifs appear to play two roles . They are binding sites for alpha-catenin , stabilizing ßcat association with APC and preventing its dephosphorylation [21] . After Axin-mediated βcat phosphorylation , these APC motifs are also phosphorylated , triggering a regulated conformational change that transfers βcat out of the destruction complex to the E3 ligase , to restart the catalytic cycle [17] . These data fit with other studies suggesting that Wnt signaling does not totally turn off the destruction complex , but reduces the rate of destruction . Instead , the destruction complex remains intact and capable of phosphorylating βcat , but βCat transfer to the E3 ligase is inhibited [11 , 12 , 13] . However , this work was largely done in cultured cells , which provide a simple place to explore pathway circuitry but do not provide a physiologically relevant situation with all regulatory mechanisms intact . We thus took these insights back into the Drosophila embryonic epidermis , arguably the system where our understanding of the roles and regulation of the Wnt pathway is strongest . Stripes of cells in each body segment produce a fly Wnt , Wingless ( Wg ) , creating a field of cells experiencing high , moderate and low levels of Wg signaling . Taking advantage of new genetic approaches and high-resolution microscopy , we addressed several key issues in the field , exploring the structure , assembly and stoichiometry of the destruction complex in vivo during normal development and how it is downregulated by Wnt signaling . To understand a complex multiprotein machine , one key issue involves the relative levels of its component parts . Most current models of Wnt regulation suggest Axin accumulates at levels dramatically lower than those of other proteins in the destruction complex . This hypothesis derives from influential early work in Xenopus oocyte extracts . By adding in known amounts of recombinant Axin and measuring the resulting destruction complex activity , they estimated Axin concentrations were as much as 5000-fold lower than those of APC and other destruction complex proteins . Their mathematical model of Wnt signaling and many subsequent ones are based on these estimates [22–24] . In contrast , recent work in cultured mammalian cells suggests Axin and APC levels are more similar [25] . Thus , defining the relative levels of Axin and APC in tissues undergoing Wnt signaling in vivo is a key issue , and the Drosophila embryo provided a superb place to accomplish this end . With relative protein levels defined , different models for the function and regulation of the destruction complex can be tested by varying absolute levels of Axin or APC and their relative ratios to one another . Substantially elevating Axin levels in Drosophila embryos strongly inhibits Wnt signaling [26] . Further analyses suggested there is a threshold below which elevating Axin does not substantially alter signaling , since more subtle elevation of Axin levels ( 2–5 fold ) had little effect in Drosophila embryos or imaginal discs [27–29] and mutating tankyrase , which elevates Axin levels 2–3 fold , does not substantially perturb Wnt signaling [28 , 30] . In contrast , a 9-fold increase in Axin levels inhibited Wnt signaling in imaginal discs [27] . However , these studies used multiple tissues or systems in parallel , and left the mechanisms underlying the dose-sensitive response unclear . The Drosophila embryo provided a place to assess how altering Axin levels affects cell fate choice , Wnt-target gene expression and ßcat levels in parallel , and to directly compare effects on cells receiving and not receiving Wnt signals . It also offered the opportunity to manipulate APC levels , the other key scaffolding protein in the destruction complex . Whether APC levels are rate-limiting remains an open question , because APC has been viewed as present in substantial excess . The Drosophila embryo also allowed us to test effects of varying the Axin:APC ratio , another key parameter of any molecular model . Finally , to effectively understand destruction complex assembly and function , we need to visualize it directly . Our recent super-resolution imaging of Axin:APC puncta in cultured cells provided the first insights into the internal structure and dynamics of this multiprotein machine , but these experiments involved significant over-expression . The Drosophila embryo provided a place to assess whether similar complexes assemble at near endogenous levels . Recent advances in molecular counting technology also offered the possibility of directly assessing the number of Axin proteins assembled in a complex . Visualizing the destruction complex in the embryo would also allow us to address how Wnt signaling inactivates it . Work in cultured cells led to a model in which Wnt binding the Frizzled:LRP5/6 receptor complex triggers LRP5/6 phosphorylation , and Axin and Dvl/Dsh membrane recruitment [31] . What happens next is disputed , with many events suggested to play a part . For example , some data suggest the destruction complex is disassembled because Dsh competes for Axin [10] or Wnt signaling destabilizes Axin [32] . Interestingly , examining effects of Wg signaling on the destruction complex in Drosophila embryos led to starkly divergent conclusions . One group reported that Wg signaling strongly reduced Axin levels , as assessed both by immunofluorescence and immunoblotting [33] . A second , visualizing GFP-tagged Axin , found little or no effect of Wg on Axin levels—instead their data suggested that Wg signaling causes a Dsh-dependent relocalization of Axin from cytoplasmic puncta to the plasma membrane [26] . Finally , a third group reported that Wg signaling initially stabilizes Axin , as assessed by immunofluorescence , increasing both membrane bound and cytoplasmic pools [28 , 34] . Thus , the effects of Wg signaling on Axin , a key part of the mechanism underlying βcat stabilization , also remain an open question . Our system , allowing direct detection of fluorescently-tagged Axin expressed at near endogenous levels , allowed us to address this issue .
Most current models of Wnt regulation suggest Axin accumulates at levels dramatically lower than those of other destruction complex proteins , potentially making destruction complex activity sensitive to very small increases in its levels . However , the literature contains indications that this is not universally true ( e . g . [25] ) . To better understand how APC and Axin levels affect Wnt signaling in vivo we directly compared levels of APC family members and Axin in Drosophila embryos . We first compared mRNA levels of Drosophila axin with those encoding the two fly APC family proteins , APC1 and APC2 , using RNAseq data from staged embryos . In embryos , APC2 plays the predominant role in Wnt regulation during early to mid-embryogenesis ( [35–37]; 2–4 or 6–8 hours after egg laying , respectively ) , while APC1 is expressed at low levels early but becomes prominent later in the central nervous system [38 , 39] . Consistent with this , APC2 mRNA levels are ~19x higher than APC1 during early embryogenesis , and ~7x higher during mid-embryogenesis ( 484 versus 26 Fragments Per Kilobase of transcript per Million mapped reads ( FPKM ) , and 201 versus 27 FPKM , respectively ) . However , in late embryogenesis , as the nervous system is assembled , APC1 mRNA levels are ~5x more abundant than APC2 ( 120 vs . 23 FPKM ) . Since APC2 and APC1 can act redundantly in regulating Wnt signaling [36 , 37] , we compared axin mRNA levels with combined mRNA abundance of APC1 plus APC2 . Surprisingly , RNAseq reads for axin were roughly comparable to those of APC1 plus APC2 at three different stages of embryonic development ( Fig 1A ) , indicating that there are not dramatic differences between APC family members versus Axin at the mRNA level . These data did not rule out differences in protein translation or stability . To determine if similar mRNA levels led to similar protein levels , we compared Axin and APC2 protein levels in early to mid-embryogenesis ( 4–8 hrs ) , when APC2 is the predominant family member expressed . Since antibodies to APC2 and Axin may have different affinities , one cannot simply compare antibody-labeled endogenous proteins . To overcome this , we utilized GFP-tagged proteins expressed at near-endogenous levels . This allowed us to compare endogenous versus GFP-tagged Axin , or endogenous versus GFP-tagged APC2 proteins , using antibodies against the endogenous proteins , followed by comparing GFP-tagged Axin and GFP-tagged APC2 proteins , using anti-GFP antibodies . We used the GAL4-UAS system [40 , 41] to express Axin:GFP , using the driver that gave the lowest level of Axin:GFP expression ( act5c-GAL4 provided by male parents ) . Axin:GFP was expressed at 1 . 0±0 . 5 fold that of endogenous Axin , as assessed by immunoblotting with anti-Axin antibodies ( Fig 1B , S1 Table ) . We next used transgenic flies expressing GFP:APC2 under control of the endogenous APC2 promotor , in an APC2 null mutant background [19] . Using anti-APC2 antibodies , we re-confirmed that APC2-driven GFP:APC2 was expressed at the same level as endogenous APC2 ( 0 . 9±0 . 4 fold endogenous APC2; Fig 1C , S1 Table ) . To complete the comparison , we then compared APC2-driven GFP:APC2 to Axin:GFP driven by zygotic act5c-GAL4 . Immunoblotting with anti-GFP antibodies revealed that GFP:APC2 is expressed ~4-fold the levels of Axin:GFP ( Fig 1D; 4 . 3±1 . 4; S1 Table ) . These three comparisons—endogenous Axin to act5c-GAL4 driven Axin:GFP , act5c-GAL4 x Axin:GFP to APC2-driven GFP:APC2 , and APC2-driven GFP:APC2 to endogenous APC2—provided a reasonable estimate of the relative levels of endogenous APC2 to Axin: APC2 accumulates at a ~5-fold higher level than Axin ( 4 . 7±1 . 4 ) . This is in contrast to the 5000-fold difference in accumulation observed in Xenopus extracts that forms the basis of some current models , but is consistent with the similar levels of mRNAs revealed by RNAseq . Axin is the key scaffold on which the destruction complex is built , and thus most models of Wnt signaling suggest Axin is rate limiting for destruction complex function . Previous experiments in fly embryos and imaginal discs strongly support this , as over-expressing Axin can shut down Wnt signaling . Our knowledge of the relative levels of APC2 versus Axin in the Drosophila embryonic epidermis allowed us to confirm and extend the analysis . We first developed ways to vary Axin levels systematically , exploring how increasing Axin levels to different degrees altered viability , cell fate and expression of a Wg target gene . We next explored the underlying mechanism , by examining how different Axin levels affected destruction complex activity and ßcat levels , both in cells receiving and not receiving Wg signals . We then brought APC2 into this picture , examining effects of elevating APC2 levels , and of altering the ratios of Axin to APC2 . To manipulate Axin levels systematically , we used the GAL4-UAS system . Four crosses using two different GAL4 drivers provided different levels and timing of Axin over-expression ( S1 Fig; Methods; S1 Table ) . act5c-GAL4 is expressed during oogenesis and relatively ubiquitously during embryonic development . 1 . By crossing UAS-Axin:GFP females to act5c-GAL4/+ males , we achieved lower-level and later elevation of Axin:GFP levels , which was driven by zygotically-expressed GAL4 ( hereafter Zyg Axin ) . 2 . By crossing act5c-GAL4/+ females to UAS-Axin:GFP males ( hereafter Mat/Zyg Axin ) , we achieved relatively high-level overexpression , which began early due to maternally-contributed GAL4 and continued zygotically . The second GAL4 driver stock was MatGAL4 , which includes two GAL4 lines expressed during oogenesis; they are not expressed zygotically but maternally expressed GAL4 protein perdures in the embryo . 3 . For maternal and zygotic over-expression , we assessed progeny of females trans-heterozygous for MatGAL4 and UAS-Axin:GFP ( hereafter Mat Axin ) . 4 . To achieve levels of Axin elevation intermediate between that produced by Zyg Axin and Mat Axin , we used MatGAL4 to co-express UAS-Axin:GFP with a second UAS-driven transgene encoding RFP ( hereafter Mat RFP&Axin ) . When two different UAS-driven transgenes are present , this reduces expression of both transgenes . We directly measured protein levels by immunoblotting with antibodies to either GFP or to endogenous Axin . These four schemes produced an excellent range of Axin expression levels in stage 9 embryos , when Wnt signaling is at its peak . Zyg Axin effectively tripled normal Axin levels in embryos in which it was expressed ( Fig 2A–2C , S1 Table; taking into account endogenous Axin and the fact that only 50% of embryos inherit the GAL4 driver ) . Mat RFP&Axin led to an ~4-fold increase , while both Mat/Zyg Axin and Mat Axin led to 8–9 fold elevation in total Axin levels ( Fig 2A–2C , S1 Table ) . When we examined the pattern of Axin:GFP accumulation , we noted Mat/Zyg Axin led to substantially more variable expression from cell-cell than MatGAL4-driven Axin:GFP . Thus , in most subsequent functional assays we used Mat Axin for high-level overexpression . In addition to differences in expression levels , these lines also differed in timing of Axin:GFP expression ( Fig 2F ) . Zyg Axin levels started very low ( as expected with no maternal GAL4 expression ) and continued to rise throughout development . Mat Axin levels started somewhat higher ( driven by maternal GAL4 ) , increased during stages 9–11 ( 4–9 hrs ) and then slowly decayed . Mat/Zyg Axin exhibited initially modest Axin:GFP levels , which continued to rise throughout development . Mat RFP&Axin accumulation followed a similar expression pattern as Mat Axin , but at decreased levels due to the presence of two UAS-driven transgenes ( Fig 2F , right ) . These tools allowed us to vary Axin levels systematically , and we thus used them to assess how altering Axin levels and timing of accumulation affect Wg signaling and its regulation , by assessing effects on embryonic viability , cell fate choice , Wg target gene expression , and Arm ( fly βcat ) levels . We first assessed effects of elevating Axin levels on embryonic viability and cell fate choice—these assays integrate effects on Wnt signaling across embryonic development , and thus have to be interpreted in light of effects on Axin levels both at stage 9 and later , as Wg signaling affects cell fate choice through stage 11 ( 9 hours; [42] ) . The relatively subtle ( 3-fold ) Axin elevation produced by Zyg Axin at stages 9–11 did not result in embryonic lethality ( Fig 3A , S2 Table; 5% lethality vs . 3% lethality of wildtype controls ( controls carried UAS-Axin without a GAL4 driver ) ) . We then examined larval cuticles to look for more subtle effects on Wg signaling . Reducing Wg signaling affects cell fate , causing loss of naked cuticle fates and merger of denticle belts—Fig 3C illustrates the graded series of defects with successively reduced Wg signaling . Most Zyg Axin embryonic cuticles ( 3-fold increase ) were near wildtype ( Fig 3B and 3C , S3 Table ) , though the occasional defects seen suggest subtle reduction of Wg signaling in some embryos . Consistent with this possibility , no hatching Zyg Axin larvae survived to adulthood—this may reflect the fact that as a consequence of zygotic GAL4 expression Axin levels continued to rise throughout development ( Fig 1F ) . The slightly higher level expression of Axin:GFP in Mat RFP&Axin embryos ( 4-fold increase ) and the earlier onset of expression led to some embryonic lethality ( 32% lethal; Fig 3A; S2 Table ) , and a larger fraction of embryos had moderate inhibition of Wg signaling , as assessed by cell fate choices ( Fig 3B and 3C , S3 Table ) . In contrast , higher-level , earlier overexpression of Axin ( 8–9 fold ) led to substantial embryonic lethality—90% lethality for Mat/Zyg Axin and 78% lethality for Mat Axin ( Fig 3A; S2 Table ) . In both crosses , there were two genotypes of embryonic progeny; for Mat/Zyg Axin these differed by whether or not they had a zygotic copy of act5c-GAL4 and for Mat Axin by whether they had one or two copies of the UAS-Axin:GFP transgene zygotically ( S1 Fig ) . Cuticle analysis of cell fates revealed that Wg-signaling was strongly reduced in many Mat/Zyg Axin and Mat Axin progeny ( Fig 3B and 3C , S3 Table ) , but there were variations in the strength of this effect that likely reflect the two different zygotic genotypes in each cross . Thus , when levels of Axin exceed ~4–5 fold endogenous during stages 9–11 , this led to embryonic lethality and strong inhibition of Wg-regulated cell fates . To assess effects of Axin levels on a Wg-regulated target gene , we examined engrailed ( en ) expression , using antibodies to its protein product . En usually accumulates in the two most posterior cell rows in each body segment ( Fig 3D and 3E , S4 Table ) , and maintenance of En expression requires Wg signaling—thus in wg mutants En stripes are narrowed ( [42]; Fig 3D , S4 Table ) . In contrast , in APC2g10 null mutants or after Axin RNAi , En expression expands to additional cell rows ( Fig 3D , 3F and 3G , S4 Table ) . The 3-fold elevation of Axin levels via ZygGAL4 did not affect En expression ( Fig 3D and 3H , S4 Table ) . In contrast , the 9-fold increase of Axin via MatGAL4 led to partial loss of En expression ( Fig 3D and 3I , S4 Table ) , though on average this was not as severe as that seen in wg mutants . Thus , mildly elevating Axin levels during the critical period ( stage 9–11 ) has little effect on embryonic viability , Wg regulated cell fates or target genes , but when Axin levels are elevated ≥ 8-fold , Wg signaling is strongly inhibited , consistent with previous data suggesting that Axin is rate-limiting . The primary role of the Axin/APC2 based destruction complex is to regulate levels of Arm/βcat . We thus measured effects of different Axin levels on Arm accumulation . Arm has two roles: as part of the cadherin-based cell adhesion complex and as a transcriptional co-activator in the Wnt pathway . Thus , all cells have a pool of Arm at the cortex in adherens junctions . In wildtype , Wg is expressed by one row of cells in each segment , and moves to neighboring cells , resulting in a gradient of Wg signaling across the segment . In cells not receiving Wg , the destruction complex binds to newly synthesized Arm , which targets it for destruction ( Fig 4A and 4B ) . Thus , levels of cytoplasmic Arm are low . However , they are not zero; instead Arm that is not immediately destroyed is retained in the cytoplasm by binding to the multiple Arm binding sites on APC2 [19] . Together cytoplasmic retention and destruction mean little or no Arm can translocate to the nucleus and co-activate Wnt target genes ( Fig 4B , arrows ) . In cells receiving Wg , the destruction complex is turned down , and Arm accumulates in both the cytoplasm and nucleus , leading to activation of Wnt target genes [43] . Together , these inputs create a gradient of Arm accumulation across the segment , with the highest level of cytoplasmic/nuclear Arm accumulation in Wg-expressing cells and their immediate neighbors , and gradually decreasing levels of cytoplasmic/nuclear accumulation in cells more distant from the Wg source ( Fig 4A and 4B; diagrammed in 4B’ ) . In wg mutants the destruction complex downregulates Arm in all cells , eliminating the stripes of Arm accumulation [43] . Because different GAL4 drivers changed both the level and the timing of Axin expression , we focused our attention on stage 9 , when Wnt signaling is maximal , to alleviate the complication of differences in timing . We developed methods to quantify the effects of elevating Axin levels on two different aspects of Arm stabilization . To quantify the graded effects of Wg signal across the segments ( Fig 4B’ ) , we used a digital image mask ( S2A’ Fig ) to remove the cortical Arm in cell-cell adherens junctions ( S2A vs . S2A” Fig ) , and then measured fluorescence levels of cytoplasmic/nuclear Arm pixel by pixel across two to three body segments ( S2A” Fig box; two wildtype examples are in Fig 4G left ) . In wildtype embryos , both our images and quantitative analysis revealed a smooth gradation of Arm accumulation , from peaks centered on Wg stripes to troughs in the interstripes ( Fig 4A , 4B and 4G ) . As a control , we examined wg null mutants , in which Arm levels were not elevated in any cells ( Fig 4G center; each mutant was analyzed in parallel with the wildtype shown to its left ) . 9-fold elevation of Axin ( Mat Axin ) led to either complete loss of this graded stabilization of Arm in cells receiving Wg signal , or a reduction in the height of the peaks , relative to wildtype ( Fig 4C and 4D , quantified in G ) . The changes in Arm peak heights were dependent on the level of Axin:GFP expression; this was best visualized in Mat RFP&Axin embryos where the lower level Axin expression only partially flattened the Arm distribution ( Fig 4E and 4F , quantified in 4H ) . To measure absolute levels of Arm stabilization by Wg signaling , we assessed Arm fluorescence in two groups of cells: 1–2 cell rows centered on cells expressing Wg ( the Wg stripes; S2B and S2B’ Fig , yellow boxes ) and 1–2 cell rows farthest from the Wg-expressing cells ( the interstripes; S2B Fig , white boxes ) . Wildtype embryos were included on the same slides as a control . We quantified absolute Arm levels in both Wg stripes and interstripes ( Fig 4B’ , black arrows , Fig 4I , S5 Table ) and also the difference in levels between these two cell types ( Fig 4B’ red arrow , Fig 4J , S6 Table ) . 9-fold overexpression of Axin ( Mat Axin ) substantially reduced Arm accumulation in Wg stripes , to levels similar to those normally seen in interstripes ( Fig 4C and 4D vs . 4A; quantified in 4I and 4J , S5 and S6 Tables ) . However , strikingly , Arm accumulation in interstripes was unaffected . The 4-fold Axin overexpression in Mat RFP&Axin embryos also reduced Wg-stabilization of Arm , but when we sorted embryos by level of Axin:GFP expression , this was less pronounced in embryos with lower levels of Axin:GFP ( Fig 4E and 4F vs . 4A , quantified in 4K and 4L , S5 and S6 Tables ) . To complete this analysis , we examined whether elevating Axin levels affected only the signaling pool of Arm ( cytoplasmic plus nuclear ) or also affected the pool at cell junctions . Using a membrane-mask , we separately assessed these two pools . Elevating Axin levels 9-fold ( Mat Axin ) reduced Arm accumulation in both the junctional and cytoplasmic/nuclear pool in Wg-ON cells , without significantly affecting either pool in Wg-OFF cells , relative to wildtype embryos ( S3A–S3C Fig , S7 Table ) . Together , these data suggest that when Axin levels are elevated ≥4–5 fold , the destruction complex cannot be effectively inactivated by physiological levels of Wg signaling , confirming previous observations that Axin is rate-limiting in this regard . However , it was also striking that elevating Axin levels did not further increase Arm destruction in cells not receiving Wg signal ( Fig 4I , S5 Table , S3A–S3C Fig . S7 Table ) , suggesting that Axin is not rate-limiting for destruction complex activity in those cells . We next investigated whether Wg signaling was similarly affected by altered APC2 levels—since it is the other key component of the destruction complex and our data revealed that its levels are not substantially different from those of Axin , we suspected it might also be rate-limiting and thus over-expression would inhibit Wg signaling . We used a similar approach to elevate GFP:APC2 levels . Using the MatGAL4 driver , we achieved an ~12-fold increase in APC2 levels ( Fig 2D and 2E; S1 Table; hereafter Mat APC2 ) . As we observed with Mat Axin , in Mat APC2 progeny GFP:APC2 levels started high and slowly decreased ( Fig 2G ) . Strikingly , elevating APC2 levels 12-fold had no effect on embryonic viability ( 94% viable; Fig 3A , S2 Table ) ; in fact , these embryos could develop to adulthood and produce viable offspring . We next examined whether elevating APC2 levels affected Wg-regulated cell fate choices , as assessed by cuticle phenotype . Little or no effect on embryonic patterning was seen ( Fig 3B and 3C , S3 Table ) , and the few denticle belt fusions observed were in hatched larvae . Finally , we examined effects on expression of the Wg target gene en . This was also unaffected by overexpression of APC2 ( Fig 3D and 3J , S4 Table ) . Thus , in stark contrast to Axin , embryonic viability , cell fate choice and Wg target gene expression are not sensitive to substantially elevated levels of APC2 . As a final exploration of the effects of elevating APC2 levels , we examined Wg-regulation of Arm stability , using the same assays we employed for analyzing effects of altering Axin levels ( S2 Fig ) . We were surprised to find APC2 overexpression led to a striking change in Arm levels , suggesting reduced activity of the destruction complex . Levels of Arm in Wg-expressing cells and their immediately adjacent neighbors were strongly elevated ( Fig 5B–5D vs . 5A ) , leading to a more defined pattern of stipes in Arm accumulation across each segment . Quantification confirmed that while interstripe Arm levels were unchanged , Arm levels in Wg stripes were significantly higher ( Fig 5E and 5F , S5 and S6 Tables ) . The sharpened stripes and elevated Arm levels in Wg-ON cells were also apparent in our analysis of Arm levels across each segment ( Fig 5G ) . Finally , the same differences were also apparent when we used a membrane mask to examine only cytoplasmic/nuclear Arm or only the membrane pool of Arm ( S3A–S3C Fig ) . These data were quite surprising , as they were the exact opposite of the effects of elevating Axin levels . We examined whether these effects result from reducing Axin levels , a function previously suggested for APC2 [27] , but immunoblotting suggested this was not the case ( Fig 5H , S2 Table ) . Instead , these data suggest that when APC2 levels are elevated in a way that accentuates the endogenous APC2:Axin ratio , stabilization of Arm by Wg signaling is enhanced . This could occur by direct effects on the ability of Wg signaling to downregulate the destruction complex , or via the ability of APC2 to bind and sequester Arm [19]—we consider these possibilities more completely in the Discussion . However , this further elevation of Arm levels in cells already receiving Wg signals had little effect on Wnt-target gene expression or cell fate ( Fig 3A , 3B and 3D , S2–S4 Tables ) . Finally , elevating APC2 levels did not alter destruction complex activity in cells not receiving Wg signals , similar to what we observed with Axin ( Fig 5E , S3A Fig ) . These effects on Arm levels—sharpened and enhanced Arm stripes—were reminiscent of effects previously seen when analyzing APC2 mutants in which the motifs that act as binding sites for Arm ( the 15- and 20-amino acid repeats ) were reduced in number or eliminated ( [44]; S3D Fig ) . These APC2 mutants were expressed at endogenous levels in a null APC2g10 background , rather than overexpressed . We analyzed the most extreme of these—APC2Δ15Δ20R1 , R3-R5 , which deletes all of the ßcat binding sites , and expressed it in the APC null background = APC2g10 APC1Q8 . This allele has a paradoxical phenotype: it strongly reduces APC2 function in Wnt regulation , as assessed by cell fates , but still promotes destruction of Arm in Wg-OFF cells [44] . To determine if the quantitative effects on Arm levels paralleled those we saw after elevating levels of wildtype APC2 , we applied our quantitative toolkit to measure Arm levels on Wg-ON and Wg-OFF cells in that mutant . Intriguingly , interstripe Arm levels were unchanged , while Arm levels in Wg stripes were significantly higher ( S3E and S3F Fig , S5 and S6 Tables ) . This phenotype mimics what we observed after elevating levels of wildtype APC2 . This may suggest that Wg-ON cells are more sensitive to any perturbation that reduces the function of the destruction complex . We consider the interpretation of this similarity further in the Discussion . These data reveal that elevating Axin levels or elevating APC2 levels had opposite effects on the ability of Wg signaling to regulate destruction complex function . To explore this further , we varied the levels of both proteins simultaneously , and also varied the ratios of their expression levels . We began by expressing both Axin:GFP and GFP:APC2 simultaneously ( progeny of GFP:APC2/MatGal4; Axin:GFP/Mat Gal4 females crossed to GFP:APC2; Axin:GFP males; hereafter , Mat APC2&Axin ) . The progeny of this cross differ in their zygotic genotypes and thus in the relative levels of Axin:GFP and GFP:APC2 ( Fig 6A ) . We first examined the average overexpression levels in embryos including all four zygotic genotypes combined . Immunoblotting revealed that , on average , they accumulate Axin:GFP at levels 4-fold above endogenous Axin ( Fig 2H–2J , S1 Table ) similar to Mat RFP&Axin ( which also contains two UAS transgenes ) , and accumulate GFP:APC2 at ~20x endogenous levels ( S1 Table ) . However embryonic lethality of embryos overexpressing both Axin:GFP and GFP:APC2 was substantially higher than that of Mat RFP&Axin embryos ( 63% versus 32% lethal; Fig 6C vs . Fig 3A; S2 Table ) , despite similar average levels of Axin:GFP accumulation ( Fig 2H–2J , S1 Table ) . In parallel , cell fates were shifted more towards the wg null phenotype ( Fig 6D , S3 Table ) than was seen in Mat RFP&Axin embryos ( Fig 3B , S3 Table ) . Therefore , co-expressing APC2 and Axin inhibits Wg signaling to a greater extent than expressing either Axin or APC2 alone , despite similar average levels of Axin:GFP and GFP:APC2 accumulation ( Fig 2H–2J , S1 Table ) . These data are consistent with the hypothesis that co-expressing Axin and APC2 enhances the resistance of the destruction complex to inactivation by Wg signal . We suspected that these averages hid differences in outcome among the four different genotypes present among the progeny ( Fig 6A ) , which would express different ratios of APC2 and Axin . To determine which genotypes exhibited elevated embryonic lethality and defects in Wg-regulated cell fates , we set up two additional crosses , in which the relative zygotic expression of Axin and APC2 differed ( Fig 6B ) : 1 ) APC2>>Axin = average zygotic dose of GFP:APC2 is higher than that of Axin:GFP , and 2 ) Axin>>APC2 = average zygotic dose of GFP:APC2 is lower than that of Axin:GFP . These two crosses had strikingly different results . APC2>>Axin progeny had only 20% embryonic lethality and Wg-regulated cell fates were only mildly affected ( Fig 6C and 6D , S2 and S3 Tables ) , while Axin>>APC2 progeny had 70% embryonic lethality and had very strong effects on Wg-regulated fates , with 39% having a wg null phenotype ( Fig 6C and 6D , S2 and S3 Tables ) . Thus , while APC2 overexpression alone does not affect cell fates , elevating levels of both APC2 and Axin levels inhibits Wg signaling to a greater degree than elevating levels of Axin alone , suggesting both total levels and the relative ratios of Axin and APC2 are important . These data made strong predictions about how different relative levels of Axin and APC2 would affect Arm destruction . While we could not directly determine genotypes of fixed and stained embryos , we developed a method to infer genotypes from levels and localization of GFP-tagged proteins . Since total protein levels of GFP:APC2 were , on average , higher than those of Axin ( Fig 2I , S4A vs . S4B Fig ) , we first separated embryos into two categories , directly quantifying total GFP expression by immunofluorescence and using low versus high GFP levels as a surrogate for zygotically UAS-GFP:APC2/+ versus zygotically UAS-GFP:APC2/UAS-GFP:APC2 embryos ( e . g . , Fig 7A’” and 7B’” vs . 7C’” and 7D’” ) . To further subdivide the embryos , we made use of the assembly of Axin:GFP into cytoplasmic puncta in interstripes [26] . If we could easily visualize cytoplasmic puncta ( Fig 7B’” and 7D’” insets ) , we categorized embryos as zygotically UAS-Axin:GFP/UAS-Axin:GFP rather than zygotically UAS-Axin:GFP/+ . This produced four presumptive genotypes with different degrees of overexpression of Axin and APC2 ( Fig 6A ) : 1 ) APC2+Axin+ . Presumptive zygotic genotype = UAS-GFP:APC2/+; UAS-Axin:GFP/+ , 2 ) APC2+Axin++ . Presumptive zygotic genotype = UAS-GFP:APC2/ +; UAS-Axin:GFP/UAS-Axin:GFP . 3 ) APC2++Axin+ . Presumptive zygotic genotype = UAS-GFP:APC2/UAS-GFP:APC2; UAS-Axin:GFP/+ 4 ) APC2++ Axin++ . Presumptive zygotic genotype = UAS-GFP:APC2/UAS-GFP:APC2; UAS-Axin:GFP/UAS-Axin:GFP . We then analyzed Arm accumulation in these four embryo categories , using the quantitative tools described above to assess absolute Arm levels in Wg stripes and interstripes relative to wildtype controls . To our surprise , despite the four presumptive genotypes , the embryos divided into two phenotypic categories with regard to Arm accumulation . In embryos of the two genotypes that overexpressed Axin at the highest levels ( APC2+Axin++ ( Fig 7B ) ; and APC2++Axin++ ( Fig 7D ) ) , Arm levels were strongly reduced in the Wg stripes ( Fig 7B and 7D: quantified in Fig 7E and 7F; S5 and S6 Tables ) . Thus , they resembled embryos overexpressing only Axin ( Fig 4C , 4D , 4I and 4J ) . In contrast , the two genotypes that overexpressed APC2 but had lower levels of Axin elevation ( APC2+Axin+ ( Fig 7A ) and APC2++Axin+ ( Fig 7C ) ) , Arm levels were strongly elevated in the Wg stripes ( Fig 7A and 7C: quantified in Fig 7E and 7F , S5 and S6 Tables ) . Thus , they resembled embryos overexpressing APC2 alone ( Fig 5B and 5C ) . Combined with the phenotypic data above , these data suggest that the ratio of APC2 to Axin plays a very important role in determining sensitivity of the destruction complex to being inactivated by Wg signaling , such that when Axin is expressed at or over the levels of APC2 , the destruction complex is resistant to inactivation , but when levels of APC2 are substantially higher than those of Axin , the destruction complex is more easily inhibited . One major question still debated in the Wnt field is what happens to the destruction complex after Wnt stimulation . Wnt signaling leads to Axin recruitment to the transmembrane receptor LRP5/6[31] . Work in both cultured human cells and Drosophila embryos suggest that both core components of the destruction complex , APC and Axin , can be recruited to the membrane after Wnt stimulation [13 , 26] . However , three studies of the resulting effects of Wg signaling on Axin levels and localization in the Drosophila embryonic epidermis yielded to three distinct conclusions: 1 ) Wg signaling destabilizes Axin [33] , 2 ) Wg signaling initially stabilizes Axin [28] , or 3 ) Wg signaling leads to Axin membrane recruitment [26] . We thus revisited the question , taking advantage of our ability to express Axin:GFP at defined levels below those at which it significantly inhibits Wg signaling . We first verified that GFP-tagging does not alter physiological roles of Axin: the ability of Axin:GFP to downregulate Arm levels , or its ability to be inactivated in cells that receive Wg . To do so , we expressed Axin:GFP in embryos in which endogenous Axin was knocked down by RNAi ( we co-expressed UAS-RFP to account for effects of different copy numbers of UAS–driven transgenes ) . Axin RNAi led to highly penetrant embryonic lethality ( S5A Fig , S2 Table ) , transformation of cell fates toward Wg-ON fates ( as assessed by cuticle analysis; S5 Fig , S3 Table ) , strong elevation of Arm levels and altered Wg expression ( S5C vs S5D Fig ) . Axin:GFP substantially restored embryonic viability and Wg-regulated cell fate choices ( S5A and S5B Fig , S2 and S3 Tables ) , downregulated Arm and restored normal Wg expression ( S5E Fig ) . Most embryos had a wildtype cuticle , though a small fraction had Wg-signaling inhibited ( S5B Fig , S3 Table ) . Together , these data suggest GFP-tagging does not substantially affect Axin function or its ability to be downregulated by Wg signals . To further verify that the GFP tag on Axin does not alter its function , we analyzed Axin self-assembly in cultured colorectal cancer cells , where Axin self-assembles into multiprotein “puncta” and recruits APC into these structures [45] . We hypothesize these puncta are larger versions of the normal multiprotein destruction complex [17 , 19] . Because GFP can dimerize under some conditions , we verified that similar puncta form and recruit APC2 when Axin is tagged with a Flag-epitope rather than with GFP or one of its derivatives ( S6A–S6F Fig ) . We also created a version of Axin tagged with a monomeric mutant of GFP [46] , and observed no difference in Axin-self-assembly into puncta or recruitment of APC2 ( S6G and S6H Fig ) . Similar puncta were previously observed in Drosophila embryos when Axin:GFP was significantly overexpressed , at levels that inhibit Wg signaling [26] . Membrane-associated endogenous Axin puncta were also seen in imaginal discs , and Axin tagged with the V5 epitope also accumulated in membrane-associated puncta and in the cytoplasm of cells in embryos that received Wg signal [34] . Our system allowed us to directly visualize Axin:GFP localization in embryos expressing it at levels near those of endogenous Axin ( Mat RFP&Axin = 4-fold elevated ) ; ~70% of these embryos are viable and >60% have no disruption of Wg-regulated cell fates ( Fig 3A and 3B , S2 and S3 Tables ) . Visualizing Axin:GFP directly avoided issues with antibody accessibility to Axin assembled into large multiprotein complexes versus protein diffuse in the cytoplasm , an issue we observed in cultured colorectal cancer cells ( [17]; S6D’ Fig insets ) . We examined Axin:GFP localization throughout the stages at which Wg signals regulate cell fate . wg mRNA expression initiates at the blastoderm stage . As germband extension starts , Wg protein is just beginning to accumulate in stripes ( Fig 8A and 8B; [42] ) . At this stage , most cells had small puncta of Axin:GFP , both membrane-proximal and cytoplasmic , along with a cytoplasmic pool . In some cells near to those initiating Wg expression , Axin:GFP containing puncta were beginning to be enriched at the cortex ( Fig 8B arrows ) . In contrast , at stage 9 , when Wg signaling begins to regulate Arm levels and shape cell fate , we observed a prominent difference in Axin:GFP localization in cells receiving or not receiving Wg signal ( Fig 8C and 8D ) . In cells far from the source of Wg , much of the Axin:GFP was assembled into bright cytoplasmic puncta , with relatively low levels in the cytoplasm ( Fig 8D and 8E yellow arrows ) . In contrast , in cells receiving Wg signal , Axin:GFP assembled into less bright membrane-associated puncta , and elevated levels of Axin:GFP were seen in the cytoplasm ( Fig 8D and 8E magenta arrows ) . A similar pattern was observed using GAL4 drivers that led to higher levels of Axin:GFP ( act5c-GAL4 = Mat/Zyg Axin or MatGAL4 without RFP = Mat Axin ) . This resembled the pattern previously observed by Cliffe et al . ( 2003 ) using a strong GAL4 driver [26] . During stage 10 , when the Wg stripes become interrupted , with separate midline and lateral stripes ( Fig 8G , brackets ) , the pattern of Axin:GFP localization became more complex in parallel . Differences in intracellular localization remained between cells near those expressing Wg ( Fig 8G , magenta arrows ) and those farther away ( yellow arrows ) . To quantitatively assess levels of Axin:GFP in different subcellular structures , we thresholded our images to different degrees , assessing which structures were brightest and thus likely contained the highest density of Axin:GFP proteins . The results were quite striking . The brightest 0 . 1% of pixels and most of the brightest 0 . 3% of pixels , which represent the highest levels of Axin:GFP accumulation , were located in the cytoplasmic puncta in Wg-OFF cells ( Fig 8F and 8F’ ) . When we lowered the threshold intensity to visualize the brightest 1% of pixels , the next structures to appear were the membrane-associated puncta in Wg-ON cells ( Fig 8F” ) . It was only when we visualized the brightest 15% of the pixels that the relatively high levels of diffuse cytoplasmic Axin:GFP in the Wg-ON cells were revealed ( Fig 8F”‘ ) . This contrasted with the lower cytoplasmic levels of Axin:GFP in Wg-OFF cells . We next sought to reconcile our observations with recent publications , whose data suggested that the primary effect of Wg signaling was to stabilize Axin in both the cytoplasm and at the membrane [28 , 34] . These studies used an antibody to an epitope to visualize epitope-tagged Axin . We therefore used a GFP-antibody to visualize Axin:GFP expression ( S7A–S7E Fig ) . Intriguingly , the bright Axin cytoplasmic puncta in the interstripe regions were less apparent ( e . g . , Fig 8C’ vs . S7A’ or S7C’ Fig ) —thus use of an antibody emphasized the stronger cytoplasmic signal in Wg-ON cells , reproducing the earlier observations . This suggested that directly visualizing Axin:GFP provides a more complete picture of the effects of Wg signaling on Axin localization and levels . Earlier work suggested that when Axin is significantly over-expressed , Axin puncta also contain APC2 [26 , 47] . We revisited this issue , using our ability to visualize Axin puncta at near endogenous levels ( 4x-elevated; Mat RFP&Axin ) in embryos where Wnt signaling is not substantially inhibited . In wildtype embryos , APC2 is cortically enriched , with a strong cytoplasmic pool ( Fig 8H; [35] ) . Expressing Axin:GFP at 4x endogenous levels significantly altered APC2 localization ( Fig 8I and 8J ) . APC2 was now recruited into both the large cytoplasmic puncta in Wg-OFF cells ( Fig 8I and 8J , yellow arrows ) and to the smaller , membrane-bound puncta in Wg-ON cells ( Fig 8I and 8J , magenta arrows ) . Intriguingly , recruitment of APC2 into Axin puncta seemed more robust in Wg-ON than in Wg-OFF cells ( Fig 8J and 8J’-yellow vs . magenta arrows ) . Together , these data suggest that in the absence of Wg signals , Axin self-assembles into large cytoplasmic multiprotein destruction complexes and diffuse cytoplasmic levels of Axin are reduced . Axin recruits APC2 into these puncta and thus they are likely to represent active destruction complexes . In contrast , in cells receiving Wg signal , Axin:APC2 puncta are recruited to the plasma membrane , these puncta diminish in intensity , and the cytoplasmic pool of Axin is correspondingly increased—these changes occur in parallel with and may cause the reduction in destruction complex activity . These data suggest that Wg signaling leads to destruction complex recruitment to the plasma membrane , as was observed in cultured cells . To confirm that Wg was required for this response , we visualized Axin:GFP localization in embryos zygotically mutant for the genetically null allele wgIG22 ( these mutants produce reduced levels of a non-functional protein , allowing us to identify mutants by reduced Wg accumulation and loss of Arm destruction ) . Consistent with the hypothesis that reception of Wg triggers membrane recruitment of Axin:GFP puncta , Axin:GFP localized to cytoplasmic puncta in all cells in wgIG22 mutants ( Fig 9A vs . 9B ) , while levels of diffuse cytoplasmic Axin:GFP were relatively low in all cells . These data are consistent with what Cliffe et al . ( 2003 ) observed when expressing Axin:GFP at higher levels . We also carried out the converse experiment , using the matGAL4 driver to ubiquitously express UAS-Wg:HA [48] , and examined effects on localization of Axin:GFP . Ubiquitous Wg expression led to highly penetrant embryonic lethality and strong expansion of the Wg-regulated naked cuticle fates ( S8A , S8B and S8E Fig , S2 and S3 Tables ) . Ubiquitous Wg expression led all cells to accumulate Axin:GFP in membrane puncta , with elevated levels of Axin:GFP in the cytoplasm ( Fig 9C ) . These data confirm that the alterations of Axin:GFP localization are driven by reception of Wg signal . The kinase GSK3 , encoded in Drosophila by the zw3 gene , plays multiple roles in Wnt signaling [2 , 31] . In addition to its essential role in regulating Arm/ßcat levels [43 , 49] by phosphorylating its degron [50 , 51] , GSK3 also phosphorylates the LRP5/Arrow co-receptor , creating Axin binding sites [8] . GSK3 also phosphorylates Axin to regulate its stability and association with ßcat [52 , 53] , and phosphorylates APC on distinct sites to increase its affinity for ßcat [54–56] or to promote ßcat release to the E3 ligase [17] . The most upstream of these roles is in membrane recruitment of the destruction complex via receptor phosphorylation—we thus explored whether reducing GSK3 activity would alter this . We knocked down maternal/zygotic zw3 by RNAi , and examined Axin:GFP localization . Strikingly , the membrane recruitment of Axin:GFP observed in Wg-ON cells was lost—instead Axin:GFP formed cytoplasmic puncta in all cells ( Fig 9D–9F ) . As expected , Arm levels were strongly elevated ( Fig 9D ) . Moreover , we also observed notable Arm enrichment in the Axin:GFP puncta ( Fig 9F ) . This is intriguing; it is consistent with the role of GSK3 in phosphorylating Arm/ßcat to create an E3 ligase binding site , and with the proposed role of GSK3 in phosphorylating APC2’s R2/B motifs to stimulate transfer of Arm/ßcat from the destruction complex to the E3 ligase [17] . Finally , we saw no significant changes in Axin levels ( Fig 9G ) , suggesting any effects on Axin stability were not substantial . Our phenotypic data above suggest that simultaneously elevating levels of both Axin and APC2 leads to synergistic inhibition of Wnt signaling . We thus examined how elevating levels of both Axin and APC2 altered destruction complex assembly and localization . GFP:APC2 expressed alone was primarily cortical ( Fig 5B and 5C ) , as observed for endogenous APC2 [35] . In embryos expressing both GFP:APC2 and Axin:GFP at strongly elevated levels ( APC2++Axin++ embryos; Fig 8K and 8L ) , we observed two notable differences from what we observed when each was expressed alone . First , the cytoplasmic puncta in Wg-OFF cells were brighter ( Fig 8K and 8L yellow arrows ) , likely due at least in part to accumulation of two different GFP-tagged proteins into the puncta . Second and more interesting , the region occupied by bright cytoplasmic puncta became much broader , expanding right up to the Wg-expressing cells ( Fig 8L , blue arrows ) , and the region with membrane-associated puncta became narrower , now largely restricted to the single row of Wg-expressing cells ( Fig 8L , magenta arrows ) . Together with the phenotypic data above ( Figs 6 and 7 ) , these data suggest that if Axin levels are limiting relative to those of APC2 ( i . e . , APC2>>Axin ) , the destruction complex is more susceptible to being turned down by Wg signaling . In contrast , if Axin levels are not limiting relative to those of APC2 ( Axin≈APC2 ) , then elevating APC2 levels makes the destruction complex less susceptible to being turned down by Wg signaling . This state correlates with accumulation in large cytoplasmic puncta , consistent with the idea that this occurs by stabilizing destruction complex assembly to the effects of Wg signaling . Data from both cultured cells and Drosophila suggest the ability of Axin and APC to polymerize into a large multimeric complex is critical for targeting βcat for destruction . Polymerization is driven by DIX-domain-mediated head-to-tail Axin polymerization ( previously visualized by crystallography and SEM [16] ) and by APC’s ability to oligomerize via its N-terminal region and Arm repeats [18] . Overexpressing Drosophila Axin in colorectal cancer cells leads to assembly into large “puncta” , which we hypothesize are enlarged versions of the normal destruction complex . APC2 is recruited into these . We used super resolution microscopy to begin to look inside these puncta , revealing that APC2 and Axin form intertwining filaments [17] . To fully understand destruction complex assembly and function , one key parameter is to estimate the number of proteins assembled into active destruction complexes . This has not been possible , either with respect to the large puncta observed after overexpression in colorectal cancer cells , or the presumably smaller complexes produced when Axin and APC2 are expressed at endogenous levels . To estimate the number of APC2 or Axin molecules within an active destruction complex , we adapted a fluorescence comparison technique developed to quantify numbers of GFP-tagged proteins in multimeric complexes [57 , 58] . This technique utilized macromolecular structures containing a known number of GFP molecules as standards ( e . g . , purified eGFP = 2 molecules and a virus-like particle = 120 molecules ) , and from these developed methods to define the number of proteins in yeast multiprotein complexes where molecule number had not been previously defined . We used 2 yeast strains from this study as standards ( Fig 10A ) : one expressing Ndc80:GFP ( calculated to have 306 molecules ) and the other expressing Mif2:GFP ( calculated to have 58 molecules ) [57] . Since we thought it likely that destruction complexes did not have a fixed size , our goal was to get an order of magnitude estimate of the number of proteins in each destruction complex punctum . We first examined GFP-tagged Drosophila Axin over-expressed in SW480 cells . Axin uses its DIX domain to polymerize , forming cytoplasmic puncta in a large range of sizes and brightnesses [17] . We compared living yeast and Axin-expressing SW480 cells in parallel ( Fig 10A ) , using identical imaging conditions ( see Methods for details ) . Puncta size in these cells varies over several orders of magnitude [17] , and thus the brightest puncta in each cell exceeded the linear range of our yeast standards and could not be analyzed . We determined brightness of individual puncta and used the two yeast standards to estimate relative brightness and thus relative molecule number . This allowed us to obtain order-of magnitude estimates of the number of Axin molecules per punctum . In the set we analyzed , the number of Axin:GFP molecules per punctum ranged from 163–1327 ( mean ~700; Fig 10C; S8 Table ) . When APC2 is expressed along with Axin in SW480 cells , it is recruited into the Axin puncta [19] . We thus also examined SW480 cells coexpressing both to get order of magnitude comparisons of the number of Axin or APC2 molecules in puncta . In cells co-transfected for Axin:GFP and RFP:APC2 , the number of Axin:GFP molecules ranged from 104–2041 ( Fig 10C; S8 Table ) , while in cells transfected with a GFP:APC2 and Axin:RFP , the number of GFP:APC2 molecules per punctum ranged from 162–3297 ( Fig 10C; S8 Table ) , suggesting puncta contain roughly comparable numbers of both proteins . Because the brightest puncta were outside the dynamic range of our camera , and thus were not quantifiable using our yeast standards , these data provide a lower bound for molecule number in the largest puncta . These data suggest that when over-expressed in SW480 cells , APC2 and Axin can assemble into destruction complexes containing at least 100s to 1000s of each protein , and within the complex are likely to be present at the same order of magnitude in molecule number . While this offered insights into the assembly ability of Axin and APC2 , it involved very significant overexpression in an APC mutant colorectal cancer cell line . To assess molecule numbers in an active destruction complex in a natural context and at more normal expression levels , we turned to live Drosophila embryos from the Mat RFP&Axin line . They express Axin:GFP at 4x-endogenous levels and >60% of these embryos are viable with no or subtle defects in Wg-regulated cell fates ( Figs 2A , 2B , 3A and 3B , S1–S3 Tables ) . We imaged Mat RFP&Axin embryos live , in parallel with yeast expressing each of our two protein number standards ( Fig 10A ) . In embryos even the brightest puncta were within the dynamic range of the camera , and thus could be accurately compared to our yeast standards . Fluorescence comparison revealed that the Axin:GFP puncta range from 46–931 Axin molecules per punctum ( at stage 9; average ~200; Fig 10D , S8 Table ) . As noted above , subcellular localization and apparent brightness of Axin:GFP puncta changed in response to Wg signaling , with the brightest puncta in the cytoplasm of Wg-OFF cells and dimmer , membrane-bound puncta in Wg-ON cells . This difference across the segment was apparent in our live Mat RFP&Axin flies ( Fig 10B ) . We used these criteria to separate the puncta into those in Wg-ON versus Wg-OFF cells . There was a significant difference between the numbers of Axin:GFP molecules in puncta found in Wg-OFF ( average ~260 molecules ) versus Wg-ON regions ( average ~130 molecules; Fig 10E; S8 Table ) , although the distributions overlapped . These data provide the first insight into the scale of macromolecular assembly in an endogenous destruction complex , suggesting each contains 10s to 100s of Axin molecules . They also support the idea that the number of Axin molecules per destruction complex decreases in response to Wg signaling . Dsh is a key positive effector of Wnt signaling , acting downstream of the receptors to downregulate the destruction complex . Dsh can co-polymerize with Axin , competing with Axin self-polymerization [10] . Data in vivo suggest Dsh and APC can compete for Axin interaction [47] . This suggested the possibility that these two forms of competition might be part of the mechanism by which Dsh downregulates destruction complex activity . One key factor in evaluating this possibility are the relative levels of the three proteins . Our analysis above revealed that Axin and APC2 accumulate at levels within a few-fold of one another . We adopted a similar strategy to assess the relative levels of Dsh . We obtained a set of Dsh:GFP transgenes driven by the endogenous promotor [59] , and used immunoblotting with an anti-Dsh antibody [60] to explore their levels of expression ( Fig 11A ) . We chose the line that accumulated Dsh:GFP at levels closest to endogenous Dsh ( line Dsh:GFP 2 ) and compared accumulation of Dsh:GFP to that of GFP:APC2 and Zyg Axin:GFP ( Fig 11B , S1 Table ) . These data revealed that Dsh accumulates at levels 2 . 4±0 . 5 times that of Axin , suggesting that all three proteins are within a few-fold of one another in abundance and thus competition for oligomerization or binding among them are plausible . We next examined whether endogenous Dsh is recruited into the Axin puncta , either in Wg-ON cells , where it acts to antagonize destruction complex activity , or in Wg-OFF cells where it does not have a known role . We used an antibody to Dsh [60] . In wildtype embryos Dsh was largely cytoplasmic , with weak membrane recruitment in Wg-ON cells ( Fig 11C double arrows ) . However , when we elevated Axin expression ( Mat Axin; 9x overexpressed ) , we saw a striking enhancement of Dsh membrane recruitment in Wg-ON cells . In those cells , Dsh was recruited into membrane-associated puncta that largely co-localized with Axin:GFP ( Fig 11D white arrows ) . In contrast , in Wg-OFF cells Dsh localization remained diffusely cytoplasmic , with little or no enrichment in the cytoplasmic Axin puncta ( Fig 11D , blue arrowheads ) . We saw similar preferential co-localization of Dsh with Axin:GFP in Wg-ON cells when we expressed Axin at near endogenous levels ( 4x-elevated ) such that Wnt signaling is not substantially inhibited ( Fig 11E ) . These data are consistent with co-recruitment of Dsh and the destruction complex to the Wnt receptors in response to Wg signal . Earlier work revealed that substantially overexpressing Dsh can lead to activation of Wg signaling ( e . g . [26 , 61 , 62] ) . Given our new knowledge revealing that relative Dsh and Axin are within a few-fold of one another in abundance , we examined the effect of elevating Dsh levels , using the MatGAL4 driver to express UAS-myc-tagged Dsh ( [63]; here after Mat Dsh ) . This elevated Dsh levels roughly 7-fold ( Fig 11F vs 11G and 11H ) , but had only a modest effect on embryonic viability , reducing it to 83% ( S8C Fig , S1 and S2 Tables ) . Cell fates in most embryos were wildtype , with 29% of the embryos showing mild expansion of naked cuticle ( S8D and S8E Fig , S3 Table ) . At this level of overexpression there was little or no effect on the regulation of Arm levels , as assessed by immunofluorescence ( Fig 11F’ vs 11G’ ) , and no apparent effect on levels of Axin , as assessed by immunoblotting ( Fig 11I , S1 Table ) . These data suggest that at endogenous levels of Axin and APC2 , Dsh is not strongly rate-limiting for Wnt signaling .
Pioneering work in Xenopus egg extracts defined key parameters underlying the biochemical action of the destruction complex , by assembling and measuring destruction complex activity . These studies lacked reagents to directly measure protein levels of all components , and thus used addition of purified Axin to estimate its relative levels . These data suggested Axin is present at levels much lower than the other components of the destruction complex , with an APC:Axin ratio ~5000:1 [22 , 23] . Many mathematical and other models in the field use these data as an underlying premise , and thus they have been influential in thinking about Wnt signaling . However , work in cultured mammalian cells cast doubt on the universality of this ratio—in some cell lines APC levels are much more similar to those of Axin ( <2-fold higher ) while in others Axin was actually present at higher levels than APC [25] . We thus used a well-characterized model where the consequences of Wnt signaling are well known: the Drosophila ectoderm during mid-embryogenesis , when cell fate is tightly regulated by Wg signaling . Using both RNAseq and direct comparisons of protein levels , we found that , in contrast to Xenopus oocyte extracts , APC and Axin levels are quite similar: our protein data suggest the APC2:Axin ratio is 5:1 . In fact , this may overestimate available levels of APC2 , the primary APC family member at this time . APC proteins have distinct cytoskeletal roles [64] at times including those just prior to those we examined [65] , and thus the pool of APC2 available for Wg signaling may be even lower . While it is possible that the difference in results involve the species used ( Xenopus vs . Drosophila ) , our data and the mammalian cultured cell data suggest that the difference may be in comparing tissues where Wnt signaling is active , versus those , like Xenopus egg extracts , in which Wnt signaling is not yet active . Thus , future mathematical modeling of Wnt signaling should include states in which APC and Axin are present at similar levels . Previous work provided conflicting results on Drosophila Axin localization in the absence and presence of Wg signaling . In interstripe cells , which receive little or no Wg signal , the destruction complex is in an active state . Cliffe et al . ( 2003 ) suggested that Axin and APC2 co-localize in cytoplasmic puncta in these cells [26] , while others did not see any notable subcellular localization of Axin in Wg-OFF cells [28 , 33] . To address this , we examined Axin:GFP localization in embryos expressing Axin below the level that substantially alters embryonic viability or cell fates ( Mat RFP&Axin; 4xendogenous ) . Our data confirm and extend the work of Cliffe et al . [26] . In Wg-OFF interstripe cells , Axin assembled into large cytoplasmic puncta , presumably driven by DIX-domain mediated head-to-tail Axin polymerization , as previously visualized by crystallography and SEM [16] . In these cells , levels of cytoplasmic Axin were relatively low , suggesting that much of the Axin self-assembles into puncta . Earlier data and our work reveal that these puncta also contain APC2 ( [26]; Fig 8 ) , and thus APC2’s ability to multimerize may also be relevant [17 , 18] . Our molecular counting experiments also provided the first assessment of the number of Axin molecules in the multiprotein destruction complex . These data suggest active destruction complexes contain tens to low hundreds of Axin proteins , thus helping explain the critical role of the Axin DIX domain [15 , 29] , which mediates Axin polymerization [16 , 66] . Our recent work to engineer the minimal Wnt regulatory machine confirmed that both Axin’s DIX domain and APC2’s Arm repeats , implicated in polymerization , are among the domains most critical for destruction complex function [67] . One key and controversial question in the field involves the mechanism ( s ) by which Wg signaling turns down the destruction complex . Different studies in cultured mammalian cells and Drosophila ( see Introduction ) led to quite different conclusions , ranging from total destruction complex disassembly to inactivation of an intact complex to Axin stabilization . Our new tools allowed us to examine Axin localization directly using a GFP-tagged protein expressed at near endogenous levels , in a tissue where we can examine cells before the onset of Wg signaling , as well as in side-by-side cells experiencing high or low levels of signaling . Our data suggest that in this tissue , Wg signaling leads to membrane recruitment of the destruction complex and are consistent with the idea that it destabilizes assembly , thus increasing the cytoplasmic Axin pool . Before Wg signaling is initiated , Axin:GFP was in cytoplasmic puncta in all cells . However once Wg signaling initiated , Axin:GFP localization differed between cells . In cells not receiving Wg , much of the Axin assembled into large cytoplasmic puncta , leaving relatively low levels diffuse in the cytoplasm . However , in cells receiving Wg signal , the Axin puncta were recruited to the membrane . Our molecular counting and image thresholding experiments suggest these dimmer membrane-proximal puncta contain fewer Axin molecules . Our image thresholding experiments further suggest that in Wg-receiving cells , diffuse cytoplasmic levels of Axin are elevated in Wg-ON relative to Wg-OFF cells . These data support and extend the earlier work of Cliffe et al . ( 2003 ) [26] , who expressed GFP-tagged Axin at more elevated levels . Our observation of elevated cytoplasmic levels of Axin in Wg-ON cells is also consistent with earlier work [28] . However , we did not see clear evidence that this results from Axin protein stabilization , as none of our manipulations of Wnt signaling components significantly altered total Axin levels . Our data further suggest that previous use of antibody staining of an epitope-tagged protein rather than direct visualization emphasized the diffuse cytoplasmic pools of Axin in Wg-ON cells while simultaneously de-emphasizing the larger cytoplasmic puncta in Wg-OFF cells , due to differential antibody accessibility . Thus , the stabilization of Axin proposed previously [25] may largely involve a change in protein localization , rather than a change in total Axin levels . This is consistent with the immunoblotting experiments of Cliffe et al ( 2003 ) , who did not detect altered Axin:GFP levels upon ubiquitous expression of Wg . While Yang et al . 2016 observed an increase in Axin levels by immunoblotting after the onset of Wg signaling in wildtype embryos [28] , this simply reflects activation of the zygotic genome . Our data and earlier data also help define how different components of the Wnt pathway regulate destruction complex localization and assembly . APC2 and Axin co-assemble into large cytoplasmic puncta in Wg-OFF cells . In response to Wg signaling these puncta are recruited to the membrane and reduced in Axin protein number—this recruitment does not occur in wg mutants whereas Wg overexpression triggers membrane recruitment in all cells ( [26]; Fig 9 ) . Mendoza-Topaz et al . ( 2011 ) found that APC2 is critical for assembly of Axin puncta in both Wg-ON and Wg-OFF cells [47] . GSK3/Zw3 is important for membrane-recruitment of Axin puncta , labeling Arm to be recognized by the E3 ligase , and for release of Arm from these puncta ( Fig 9 ) . Dsh is specifically recruited into Axin puncta at the membrane of Wg-ON cells ( Fig 11 ) , and its overexpression accentuates membrane recruitment of both APC2 and Axin [26] . Integrating our data with this earlier work , we hypothesize that in the presence of Wg signaling , Axin puncta are recruited to the membrane , presumably by binding to the activated Wg-receptor . We further hypothesize that Wg signaling , acting at least in part via Dsh , either destabilizes puncta or inhibits puncta assembly , increasing the relative amount of Axin in the cytoplasmic pool . Previous work revealed that sufficiently elevating Axin levels could inactivate Wnt signaling either in cultured mammalian cells [68] or in Drosophila embryos [26 , 69] , demonstrating that Axin is rate-limiting . More recent work revealed that this only occurred when Axin levels were elevated over a certain level [27 , 29] . Our knowledge of absolute levels of APC2 and Axin allowed us to vary levels of each individually or together , thus varying both levels and ratios of the two proteins in the Drosophila embryo where effects of Wg signaling are well characterized . By assessing the effect on embryonic viability , expression of the target gene en , and cell fates choices , we defined the effects of different Axin levels and different timing of Axin accumulation . When Axin:GFP was expressed at ≤4x endogenous Axin , we observed little or no effect on any of these parameters , while at >8x endogenous Axin there was a dramatic increase in embryonic lethality , reduced En expression , and a shift towards a more wg-null like phenotype . These data confirm that Axin can be rate-limiting in vivo . Our data also provided insight into the underlying mechanism: increasing Axin ≥4x rendered the destruction complex less sensitive to inactivation by Wg signaling , and thus decreased Arm levels specifically in Wg-ON cells . Further mechanistic insights remain to be determined , but a key parameter may be the levels of “active Dsh” protein , which is activated by Wg signaling and then can heteropolymerize with Axin and compete with APC [16 , 26 , 47] . Our data reveal that Dsh levels are in the same order of magnitude as those of Axin , making competition plausible . If down-regulation involves a competition between Axin homo-multimerization and Dsh hetero-multimerization , sufficient elevation of Axin levels may saturate the available Dsh molecules and therefore inhibit its ability to inactivate the destruction complex , thus rendering a subset of destruction complexes immune to downregulation . In contrast , our data suggest Axin is not rate-limiting in Wg-OFF cells—Arm levels there were not further decreased by elevating Axin levels . We next asked whether APC2 , the second core component of the destruction complex , is also rate-limiting for destruction complex activity . Expressing GFP:APC2 at >10x endogenous levels had little to no effect on embryonic lethality , En expression , or Wg-regulated cell fate choices . This might be because Axin is rate-limiting—thus additional APC2 would not trigger assembly of additional destruction complexes once it exceeded the available pool of Axin . Intriguingly , however , we observed an unexpected effect of elevated levels of APC2 . In wildtype the normal gradient of Wg creates a gradient of Arm accumulation . In contrast , in embryos with high APC2 expression there is an essentially binary change in Arm accumulation in response to Wg signaling . Wg–expressing cells and their immediately adjacent neighbors accumulate Arm at levels ~1 . 5x higher than the same cells in wildtype . However , in cells more distant from the Wg-expressing cells , Arm levels are unchanged from wildtype . These data suggest a potential positive role for APC2 in turning the destruction complex down in the presence of Wg signaling . While this paper was under review , another paper was published , suggesting a different role of APC in inhibiting the destruction complex in response to Wg signal , by modulating Axin phosphorylation by GSK3 [70] . These paradoxically opposite effects of elevating levels of Axin or APC2 on the ability of Wg signaling to inactivate the destruction complex were surprising . Our dual over-expression of APC2 and Axin provided potential insight into the underlying mechanisms , emphasizing the importance of the relative ratios of different destruction complex proteins . In embryos with elevated expression of both APC2 and Axin , the destruction complex was even less effectively turned down by Wg signals than after overexpression of Axin alone , and large cytoplasmic Axin/APC2 puncta were found in cells immediately adjacent to those expressing Wg , rather than being confined to cells with lower levels of Wg signaling . As noted above , this may occur in situations where Axin levels exceed those of active Dsh . These data are consistent with the known role of APC proteins in promoting Arm/ßcat destruction , and fit well with our work in cultured mammalian cells , which demonstrated that APC can stabilize Axin multimerization , thus increasing destruction complex size and its effective activity [17] . However , if Axin was limiting , effects of APC2 elevation were quite different . Now elevated levels of APC2 allowed Wg signaling to more effectively turn down the destruction complex , thus elevating Arm levels . It is possible that when APC2 levels exceed those of Axin , it forms incomplete subcomplexes with other destruction complex proteins , thus titrating their levels . Since Axin directly interacts with GSK3 and CK1 while APC2 does not , we do not think the effects of elevating APC2 occurred solely by sequestering those proteins in partially assembled and inactive complexes . One potential explanation is that if APC2 levels exceed those of Axin , it binds and sequesters Arm in binary APC2:Arm complexes , protecting Arm from destruction by the remaining functional destruction complexes . Previous work revealed that the ability of APC2 to retain Arm in the cytoplasm fine-tunes Wg signaling independent of its role in destruction [19 , 44] . Alternately , these data are consistent with a model in which APC2 has dual positive and negative roles in Wnt regulation , as was previously suggested to occur in the eye and wing imaginal discs [71] . However , the mechanism by which this occurs remains unclear . Our immunolocalization and immunoblotting data do not support the hypothesis that APC2 overexpression promotes Axin turnover , as was previously suggested based on loss-of-function analysis [27 , 71] . Perhaps the membrane-associated pool of APC2 brings both Axin and the destruction complex in proximity to the Wg receptors and Dsh , allowing more rapid and effective turndown of destruction complex function and assembly . Our data are also consistent with the hypothesis that relative levels of Axin and active Dsh are key , as we found Axin and Dsh accumulate at quite similar total levels . It would be interesting to examine the effect of varying Dsh levels on destruction complex assembly , localization and function . If Dsh levels are a rate-limiting factor in turning off the destruction complex , then increasing Dsh may balance the effects of elevating Axin . We only saw modest effects of elevating Dsh levels 7-fold , suggesting that elevating Dsh levels alone may not be sufficient , if the ability of the Wg receptor to “activate” the pool of Dsh is limiting . Our data on Dsh localization suggest that its ability to co-localize with Axin requires “activation” by Wg signaling . Moving forward , it would be interesting to count the number of Dsh molecules in the membrane-associated Axin puncta to determine how Dsh molecule numbers within puncta relate to those of Axin or APC2 . We also need to explore the ratio of APC2:Axin within the destruction complex in vivo , to parallel the work after overexpression in cultured mammalian cells reported here . Wg signaling may alter these ratios and thus regulate destruction complex function . Intriguingly , after Axin over-expression , recruitment of APC2 into Axin puncta seemed more robust in Wg-ON than in Wg-OFF cells ( Fig 8J and 8J’-yellow vs . magenta arrows ) . One striking aspect of our manipulations was that the effects of elevating levels of APC2 or Axin appeared largely confined to the cells receiving Wg signal . These data are consistent with the idea that normal levels of Axin and APC2 are not rate-limiting in Wg-OFF cells , and thus the alterations in APC2:Axin ratio we created do not alter the efficiency of Arm turnover there . In contrast , the cells receiving Wg signal seem much more sensitive to any manipulation that reduced destruction complex function . These include the elevated APC2:Axin ratio created by elevating APC2 levels , or the reduced ability of APC2Δ15ΔR1 , ΔR3-5 to bind and retain Arm in the cytoplasm [44] . One speculative explanation for this is that in those cells , the system is finely balanced , due to competitive interactions between Dsh and Axin , Axin and Axin , and Axin and APC2 . In such a scenario , relatively small changes in the levels of any one of these proteins could change the balance of active versus inactive destruction complexes . Our data suggest that the relative levels of different Wnt signaling regulators are a critical determining factor , and that modulating relative levels of different components may make cells more or less sensitive to Wnt signaling . Integrating our data with data from many labs using Drosophila and cultured mammalian cells , we propose the following model . During embryogenesis , cells begin with relatively similar levels of the two core scaffolds of the destruction complex ( 4-5x more APC2 than Axin ) , and levels of total Dsh protein are in the same range . When Wg signaling is off , Axin self-assembles into cytoplasmic complexes of tens to hundreds of molecules , which we believe represent the functional destruction complex . In this state , much of the Axin in the cell is assembled into puncta , with less free in the cytoplasm . In these cells , we propose that there are 2 pools of APC2 , a pool localized to the cortex that mediates APC2’s cytoskeletal functions , and another that is associated with and stabilizes the assembly of the Axin puncta . This would represent a high activity state of the destruction complex , and it would rapidly bind , sequester and turnover all newly synthesized Arm that is not assembled into adherens junctions . In these cells , which lack activation of the Wg receptors , Dsh is not competent to integrate into Axin complexes at levels sufficient to antagonize destruction complex function . In the presence of Wg signaling , LRP5/6 is recruited to the Wg receptor Frizzled , and phosphorylation of LRP5/6 by GSK3 and other kinases recruits Axin and Dsh to the membrane , in the process activating Dsh so it can be incorporated into Axin complexes . We hypothesize that Axin membrane-recruitment involves largely intact destruction complexes . Our observations are consistent with recent data suggesting that the destruction complex is not fully disassembled in response to Wg signaling nor is Arm phosphorylation by the destruction complex completely inhibited [11–13 , 17] . Instead the ability of the destruction complex to target Arm for destruction is reduced , perhaps in part by blocking the ability to transfer Arm to the E3 ligase . Dsh contains a DIX domain , like Axin , which allows Dsh to hetero-dimerize with Axin . We hypothesize that this Dsh:Axin interaction aids in puncta re-localization and stimulates the decrease in destruction complex size and function . Dsh may actively reduce the size of destruction complexes by competition with Axin:Axin multimerization , or it may compete with APC2 for access to Axin . Other longer-term effects may then reinforce this initial event , including ubiquitination and destruction of Axin or inhibition of GSK3 kinase activity .
All crosses were performed at 25°C . Wildtype was either y w or act5c-Gal4/CyO . The following stocks were obtained from the Bloomington Stock Center: act5c-GAL4 ( 4414 ) , Maternal alpha tubulin GAL4 ( referred to as MatGAL4; a stock carrying both of the GAL4 lines in 7062 and 7063 ) , UAS-Axin:GFP ( 7225 ) , UAS-Dsh:Myc ( 9453 ) , UAS-RFP ( 30556 ) , wgIG22 ( 5351 ) , UAS-Axin-RNAi ( 62434 ) , UAS-zw3-RNAi ( 35364 ) , and UAS-Wg:HA ( 5918 ) . We also used UAS-GFP:APC2 [19] and an APC2 transgene which expresses APC2 under its endogenous promoter [19] . Dsh:GFP 2 , a Dsh:GFP transgene expressed under its endogenous promoter is from [59] . Dsh:GFP 2 . 33 and Dsh:GFP 2 . 35 , which are derivatives of Dsh:GFP after local gene hopping , were both kind gifts from J . Axelrod ( Stanford ) . Cross Abbreviations ( Female x Male ) : GFP:APC2 = APC2 promoter-GFP:APC2; APC2g10 x APC2 promoter-GFP:APC2; APC2g10 Zyg Axin = UAS-Axin:GFP x act5c-GAL4/+ Mat RFP&Axin = UAS-RFP/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-RFP; UAS-Axin:GFP Mat Axin = +/MatGAL4; UAS-Axin:GFP /MatGAL4 x UAS-Axin:GFP Mat/Zyg Axin = act5c-GAL4/+ x UAS-Axin:GFP . Mat APC2 = UAS-GFP:APC2/MatGAL4; +/MatGAL4 x UAS-GFP:APC2 Mat APC2& Axin = UAS-GFP:APC2/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-GFP:APC2; UAS-Axin:GFP APC2 >> Axin = UAS-GFP:APC2/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-GFP:APC2 Axin >> APC2 = UAS-GFP:APC2/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-Axin:GFP Axin RNAi = UAS-Axin-RNAi/MatGAL4; +/MatGAL4 x UAS-Axin-RNAi/+ Mat Axin-RNAi x Axin:GFP = UAS-Axin-RNAi/MatGAL4; +/MatGAL4 x UAS-Axin:GFP Mat Axin-RNAi x RFP = UAS-Axin-RNAi/MatGAL4; +/MatGAL4 x UAS-RFP Mat RFP x Axin:GFP = UAS-RFP/MatGAL4; +/MatGAL4 x UAS-Axin:GFP wgIG22 mutant = wgIG22/MatGAL4; +/MatGAL4 x wgIG22/+; UAS-Axin:GFP/+ zw3 RNAi = UAS-zw3-RNAi/MatGAL4; +/MatGAL4 x UAS-zw3-RNAi zw3 RNAi x Axin = UAS-zw3-RNAi/MatGAL4; +/MatGAL4 x UAS-Axin:GFP Mat x Wg = UAS-Wg:HA/MatGAL4; +/MatGAL4 x UAS-Wg:HA/UAS-Axin:GFP Mat Dsh = UAS-Dsh:Myc/MatGAL4; +/MatGAL4 x UAS-Dsh:Myc Embryonic lethality assays and cuticle preparations were as in [72] . Inhibition of Wg signaling was assessed by analyzing embryonic and first instar larvae cuticles with the scoring criteria found in Fig 3C and S8 Fig . Embryos were prepared as in [73] . Briefly , flies were allowed to lay eggs on apple juice/agar plates with yeast paste for up to 7 hours . Embryos were collected in 0 . 1% Triton-X in water using a paintbrush , then dechorionated for 5 minutes in 50% bleach . Embryos were fixed for 20 minutes in 1:1 heptane to 9% formaldehyde , with 8mM EGTA added to preserve GFP expression . Embryos were then devitillenized by vortexing in 1:1 heptane to methanol . Embryos were then washed in methanol followed by 0 . 1% Triton-X in PBS , then incubated in blocking buffer ( 1:1000 normal goat serum diluted in 0 . 1% Triton-X in PBS ) for 30 minutes . Embryos were incubated in primary overnight at 4°C , washed in 0 . 1% Triton-X in PBS , then incubated in secondary antibody for 1 hr at room temperature . Embryos were mounted in Aqua polymount ( Polyscience ) . Primary antibodies were: Wingless ( Wg , Developmental Studies Hybridoma Bank ( DSHB ) :4D4 , 1:1000 ) , Arm ( DSHB:N27 A1 , 1:75 ) , phospho-tyrosine ( pTyr , Millipore:4G10 , 1:1000 ) , En ( DSHB:4D9 , 1:50 ) , GFP ( Abcam:ab13970 , 1:10 , 000 ) , Neurotactin ( Nrt , DHSB:BP 106 , 1:100 ) , APC2 [35] , 1:1000 ) , and Dsh ( [60]; 1:4000 ) . Stage 9 embryos were stained with antibody to Engrailed and imaged on a Zeiss LSM 710 or 880 scanning confocal microscope . Images were processed using FIJI ( Fiji Is Just ImageJ ) as follows: maximum intensity projections 8μm thick were created and thresholded to highlight cells expressing Engrailed and eliminate background noise . Three lines parallel to the midline were drawn to intersect with bands 2 through 5 of Engrailed expressing cells relative to the head , two on either side of the embryo and one just to the left of the midline . The cells in each Engrailed band which were intersected by each line were included in our measurements . The number of cells per Engrailed stripe was then determined by averaging these three values . Embryos were scored blind . Significance was assessed using a one-way ANOVA test . Wg-Stripe and Interstripe Arm level values were generally normally distributed , as tested by the D’Agostino-Pearson omnibus normality test as well as the Shapiro-Wilk normality test , and thus parametric tests were employed in statistical analysis . The Paired t-test was used to determine the significance between intragroup values , and an unpaired t-test was used to determine the significance between intergroup values . For multiple comparisons , ordinary one-way ANOVA followed by Dunnett's multiple comparisons test were applied . 4-8hr old embryos were collected in 0 . 1% Trition-X100 , dechorionated in 50% bleach , and then homogenized with a pestle in RIPA buffer ( 1% NP-40 , 0 . 5% Na deoxycholate , 0 . 1% SDS , 50mM Tris pH 8 , 300 mM NaCl; 1x Halt Protease and Phosphatase Inhibitor ( Thermo Scientific ) ) . Protein concentrations were calculated using Protein Assay Dye ( BioRad ) following the manufacturer’s recommendations . Samples were mixed with SDS-PAGE sample buffer , boiled for 5 minutes and then run on an 8% SDS-PAGE gel and transferred to a nitrocellulose membrane . Westerns were visualized using a CLX Licor machine which allowed blots to be imaged over a 4-log range . Band densitometry was calculated using LICOR Image Studio and significance was assessed using a one-sample t test using GraphPad . When band densitometry values differed by more than 5-fold , serial dilutions of samples were used to verify values . Values acquired from these dilutions were reported . Primary Antibodies: anti-GFP ( JL-8 Clontech , mouse monoclonal , 1:1000 ) , anti-Axin ( a kind gift from Y . Ahmed , guinea pig polyclonal , 1:1000 ) , anti-γ-tubulin ( Sigma-Aldrich , mouse monoclonal , 1:2000 ) , anti-APC2 ( Rabbit polyclonal , a kind gift of M . Bienz [75] , 1:1000 ) , and anti-Dsh ( [60] , 1:1000 ) . Secondary Antibodies: IRDye680RD anti-Rabbit ( Licor , 1:10 , 000 ) , IRDye680RD anti-Guinea pig ( Licor , 1:10 , 000 ) , and IRDye800CW anti-Mouse ( Licor 1:10 , 000 ) . mRNA collection and RNAseq analysis are described in [76] ( GEO accession number GSE38727 ) . SW480 cells were cultured at 37° C at normal atmospheric levels of CO2 in L15-media ( Cellgro ) supplemented with 10% FBS and 1x penicillin–streptomycin . Drosophila APC2 or Axin protein constructs were transfected into SW480s using Lipofectamine 2000 ( Invitrogen ) as recommended by the manufacturer . Cells were imaged 24 hours later . Full length Drosophila APC2 and Axin were cloned with either a GFP , RFP , or Flag tag as in [17] . To verify that Axin:GFP polymerization is not simply a result of di- or oligomerization of the GFP protein , we created a monomeric GFP ( mGFP ) by changing Alanine 206 to Leucine [46] . We created the A206K amino acid change in our base plasmid ( pCMV-Axin:GFP; [17] ) using QuikChange ( Agilent Technologies ) following manufacturer’s recommendations . The full plasmid was sequenced to verify the amino acid change in GFP and to ensure no other detrimental mutations were induced in the plasmid . Yeast Fluorescence comparison analysis was performed as described in [57 , 58] . Briefly , yeast were grown at 24°C in YPD media until they reached an OD between 400–600 . Yeast cells were then pelleted and resuspended in YC complete media for live imaging . To adjust for possible background caused by media , both SW480 cells and embryos were also suspended in YC complete medium for live imaging . Images were taken using the same settings on the same day for each experiment . Each slide was imaged for no longer than 20 minutes at room temperature ( ~25°C ) . For analysis , a 15 x 15pixel ROI was created around each punctum , then a 21 x 21 pixel ROI was made around the smaller ROI for background subtraction . Puncta distance from the coverslip and depth of field ( number of Z slices containing a single punctum ) were both taken into account when calculating the molecule numbers ( as in [58] ) . To verify molecular counting , each punctum was compared to 2 different yeast strains: Ndc80:GFP ( ~306 molecules of GFP ) and Mif2:GFP ( ~58 molecules of GFP; both kind gifts from K . Bloom [57] ) . Data sets were only used when molecular numbers were consistent ( +/- 15 molecules ) when calculated with both the Ndc80 and Mif2 standards . Due to the dynamic range of the camera , we were limited in the brightness of puncta that we were able to image in SW480 cells . Therefore , the brightest ( presumed largest ) puncta were omitted from analysis , suggesting that our molecular counts in SW480 cells are an underestimate , as noted in the Results . However , all puncta in embryos were within the camera’s dynamic range . Images were taken on a Zeiss LSM 710 confocal microscope using a Licor LSM-T-PMT camera , with a 488nm diode for stable illumination on a 100x/1 . 4 NA objective lens . Images were analyzed using FIJI . For a more detailed description of how to calculate molecular number see [58] . | Cell-cell communication is critical for cells to choose fates during embryonic development and often goes wrong in diseases like cancer . The Wnt cell signaling pathway provides a superb example . Loss of negative regulatory proteins like APC and Axin takes the brakes off cell proliferation and thus contributes to colon cancer . We study how APC , Axin and their partners keep cell signaling off , and how cell-to-cell Wnt signals reverse this . We use the fruit fly embryo , combining biochemical and genetic tools with advanced microscopy . We found that the destruction complex proteins APC2 , Axin , and their antagonist Dishevelled are present at similar levels , allowing them to effectively compete with one another . We further find that the ability of Wnt signaling to turn off the negative regulatory destruction complex machine is influenced both by the levels of Axin and APC2 and by the ratio of their levels . We visualize the active destruction complex in the animal , and count the number of Axin proteins in this complex . Finally , we find that Wnt signals have two effects on the destruction complex—recruiting it to the plasma membrane and altering its assembly/disassembly . We then propose a new model for how this important signaling pathway is regulated . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"phosphorylation",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"molecular",
"probe",
"techniques",
"antigen-presenting",
"cells",
"immunology",
"immunoblotting",
"vertebrates",
"animals",
"mammals",
"animal",
"models",
"developmental",
"biology",
"signal",
"inhibition",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"amniotes",
"molecular",
"biology",
"techniques",
"embryos",
"drosophila",
"research",
"and",
"analysis",
"methods",
"embryology",
"animal",
"cells",
"proteins",
"molecular",
"biology",
"insects",
"arthropoda",
"biochemistry",
"signal",
"transduction",
"eukaryota",
"cell",
"biology",
"cats",
"post-translational",
"modification",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"wnt",
"signaling",
"cascade",
"cell",
"signaling",
"organisms",
"signaling",
"cascades"
]
| 2018 | Supramolecular assembly of the beta-catenin destruction complex and the effect of Wnt signaling on its localization, molecular size, and activity in vivo |
Polyphenism is a successful strategy adopted by organisms to adapt to environmental changes . Brown planthoppers ( BPH , Nilaparvata lugens ) develop two wing phenotypes , including long-winged ( LW ) and short-winged ( SW ) morphs . Though insulin receptor ( InR ) and juvenile hormone ( JH ) have been known to regulate wing polyphenism in BPH , the interaction between these regulators remains largely elusive . Here , we discovered that a conserved microRNA , miR-34 , modulates a positive autoregulatory feedback loop of JH and insulin/IGF signaling ( IIS ) pathway to control wing polyphenism in BPH . Nlu-miR-34 is abundant in SW BPHs and suppresses NlInR1 by targeting at two binding sites in the 3’UTR of NlInR1 . Overexpressing miR-34 in LW BPHs by injecting agomir-34 induces the development towards SW BPHs , whereas knocking down miR-34 in SW BPHs by injecting antagomir-34 induces more LW BPHs when another NlInR1 suppressor , NlInR2 , is also suppressed simultaneously . A cis-response element of Broad Complex ( Br-C ) is found in the promoter region of Nlu-miR-34 , suggesting that 20-hydroxyecdysone ( 20E ) might be involved in wing polyphenism regulation . Topic application of 20E downregulates miR-34 expression but does not change wing morphs . On the other hand , JH application upregulates miR-34 expression and induces more SW BPHs . Moreover , knocking down genes in IIS pathway changes JH titers and miR-34 abundance . In all , we showed that miRNA mediates the cross talk between JH , 20E and IIS pathway by forming a positive feedback loop , uncovering a comprehensive regulation mechanism which integrates almost all known regulators controlling wing polyphenism in insects .
The phenomenon of polyphenism is that two or more distinct phenotypes are displayed by an organism with the same genotype . This phenomenon is triggered by environmental cues such as population density , host nutrition , and temperature [1] . For example , locusts show density-dependent phenotypic plasticity and have two phases: a low-density “solitarious” phenotype and a high-density “gregarious” phenotype [2] . In beetles , horn polyphenism is a nutrition-dependent trait where some male beetles have fully-developed horns , while other males are completely hornless , depending on their nutrition status and body size [3] . Eusocial insects , including members of Hymenoptera , Blattodea ( termites ) , often display caste differentiation , producing multiple types of offspring with different reproductive and morphological features [4] . Based on the physiological state of the mother , aphids produce winged adults in deteriorating environments and flightless morphs when environmental conditions are stable [5] . Brown planthopper ( BPH , Nilaparvata lugens ) is one of the most notorious planthoppers , migrating from Vietnam and the Philippines to China and Japan in summer during the rice growing season , then flying back to tropical regions during winter after rice crops have been harvested in temperate zones [6] . Wing polyphenism is a key determinant of the success of planthoppers [7] . Long-winged ( LW ) BPHs are capable of long-distance migration , while short-winged ( SW ) morphs display higher reproductive capabilities . Prior works showed that insulin/IGF-1 signaling ( IIS ) pathways involved in regulating wing polyphenism in BPHs [8] . Two insulin receptors ( InR ) , NlInR1 and NlInR2 , have been shown to play opposing roles in determining wing fate . NlInR2 inhibits NlInR1 , switching long-winged morphs to short-winged morphs , and vice versa . These two insulin receptors modulate wing polyphenism by regulating insulin/IGF signaling ( IIS ) pathway [9] . Wounding causes a shift to short wing through stimulation of the forkhead transcription factor ( NlFoxO ) and its downstream target Nl4EBP [10] . High glucose concentration in the rice host induces long-winged females , indicating that host nutrition quality has a direct impact on wing polyphenism in female BPH . This work shows that the nutrition is a key determination factor in controlling wing polyphenism in BPH and also raises a question of sex difference in wing polyphenism [11] . Silencing the c-Jun NH2-terminal kinase ( NlJNK ) increased the proportion of short winged female adults , suggesting that JNK signaling involve in regulating wing polyphenism in BPH [12] . In addition , topic application of juvenile hormone ( JH ) in the BPH nymphs induces SW morphs [13] . Knocking down the juvenile hormone epoxide hydrolase ( Nljheh ) prevents the degradation of JH and induces a bias towards SW BPHs [14] , providing evidence that JH might also involve in regulating wing polyphenism . However , there is a dispute over JH regulation of wing polyphenism in BPH [8 , 15] . Though various regulators have been reported to involve in modulating phenotypic plasticity of wing polyphenism in BPHs , how these regulators interact with each other remains an important unsolved question . miR-34 is an important modulator involved in the cross talk between the IIS pathway and hormone in C . elegans and Drosophila . In C . elegans , miR-34 expression is regulated by IIS via a negative feedback loop between miR-34 and DAF-16 ( the sole ortholog of FoxO in nematodes ) , providing robustness to environmental stress [16] . miR-34 is maternally inherited in Drosophila [17] , and is activated by JH but is suppressed by 20E through Broad Complex ( Br-C ) [18] . Here , we found a microRNA ( miRNA ) , Nlu-miR-34 , modulates the crosstalk between IIS , JH and 20-hydroxyecdysone ( 20E ) pathways to control wing polyphenism . Nlu-miR-34 suppresses NlInR1 , inducing the development towards SW morph . Further , Nlu-miR-34 is activated by JH through Br-C . Moreover , silencing NlInR1 induces the expression of Nlu-miR-34 , suggesting that Nlu-miR-34 , IIS , and JH form an autoregulatory positive feedback loop to control wing polyphenism in BPHs .
Insulin receptors in BPH ( NlInRs ) were reported to determine the wing polyphenism in BPH [9] . To investigate whether miRNA may play a role in regulating wing polyphenism , we carried out a bioinformatics analysis to identify miRNAs targeting at NlInR1 or NlInR2 . The results showed that only Nlu-miR-34 can target NlInR1 with two putative binding sites in the 3’untranslated region ( UTR ) predicted by all five target prediction algorithms ( Fig 1A , see methods ) . miR-34 is highly conserved in insects and nematodes ( Fig 1B , see methods ) and is predicted to have 19 target genes in BPH ( S1 Table ) . To confirm the interaction between Nlu-miR-34 and NlInR1 , we introduced the 3’UTR sequence of NlInR1 at the downstream of the firefly luciferase gene in the pMIR-REPORT vector . Constructs with or without ( negative control ) the 3’UTR of NlInR1 were both transfected into human embryonic kidney 293T ( HEK293T ) cells . The luciferase activity was significantly reduced to 30% relative to the negative control in the presence of agomir-34 ( the mimics of Nlu-miR-34 ) ( Fig 1C and 1D , Student’s t-test , p = 4 . 99e-12 , n = 300 , six replicates ) . Mutating any of the binding sites abolished the suppression effect of agomir-34 on the reporter activity of the NlInR1 target sites ( Fig 1D ) , indicating that both two binding sites are essential for the interaction between Nlu-miR-34 and NlInR1 . Next , we performed RNA-binding protein immunoprecipitation ( RIP ) assay with the antibody-Ago1 ( Argonaute 1 ) in wing buds of BPH nymphs . The transcripts of NlInR1 was significantly enriched in the Ago1-immunoprecipitated RNAs from agomir-34-injected BPHs compared with those from the control samples ( Fig 1E , 9 . 30-fold , Student’s t-test , p = 0 . 002 , n = 150 , two replicates ) , showing that Nlu-miR-34 directly interacted with NlInR1 in the wing buds in vivo . Then , we used fluorescence in situ hybridization ( FISH ) assay to examine whether Nlu-miR-34 co-localize with NlInR1 in wing buds . Indeed , both of Nlu-miR-34 and NlInR1 were detected in wing buds of the fourth instar BPHs ( S1 Fig ) . The critical stages of wing fate determination in BPH is the second to fourth instar of nymph [19] . However , the wing morphs cannot be discriminated at nymph stages . To examine the expression of Nlu-miR-34 in nymphs with different wing fate , we first developed two BPH strains by continually selecting SW or LW adults for crossing for more than 50 generations ( SW♀ x SW♂ for SW strain , and LW♀ x LW♂ for LW strain ) [14] . Almost all adults in SW strain emerge as short-winged adults while 80% of LW strain individuals become long-winged adult morphs . Because these two strains were selected from the same population , SW and LW strain should have the same genetic background . Each generation was purified in the adult stage using a previously described method [20] . We measured the expression of Nlu-miR-34 in the nymph stages of both SW and LW strain . The results show that Nlu-miR-34 was significantly highly expressed in the second and third instars nymphs in SW strain ( Fig 2 , Student’s t-test , p = 0 . 0001 for the second instar , and p = 0 . 002 for the third instar , four replicates ) , suggesting that Nlu-miR-34 might be essential to develop SW adults . However , we found that NlInR1 was also highly expressed in the second and third instar SW nymphs when using the whole body for experiments ( S2 Fig ) . To decipher the roles of Nlu-miR-34 in controlling wing polyphenism , we increased its abundance by injecting agomir-34 into the third instar nymphs at 24 h post molt in LW strain . The abundance of Nlu-miR-34 was significantly increased by 3 . 06-fold compared to the negative control , which was treated with agomir-NC ( a synthesized small RNA with a randomly shuffled sequence of Nlu-miR-34 ) ( Fig 3A , Student’s t-test , p = 0 . 026 , three replicates ) . Fluorescence in situ hybridization ( FISH ) assay showed that the abundance of Nlu-miR-34 increased in the wing bud cells at 24 h post injection ( Fig 3B ) . Next , the agomir-34 and agomir-NC treated nymphs were kept for wing observation until emergence into adults . The results showed that overexpression of Nlu-miR-34 induced the development of SW morphs 51 . 08 ± 6 . 16% in a significantly higher proportion than those in the control group 25 . 78 ± 4 . 78% ( Fig 3C , Student’s t-test , p = 0 . 032 , n = 150 , three replicates ) . In addition , overexpression of Nlu-miR-34 downregulated the expression of its target gene NlInR1 ( Fig 3D , Student’s t-test , p = 0 . 005 , three replicates ) . Having showed increased Nlu-miR-34 in LW strain induces SW morphs , we next inhibited Nlu-miR-34 in SW strain by injecting antagomir-34 ( the inhibitor of Nlu-miR-34 ) . The expression of Nlu-miR-34 was significantly decreased by 56 . 33 ± 16 . 16% ( Fig 4A , Student’s t-test , p = 0 . 004 , three replicates ) . FISH assay showed that the abundance of Nlu-miR-34 decreased in the wing bud cells at 24 h post injection ( Fig 4B ) . However , we did not observe a decrease of SW morphs in antagomir-34 treated groups comparing with the negative control ( Fig 4C , Student’s t-test , p = 0 . 897 , n = 150 , three replicates ) . Since NlInR2 suppresses NlInR1 by forming a heterodimer [9] and NlInR2 is highly expressed in SW strain ( S3 Fig ) , we speculated that NlInR2 might have a more potent suppressing effect and thus impair the effect of antagomir-34 . To test this hypothesis , we first knocked down NlInR2 with 0 . 84 ng dsNlInR2 and then inhibited Nlu-miR-34 with two different amounts of antagomir-34 ( 40 ng and 60 ng ) in SW strain . As expected , after partially knocking down NlInR2 by injecting dsNlInR2 , downregulation of Nlu-miR-34 significantly induced LW morphs ( Fig 5A , Chi-square test , 40ng: χ2 = 5 . 8 , df = 1 , p = 0 . 003; 60ng: χ2 = 7 . 9 , df = 1 , p = 0 . 042; n = 150 , three replicates ) . To test whether this suppressing effect is quantity dependent , we next inhibit NlInR2 by injecting three different amounts of dsNlInR2 ( 0 . 42 ng , 0 . 84 ng and 3 . 38 ng ) . The dsNlInR2-treated individuals were further injected with 40 ng antagomir-34 to inhibit Nlu-miR-34 . The results showed that at a low quantity of dsNlInR2 ( 0 . 42 ng ) , knocking down Nlu-miR-34 with 40 ng antagomir-34 did not lead to more LW morphs ( Fig 5B ) . However , when dsNlInR2 increased to 0 . 84 ng and 3 . 38 ng , the proportions of LW BPHs was significantly increased and SW BPHs were significantly deceased ( Fig 5B , Chi-square test , 0 . 84 ng: χ2 = 8 . 4 , df = 1 , p = 0 . 004; 3 . 38 ng: χ2 = 3 . 3 , df = 1 , p = 0 . 068; n = 150 , three replicates ) . These results suggested that both NlInR2 and Nlu-miR-34 suppress NlInR1 to control wing polyphenism . NlInR2 might have a major suppression effect whereas Nlu-miR-34 might be a “supplemental” suppressor to clear the leaky transcripts of NlInR1 in the wing bud of SW strain . Next , to further uncover the factors regulating Nlu-miR-34 expression , we amplified the 5’UTR of Nlu-miR-34 using rapid amplification of cDNA ends ( RACE ) . The transcription starting site ( TSS ) was determined by mapping the 5’UTR sequence to the BPH genome ( GenBank assembly accession: GCA_000757685 . 1 ) , showing that the TSS of Nlu-miR-34 is located at position 119 , 684 in scaffold0000007224 ( - ) . We searched for transcription binding sites ( TFBS ) nearby the TSS of Nlu-miR-34 ( -2 , 000 to +100 ) using PROMO [21] and Match [22] . Both programs identified a potential cis-response element ( CRE ) of Broad Complex ( Br-C ) , suggesting that NlBr-C might involve in regulating Nlu-miR-34 ( Fig 6A ) . To confirm this , we knocked down NlBr-C by injecting dsNlBr-C at the third instar nymphs in SW strain ( Fig 6B , Student’s t-test , p = 3 . 21e-5 , three replicates ) , resulted in the increase of Nlu-miR-34 by 1 . 44-fold ( Fig 6C , Student’s t-test , p = 0 . 011 , three replicates ) . Because Br-C is an ecdysone inducible transcription factor [23] , we then used 20-hydroxyecdysone ( 20E ) to treat the third instar nymphs in SW strain . Topical application of 20E decreased Nlu-miR-34 significantly ( Fig 6D , Student’s t-test , p = 0 . 0001 , three replicates ) . However , similar as the effect of using antagomir-34 alone without knocking down NlInR2 , 20E application did not change the proportion of wing morphs . Since JH and 20E show antagonistic actions through Br-C [24] and JH was known to regulate wing polyphenism in BPH [13 , 25] , we hypothesize that JH might involve in regulating the expression of Nlu-miR-34 . To test this mode , we used topical application of JH III analogue at the third instar of BPHs in LW strain . Acetone were used as the negative control . The results showed that increased JH titer led to a strong bias towards SW BPHs ( increasing by 23 . 09 ± 6 . 16% ) ( Fig 6E , Student’s t-test , p = 0 . 003 , n = 100 , three replicates ) . After JH application , Nlu-miR-34 was significantly increased by 1 . 91-fold while NlInR1 was decreased by 39 . 7 ± 14 . 26% , but it didn’t change the NlInR2 level ( Fig 6F–6H , Student’s t-test , p = 6 . 04e-5 and 0 . 008 , three replicates ) . Moreover , NlBr-C was also decreased by 25 . 18 ± 12 . 13% at 24h post JH treatment ( Fig 6I , Student’s t-test , p = 0 . 022 , three replicates ) . These results showed that the upregulation of Nlu-miR-34 by JH and the downregulation of Nlu-miR-34 by 20E are likely mediated by NlBr-C in a transcriptional regulation manner . We previously showed that JH activates Nlu-miR-34 . Interestingly , we found JH titer was significantly increased by the overexpression of Nlu-miR-34 ( Fig 7A , n>300 ) , suggesting that JH and Nlu-miR-34 form a positive autoregulatory loop . In addition , NlBr-C was significantly decreased by 39 . 60 ± 3 . 99% when overexpressing Nlu-miR-34 ( Fig 7B , Student’s t-test , p = 6 . 72e-5 , three replicates ) , suggesting the positive feedback loop between JH and Nlu-miR-34 might be mediated by Br-C . To test whether NlInR1 also involve in this autoregulatory loop , we next knocked down NlInR1 by injecting dsNlInR1 in the third instar nymphs in LW strain ( Fig 7C ) . The results showed that knocking down NlInR1 increased Nlu-miR-34 by 2 . 03-fold ( Fig 7E , Student’s t-test , p = 0 . 006 , three replicates ) and led to a strong bias towards SW morphs ( Fig 7D , Student’s t-test , p = 2 . 51e-7 , n = 100 , three replicates ) . Taking together , these results suggested that miR-34 , JH , 20E and IIS pathway form a positive autoregulatory loop to control wing polyphenism in BPHs ( Fig 7F ) .
Many animals with the same genetic background exhibit polyphenism to adapt to different environment conditions [1] . It has been showed that insulin-like peptides ( ILP ) , the IIS pathway , and JH involve in the regulation of polyphenism in various organisms [26–29] . However , it remains elusive how these regulators interact with each other . Here , we showed that Nlu-miR-34 is a mediator between the crosstalk between IIS and hormonal signals by forming a positive feedback loop . miR-34 is a conserved miRNA that plays an important regulatory role in a wide range of organisms . It seems the positive feedback loop mediated by miR-34 might be a conserved process because miR-34 is also regulated by IIS via a negative feedback loop between miR-34 and DAF-16 ( the sole ortholog of FoxO in nematodes ) in C . elegans [16] . We propose a regulation mode of JH -miRNA-IIS feedback loop in BPH ( Fig 7F ) . Through this positive autoregulatory loop , minor changes in environment conditions , above a certain threshold , would be amplified significantly in vivo inducing a phenotype shift . Two pathways involved in miRNA or InR regulation are redundant or at least partially redundant . It seems that InR regulation is more sensitive to nutritive environment whereas miRNA regulation is likely to be hormonally regulated . This discovery extends our understanding of the interplay between nutrition status-based signals and hormone signals in modulating phenotypic plasticity . Polyphenism is a phenomenon in which a single genotype can arise into two or more discrete phenotypes in response to environmental stimuli; these phenotypes do not segregate in F1 generations produced from parents with distinct phenotypes , suggesting that polyphenotypic changes are not determined by only one or two genes [30] . However , knocking down NlInR1 induces formation of 100% SW BPHs [9] , suggesting that the effect of insulin or hormone regulation is normally determinative in most cases . Therefore , other regulators should participate in regulation; and these regulators should be either sensitive to environmental cues or dose dependent . Here , we showed that Nlu-miR-34 mediates the cross talk between JH and IIS , and miRNAs are normally expressed at moderate levels [31] . In contrast to other regulatory molecules , miRNAs are typically very fine-tuned , specifically binding to and regulating their targets , acting as buffers to ensure developmental robustness [32] . These features enable them to act as ideal regulators of phenotype polyphenism , responding to varied environment cues . To the best of our knowledge , we presented the first evidence that miRNAs regulate developmental plasticity by modulating the cross talk between the JH and IIS pathways . It should be noted that the experiments of overexpression of Nlu-miR-34 were repeated independently in three labs in Nanjing , Wuhan and Hangzhou , ensuring the reliability of miRNA modulation of wing polyphenism in BPH ( S4 Fig ) . JH was reported to regulate wing polyphenism in BPH twenty years ago by Tojo and his colleagues but without further evidence [13 , 25] . A recent report showed that knocking down juvenile hormone epoxide hydrolase ( Jheh ) enhanced short wing formation [14] . However , Zhang and his colleagues argued that lack of direct evidence of JH regulation of wing polyphenism in BPH [8 , 15] . Here we present evidence that JH participates in the regulation of wing polyphenism in BPH by upregulating Nlu-miR-34 expression . JH has been reported to regulate miRNA expression during larvae development in D . melanogaster [18] . JH can redirect the development of LW to SW/wingless morphs in BPHs [13 , 25] , aphids [33] , and crickets [34] , but the relationship between JH and wing polyphenism is inverse with that in ants [35–37] . It is interesting to notice that both winged ants and SW BPHs have higher fecundity than other morphs , inferring JH regulation of wing polyphenism varies between different insects and JH might be involved in reproduction regulation [38] , which requires further investigations . Though JH has been well understood in regulating phenotype polyphenism [1 , 28] , the role of 20E is less understood in this field . We showed that miR-34 is activated by JH but is suppressed by 20E in planthopper . A CRE of Br-C was found in the promoter region of Nlu-miR-34 . It has been reported that downregulation of miR-34 requires Br-C in Drosophila [18] . Since we did not find any JH response element ( JHRE ) in the promoter region of Nlu-miR-34 , JH might regulate Nlu-miR-34 in an indirectly manner . JH has been reported to show antagonistic interaction with 20E through Br-C [24] , it is likely that JH may activate the expression of Nlu-miR-34 through a cross talk between JH and 20E mediated by Br-C . However , topic application of 20E did not change the proportion of wing morphs . These results open a new question whether 20E regulates wing polyphenism in BPHs , and if so , in which way the 20E may regulate wing polyphenism . In summary , we showed , for the first time , a comprehensive regulation mechanism of wing polyphenism in BPH by integrating almost all known regulators ( miRNA , IIS , JH , and 20E ) into a positive autoregulatory feedback loop ( Fig 7F ) . We also presented evidence that 20E might also involve in regulating wing polyphenism by cross talking with JH . These discoveries extend our understanding of the mechanism regulating insect polyphenism , and shed new lights on how environmental cues determine which alternative phenotypes are produced by a given genotype .
Wild BPH populations were obtained from rice fields in Wuhan , Hubei Province , China , and raised on a BPH-susceptible rice variety Taichuang Native 1 ( TN1 ) in the greenhouse . SW and LW strains were selected to propagate for more than 50 generations [14] . The percentage of SW morphs in the generated SW strain was more than 80% , while LW morphs represented approximately 80% of the LW strain . Each generation was purified in the adult stage using a previously described method [18] . BPHs were raised in growth chambers at 26°C ( ± 2°C ) under a photoperiod of 16:8h ( light: dark ) with the relative humidity of 75% ( ± 5% ) . We obtained BPH miRNAs sequences that have been previously reported [39 , 40] . The 3’UTR sequences of NlInR1 ( KF974333 ) and NlInR2 ( KF974334 ) were downloaded from NCBI ( http://www . ncbi . nlm . nih . gov/ ) . We used five algorithms , including miRanda [41] , TargetScan [42] , microTar [43] , PITA [44] , RNAhybrid [45] , to predict miRNAs that can target NlInR1 or NlInR2 . Default parameters were used for all algorithms . To investigate whether miRNA may play a role in regulating wing polyphenism , we carried out a bioinformatics analysis to identify miRNAs targeting at NlInR1 or NlInR2 . The results showed that three miRNAs target on NlInR1 and four miRNAs target on NlInR2 ( S2 Table ) . Then , we used luciferase reporter assays to confirm the interaction between miRNA and their targets , and the results confirmed the interactions between NlInRs and three miRNAs , Nlu-miR-34 , Nlu-miR-989b and Nlu-miR-989c . We overexpressed these three miRNAs separately to test their effects on wing polyphenism in BPH , showing that only Nlu-miR-34 can significantly change the ratio of wing morphs ( Fig 3A and S5 Fig ) . So , we focused on Nlu-miR-34 in further studies . miR-34-5p sequences from several insects and nematode C . elegans were downloaded from miRBase and submitted to ESPript 3 . 0 ( http://espript . ibcp . fr/ESPript/cgi-bin/ESPript . cgi ) for multiple sequence alignment and visualization . The 3’UTRs of NlInR1 and NlInR2 were amplified and introduced into the pMIR-REPORT vector ( Obio , China ) downstream of the firefly luciferase gene . The pRL-CMV vector ( Promega , USA ) , which contains the Renilla luciferase gene , was used as a control luciferase reporter vector . The predicted binding site ( 5’-CACTAGTGACCGCGTAGTTGCCTGCCG-3’ ) was completely removed for mutant 1 , and another binding site ( 5’-TGACGTTGCCGCCGCCACTGCCG-3’ ) was removed for generation of mutant 2 . HEK293T cells ( Obio , China ) were cultured at 37°C in 5% CO2 in DMEM ( Gibco , USA ) media supplemented with 10% FBS ( Hyclone ) for 24 h in 96-well culture plates . Cells were transfected with pMIR-REPORT ( 0 . 2 μg ) , pRL-CMV ( 0 . 01μg ) , 0 . 25 μl of 100 nM miRNA mimics ( RiboBio , China ) , and 0 . 25 μl Lipofectamine 2000 reagent ( Invitrogen , USA ) according to the manufacturer’s instructions . The activity of the two luciferase enzymes was measured 48 h after transfection following the manufacturer’s recommendations ( Dual-Luciferase Reporter Assay System , Promega , USA ) with Infinite M1000 ( Tecan , Switzerland ) . Three replicates were performed for each experiment . Results are expressed as the means of the ratio of firefly luciferase activity/Renilla luciferase activity . Controls were set to one , and the two-tailed t-test was used for statistical analysis . A Magna RIP Kit ( Millipore , Germany ) was used to perform the RIP assay according to the manufacturer’s instructions . Approximately 50 wing buds of the fourth instar nymphs were collected and crushed using a homogenizer in ice-cold RIP lysis buffer . Magnetic beads were incubated with 5 μg of RIPAb+ Ago-1 antibody ( Millipore , Germany ) or normal mouse IgG ( Millipore , Germany ) . The homogenates in the RIP lysates were centrifuged and the Ago 1-bound mRNAs in supernatants were pulled down by magnetic bead–antibody complex at 4°C overnight . The immunoprecipitated RNAs were released by digestion with protease K . Finally , the RNAs were purified by which methods and used for cDNA synthesis . The PrimeScript 1st Strand cDNA Synthesis Kit ( Takara , Japan ) was used and the abundance of InR1 was quantified by qPCR . An antisense nucleic acid detection probe ( 5’-ACAACCAGCUAACCACACUGCCA -3’ ) designed to detect Nlu-miR-34 was labeled with Cy3 . The probe for detecting NlInR1 ( 5’- GAACAGCCAGGACAGGCCGAAUCCUCCAUG-3' ) was labeled with FAM . The random shuffled probe ( 5’-UUCUCCGAACGUGUCACGUU-3’ ) and the probe ( 5’-UUGUACUACAAAAGUACUG-3’ ) were used as miRNA and mRNA negative controls , respectively . The wing buds were dissected from the third or fourth instar stage nymphs injected with agomir or antagomir ( RiboBio , China ) . For fluorescence in situ hybridization , wing buds were treated for 24 h , fixed in 4% paraformaldehyde for 2 hours , and then incubated with the miRNA probes at 37°C for 24 h . The samples were washed in PBS containing Triton X-100 ( 5% v/v ) and then incubated with the DAPI ( RiboBio , China ) at room temperature for 30 min . Signals were analyzed and images were recorded using a Zeiss LSM 780 confocal microscope ( Carl Zeiss SAS , Germany ) . Figures were prepared using Zeiss LSM ZEN 2010 software ( Carl Zeiss SAS , Germany ) . Total RNA was isolated from ten fourth to fifth instar nymphs of SW BPHs using the TRIzol regent ( Invitrogen , USA ) . The SMARTer RACE cDNA amplification kit ( Clontech , USA ) was used to obtain the 5’UTR of Nlu-miR-34 . 5’RACE nested-PCR was performed using Ex-Taq ( Takara , Japan ) according to the manufacturer’s instructions . The amplified products were separated by agarose gel electrophoresis and purified using a Gel Extraction Kit ( Takara , Japan ) . The purified DNA was ligated into the pMD19-T vector ( Takara , Japan ) , and sent to BGI-Shenzhen for sequencing . The transcript starting site ( TSS ) of miR-34 was confirmed by alignment to genomic scaffold sequences ( GenBank assembly accession: GCA_000757685 . 1 ) . Any unknowns “N” in the genomic sequences were confirmed by PCR validation . A 2100 bp flanking sequence ( -2000 to +100 ) upstream of the TSS was used as the promoter region for analysis by Promoter Scan web server ( http://www . cbs . dtu . dk/services/Promoter/ ) . A putative TATA-box was identified at position -1707 . To identify putative transcription factor binding sites ( TFBS ) in the promoter region , two algorithms , including PROMO 3 . 0 [18] and MATACH 1 . 0 [19] based on TRANSFAC database ( http://gene-regulation . com/pub/databases . html ) , were computationally analyzed using default parameters . A Br-C Z4 biding site ( TTTTGTTTAAATT ) at position -699 was predicted by both programs . The Nlu-miR-34 sequence , including the intact 5’UTR , was deposited in GenBank ( MG894367 ) . BPH nymphs were collected for microinjection on the first day of the third instar . 15 nl of agomir-34 ( 40 ng , 15 nl ) or antagomir-34 ( 40 ng , 15 nl ) were microinjected into the conjunctive of each nymph between the prothorax and mesothorax using a Nanoliter 2000 injector ( WPI , USA ) . 150 nymphs were used for each experiment . Ten nymphs were transferred into glass tubes 24 h post injection and placed in a growth chamber . Agomir or antagomir of random shuffled sequences was injected with equivalent volume as negative control ( RiboBio , China ) . All experiments were performed in triplicate . In order to knock down NlInR1 and NlBr-C expression , dsRNA was synthesized using the T7 RiboMAXTM Express RNAi System ( Promega , USA ) following the manufacturer’s instructions . Each nymph , at the first day of the third instar , was injected with dsRNA ( 250 ng , 15 nl ) using a Nanoliter 2000 injector ( WPI ) . Equivalent volumes of dsGFP were injected as negative controls . 100 nymphs were used for each treatment , and all experiments were carried out in triplicate . BPH nymphs were collected for microinjection on the first day of the third instar . The dsNlInR2 , the concentration of which were divided into three groups , 0 . 42 ng , 0 . 84 ng and 3 . 38 ng , was injected into nymphs respectively with 40 ng antagomir-34 . Either 40 ng or 60 ng antagomir-34 was injected into nymphs with 0 . 84 ng dsNlInR2 . The volumes of mixtures were 15 nl . All of microinjections were done using a Nanoliter 2000 injector ( WPI , USA ) . Approximately 200 nymphs were used for each experiment . Control nymphs were injected with equivalent volumes of the mixtures of dsNlInR2 and antagomir-NC . All experiments were performed in triplicate . Juvenile hormone III ( JH III , 96% ) was purchased from Toronto Research Chemicals ( Sigma , USA ) . 200 ng of JH III was dissolved in acetone and dropped 40 nl onto the mesonotum of the third instar BPH nymphs using a Nanoliter 2000 injector ( WPI , USA ) , after anaesthetizing with carbon dioxide [14] . Equivalent volumes of acetone were used as negative controls . The rearing condition and methods of observation were same as the method of RNAi and injection of agomir-34 and antagomir-34 . 100 nymphs were used for each experiment , all experiments were performed in triplicate . Total RNA , enriched for miRNAs and mRNA , was extracted using the miRNeasy Mini kit ( Qiagen , Germany ) and TransZol Up Plus RNA kit ( TransGen , China ) from whole insect bodies ( n = 30 ) , respectively . The miScript Ⅱ RT kit ( QIAGEN ) and the TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix kit ( TransGen ) were used to prepare cDNA . Quantitative real-time PCR ( qPCR ) reactions were conducted with an ABI Prism 7300 ( Applied Biosystems , USA ) using miScript SYBR Green ( Qiagen , Germany ) and SYBR® Premix Ex Taq II RR820A ( Takara , Japan ) , according to the manufacturer’s instructions . The data was analyzed using the 2-△△Ct method . Actin1 ( GenBank accession No . EU179846 . 1 ) was used as the endogenous control . This gene has been studied to show it is suitable to be used as a reference gene [46] . All treatments were carried out in triplicate . Independent reactions were performed in quadruplicate for each RNA sample , and the signal intensity of the target gene is presented as the average value . The qPCR primers used in this study are listed in S3 Table . About 300 BPHs were collected and pooled together for analysis . Extraction of JH III was performed according to a previously described method with some modifications [47 , 48] . JH III dilutions of 20 , 10 , 1 , 0 . 5 , 0 . 25 , 0 . 125 , and 0 . 0625 ng/ml were prepared in 70% methanol and used for calculation of the standard curve . The LC–MS/MS analyses were performed using an Agilent 6460 triple quadrupole mass spectrometer ( Agilent , USA ) equipped with an electrospray ionization ( ESI ) source , operated in the positive ion multiple-reaction monitoring ( MRM ) mode . Agilent Mass Hunter Workstation software was used for system operation , data acquisition , and data analysis . Chromatographic separation was carried out using a Zorbax SB-Aq column ( 100 mm × 2 . 1 mm , 3 . 5 μm ) ( Agilent , USA ) . JH III was separated using a binary gradient . Briefly , mobile phase A consisted of 0 . 1% formic acid in water and mobile phase B was comprised of 0 . 1% formic acid in acetonitrile . The gradient program was as follows: 0–5 min , 60%–85% of B; 5–8 min , 85%–95% of B . The flow rate of the mobile phase and the column temperature were set at 0 . 3 mL/min and 30°C , respectively , and the injection volume was 5 μl . The retention time of JH III was 3 . 82 min and the total run time was 10 min . Mass spectrometric detection was completed by use of an electrospray ionization ( ESI ) source in positive ion multiple-reaction monitoring ( MRM ) mode using the following parameters: precursor ( m/z ) , 267; ion product ( m/z ) , 235 . 2; fragmentor voltage ( V ) , 80; collision Energy ( eV ) , 1 . SPSS 17 . 0 software ( IBM SPSS Inc . USA ) was used for statistical analysis . The differences between treatments were compared using the Student’s t-test . The effects of antagomir on distribution of wing types in BPHs were analyzed using the Chi-square test . p < 0 . 05 was considered statistically significant . All results are expressed as means ± SEM , three replicates . | Polyphenism is a fascinating phenomenon which significantly improves the ability of a species to explore various environmental resources . Brown planthopper ( Nilaparvata lugens , BPH ) is a notorious insect pest which causes huge damages to rice . BPH has two wing phenotypes , long-winged ( LW ) and short-winged ( SW ) morphologies . LW morphs are capable of long-distance migration , while SW morphs have high reproductive capabilities . Juvenile hormone ( JH ) and insulin/IGF signaling ( IIS ) pathway have been known to participate in regulating polyphenism in various organisms . However , how these regulators interact with each other remains largely unknown . Here , we show that a conserved microRNA , miR-34 , mediates the crosstalk between JH , 20E and IIS pathway to modulate wing polyphenism in BPH . miR-34 suppresses insulin receptor-1 ( InR1 ) , increases JH titer , and induces SW morphs . On the other hand , JH increases the expression of miR-34 and induces SW morphs , while 20E decreases miR-34 but does not change proportion of wing morphs . Knocking down genes in IIS pathway changes JH titer and miR-34 abundance . Therefore , miRNA , JH , 20E and IIS form an autoregulatory feedback loop to control wing polyphenism in BPH . Our work presents a comprehensive mechanism of wing polyphenism by integrating various regulators . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"luciferase",
"molecular",
"probe",
"techniques",
"natural",
"antisense",
"transcripts",
"gene",
"regulation",
"enzymes",
"enzymology",
"hormones",
"developmental",
"biology",
"regulator",
"genes",
"micrornas",
"animal",
"anatomy",
"nymphs",
"gene",
"types",
"molecular",
"biology",
"techniques",
"zoology",
"research",
"and",
"analysis",
"methods",
"probe",
"hybridization",
"proteins",
"animal",
"wings",
"hyperexpression",
"techniques",
"gene",
"expression",
"life",
"cycles",
"oxidoreductases",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"rna",
"fluorescent",
"in",
"situ",
"hybridization",
"anatomy",
"nucleic",
"acids",
"cytogenetic",
"techniques",
"genetics",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna"
]
| 2019 | miR-34 modulates wing polyphenism in planthopper |
The mechanism for cortical folding pattern formation is not fully understood . Current models represent scenarios that describe pattern formation through local interactions , and one recent model is the intermediate progenitor model . The intermediate progenitor ( IP ) model describes a local chemically driven scenario , where an increase in intermediate progenitor cells in the subventricular zone correlates to gyral formation . Here we present a mathematical model that uses features of the IP model and further captures global characteristics of cortical pattern formation . A prolate spheroidal surface is used to approximate the ventricular zone . Prolate spheroidal harmonics are applied to a Turing reaction-diffusion system , providing a chemically based framework for cortical folding . Our model reveals a direct correlation between pattern formation and the size and shape of the lateral ventricle . Additionally , placement and directionality of sulci and the relationship between domain scaling and cortical pattern elaboration are explained . The significance of this model is that it elucidates the consistency of cortical patterns among individuals within a species and addresses inter-species variability based on global characteristics and provides a critical piece to the puzzle of cortical pattern formation .
Cerebral cortical patterns have fascinated scientists for centuries with their beauty and complexity . Numerous groups relate malformations in sulcal patterns to different diseases in humans , such as autism [1] and attention deficit/hyperactivity disorder ( ADHD ) [2] . Though many advances have occurred in cortical development and sulcogenesis , the understanding of how sulci form and what factors determine the placement of sulci is still limited . The cerebral cortex across species displays a variety of shapes and sizes and also wide array of sulcal patterning . Studying the evolutionary development of sulcal patterns might provide clues about the cortical development taking place in humans . A major advance in determining how these sulcal patterns form was the introduction of the axonal tension hypothesis [3] . This hypothesis describes a mechanically-based scenario where axonal tension , created by developing corticocortical connections in strongly interconnected regions , pulls together gyral walls and creates a folding pattern . This hypothesis furthered the concept that variability between folding patterns among individuals is genetically driven , not just the consequence of random mechanical buckling from a confined cortex . Other mechanochemical models have also been proposed to explain morphogenesis in the central nervous system [4] . Recently , it has been suggested that a cortical pattern can arise based on regional patterns of intermediate progenitor ( IP ) cells in the subventricular zone ( SVZ ) [5] . The intermediate progenitor model , which builds upon the intermediate progenitor cell hypothesis [6] , states that during the development of the cortex certain radial glial cells in the ventricular zone ( VZ ) are activated to create IP cells that travel to the SVZ . These IP cells amplify the amount of neurons created in a given radial column . Furthermore , a subset of IP cells creates a local amplification of neurons in upper cortical layers surrounded by areas of non-amplification , resulting in a wedge shape in the cortex . This wedge shape is representative of a gyrus . This new hypothesis is still being debated [7] , [8] and , if correct , could be a scenario for chemically-based pattern formation in the cortex . Here , a relatively simple and , we believe , elegant chemically-driven mathematical model is proposed to explain how IP cell subsets are distributed spatially and temporally in the developing cortex . Our model , which we call the Global Intermediate Progenitor ( GIP ) model , uses a Turing reaction-diffusion system [9] containing an activator and inhibitor on a prolate spheroidal surface to determine regional areas of activation of the production of IP cells . The GIP model allows determination of the placement of the initial sulci underlying observed complex cortical patterns . It also demonstrates that the initial folds of the arising sulcal pattern are governed by the global shape of the lateral ventricle . The dependency on the global shape provides a critical piece to the puzzle of cortical development .
The patterns created by Turing reaction-diffusion systems have been used to describe pattern formation in numerous biological systems [12] . Though biological Turing patterns have not been proven as rigorously as chemical Turing patterns [13] , recent results [14] give supporting evidence of Turing patterns formation in a biological setting . A Turing system is a reaction-diffusion system , given by ( 1a ) ( 1b ) containing an activator ( U ) and an inhibitor ( V ) that are diffusing throughout their domain and interacting with each other as described by the reaction kinetics ( F and G ) . The reaction kinetics chosen for the GIP model are from the Barrio-Varea-Maini ( BVM ) [15] system given by ( 2a ) ( 2b ) where ( u , v ) = ( U−U0 , V−V0 ) and ( U0 , V0 ) is the steady state . Not much is known about the possible interactions between u and v that regulate the production of IP cells . Hence the BVM system is ideal because the kinetic equations do not assume any prior knowledge of how the reactants ( activator and inhibitor ) interact and instead takes a phenomenological approach . These kinetics ( Equations 2a , b ) also provide control over the amount of linear ( α and γ for u; and β for v ) , quadratic ( r2 ) , and cubic ( r1 ) interactions . The diffusivity ratio and domain scaling are given by d and δ , respectively . Key aspects of the system that determine what pattern will arise include the ratio of the diffusivities of the activator and inhibitor , domain scale and shape , and quadratic versus cubic terms in the kinetic reactions [12] , [16] . In order to analyze the nonlinear BVM system , the kinetics are approximated linearly by expanding them in a Taylor series around the steady state and neglecting higher order terms resulting in ( 3 ) Solutions of Equation 3 are of the form ( u , v ) = T ( t ) X ( x ) . The temporal solution is T ( t ) = eλ ( k2 ) t , where λ ( k2 ) is the temporal eigenvalue . The spatial solution solves the Helmholtz Equation ( ∇2X+k2X = 0 ) in the given domain where k2 is the spatial eigenvalue . Turing patterns have been studied in depth in 1D [12] , 2D [17] , and spherical domains [18] . In all these domains , the solution to the domain's associated eigenvalue problem can predict which pattern will form . For a spherical domain , the eigenvalue solution yields k2 = n ( n+1 ) /r2 , where n is the spherical harmonic index and r is the radius . An increase in k2 , which depends on domain scaling when diffusion coefficients are held constant , results in an increase in n and changes the predicted Turing pattern . Here , we derive a formula that predicts the Turing pattern observed on a prolate spheroidal surface which represents the SVZ . A prolate spheroid is created by rotating an ellipse about its major axis . It has a focal distance , where a and b are the major and minor axes , respectively . Spheroidal coordinates are expressed as ( ξ , η , φ ) where ξ is the radial term; η = cos θ , where θ is the asymptotic angle with respect to the major axis; and φ is the rotation term . To predict which pattern will emerge , the Helmholtz equation is expanded with respect to the prolate spheroidal coordinate system [19] resulting in ( 4 ) where . Because the Helmholtz equation ( ∇2X+k2X = 0 ) is separable in prolate spheroidal coordinates , we rewrite X in terms of X = R ( c , ξ ) S ( c , η ) Φ ( φ ) , such that S ( c , η ) , R ( c , ξ ) , and Φ ( φ ) satisfy ( 5 ) ( 6 ) ( 7 ) where m and ρ are separation constants . Since multiple , discrete values of ρ are possible for a given m and ρ is also dependant on c , the notation will be ρmn ( c ) . Because our domain is a prolate spheroidal surface , the radially-invariant solution is needed ( i . e . ) . In order for Equation 6 to hold , must equal zero for a nontrivial solution . This results in ( 8 ) where ξ0 is the spheroidal radius of the shell that conserves a surface area of 4π ( comparable to the surface area of a unit sphere ) . The significance of the formula in Equation 8 is that it relates a given domain size ( controlled by k2 ) and domain shape ( the eccentricity of the prolate spheroid controlled by f ) to the arising pattern . To demonstrate this formula's ability to predict pattern formation , the system ( Equations 2a , b ) is discretized similar to that of a sphere [18] . A forward-Euler finite difference scheme is used and u and v are discretized such that u ( η , φ ) = ( −1+h1dη , h2dφ ) where h1 = 0 , ‥ , 34 , and h2 = 0 , ‥ , 68 . The continuity with respect to η around the north and south poles is maintained as described in [18] and periodic boundary conditions are used for φ . In Figure 2A , Amn is plotted for n = 0 , ‥ , 7 and m = 0 , ‥ , n . Numerous simulations were executed and two are shown here . The first simulation ( Figures 2B and 2C ) corresponds to k2 = 30 . When k2 = 30 ( asterisk in middle ) is plotted on the Amn vs . k graph ( Figure 2A ) , k2 corresponds with A35 and predicts a ( 3 , 5 ) pattern that agrees with the numerical simulation ( Figures 2B and 2C ) . The second simulation corresponds to k2 = 60 ( Figure 2D and 2E ) and when plotted ( top right asterisk ) on the Amn vs . k graph , predicts a ( 7 , 7 ) pattern that is observed in the numerical simulation .
The model presented here addresses the directionality of the initial sulci formed . In order to use the predictive power of the proposed prolate spheroidal harmonic system , sulci need to be formulated in terms of prolate spheroidal harmonics . Since the initial sulcal formations mimic stripes , only the prolate spheroidal harmonics resulting in striped patterns were studied . In order to form a sulcus , the gyral banks on either side of the sulcus need to be created . In terms of the production of IP cells , the areas on either side of the sulcus will need to be ‘activated’ while the area of the sulcus is ‘not activated’ ( see Figure 3B and 3E ) . The two sulcal directions considered are sectorial and transverse . Sectorial sulci extend in the direction from the frontal lobe around the Sylvian fissure to the temporal lobe . This represents the direction of the major axis of the prolate spheroid approximating the lateral ventricle as shown in Figure 3A . The alignment of sulcal pits ( deepest part of sulcus ) along the major axis of the lateral ventricle in the human has been shown [21] . In terms of spheroidal harmonics , the pattern of IP cells needed to create sectorial sulci is ( m , n ) = ( 1 , 1 ) for 1 sulcus ( Figure 3C ) , ( 2 , 2 ) for 2 sulci , and so forth . Transverse sulci form in the direction of rings around the VZ as shown in Figure 3D . This direction corresponds to ( 0 , 2 ) for 1 sulcus ( Figure 3F ) , ( 0 , 4 ) for 2 sulci , and so forth . In each species displaying a cortical pattern , a number of sectorial sulci ( or sulcal pits ) are observed . The exact number of sectorial sulci is not the focus here . Of interest , rather , are the occurrence of a transverse sulcus , the transition from transverse to sectorial sulci , and the role of lateral ventricular eccentricity . For f = 3 ( Figure 4A ) , as the domain scaling ( k2 ) increases , A11 is reached first , followed sequentially by A02 and A04 . This sequence corresponds to a sectorial sulcus forming first . If the focal distance is increased , e . g . if f = 4 ( Figure 4B ) , there is a shift in the Amn curves and , as k2 increases , A02 will now occur before A11 . This results in a transverse sulcus forming before the first sectorial sulcus . A further increase in focal distance to f = 6 ( Figure 4C ) again shifts the Amn curves , so that A04 now occurs before A11 . Two transverse sulci will now form before a sectorial sulcus is created . These scenarios illustrate how focal distance plays a role in determining the order of pattern formation .
The GIP model illustrates how sulcal placement and directionality is related to changes in focal distance . In order to determine the effect of changes in focal distance on cortical pattern formation , the evolutionary development of cortical patterns was examined . The lateral ventricle is a c-shaped cavity with an anterior horn that extends into the frontal lobe of the hemisphere and an inferior pole that enters the temporal lobe [11] . During the critical stages of brain development the volume of the lateral ventricle increases [22] which also increases the surface area of the lateral ventricle ( i . e . k2 increases ) . Also , as species have evolved the neocortex has expanded , resulting in major evolutionary advances [23] . As the frontal and temporal lobes expand , the lateral ventricle extends into the lobes increasing the lateral ventricular eccentricity resulting in changes in the cortical pattern obtained . For example , overlaying an evolutionary ladder on the scenarios described in Figure 3 implies that the cortices of species on the lower rungs of the evolutionary ladder , such as the cat , do not display transverse sulci before the formation of sectorial sulci ( Figures 4A and 5A ) . Following this evolutionary ladder , at some point the first transverse sulcus appears , as shown in Figure 4B . This second stage is representative of the formation of the calcarine sulcus in species such as the lemur ( Figure 5B ) . Further along the evolutionary ladder , the second transverse sulcus appears ( Figure 4C ) . This is representative of the central sulcus found in higher order primates such as the human ( Figure 5C ) . For humans , this predicted ordering of sulcal formation correlates well with what has been observed during development through the examination of naturally aborted fetuses [24] and MRI study on preterm infants [25] . The first sulci to appear are the anterior calcarine and central sulcus which are in the transverse direction ( blue lines in Figure 5C ) . This is followed by the formation of the superior and inferior frontal sulci , superior and inferior temporal sulci , the intraparietal sulcus and the cingulated sulcus; all which form in the sectorial direction ( red lines in Figure 5C ) . The GIP model also provides a plausible explanation for the development of the central sulcus . Lemurs and humans are both members of the primate order . The lemur is of the suborder prosimian , which is the most primitive of the primates [26] . Most prosimians can be distinguished from anthropoids , the higher primates , by the absence of the central sulcus [26] . Therefore , this model links evolutionary development , through the lateral ventricular eccentricity , to the development of the central sulcus . The GIP model is a theoretical model that builds upon the ideas of the IP model . One argument that has been presented against the intermediate progenitor model is that an “elaborately choreographed set of developmental instructions [regulating the production of IP cells] would be required to account for the tremendous complexity of human cortical convolutions” [7] . The beauty of the GIP model is that it provides an uncomplicated approach that relates to a biologically plausible mechanism of pattern formation . It uses chemical morphogens that may be governed by specific genes to control IP cell production , resulting in the ability to predict the placement and directionality of sulcal pattern formation . The GIP model reveals the role that the global shape of the lateral ventricle has on the positioning of the initial sulci during cortical development . This model explains the development of the initial folds , particularly how two transerve sulci can form before any sectorial sulci in the human . There are many sulci , such as the precentral and postcentral sulcus , that form after this event which are not in the scope of this present work . Also , we believe the Sylvian fissure is formed by the c-shape of the lateral ventricle , which is not applicable to the model . Lateral ventricular shape , or shape of any nontrivial object , is not easy to quantify . The GIP model approximates the lateral ventricle with a prolate spheroid allowing the capture of key shape characteristics in one parameter , the focal distance ( f ) . This approximation also gives the resulting patterns in terms of prolate spheroidal harmonics which contain an order based on the prolate spheroidal indices , m and n . The Helmholtz equation could also be solved on a given triangulated mesh representing the lateral ventricle resulting in a set of eigenvalues and eigenfunctions . The eigenfunction whose associated eigenvalue produces diffusion-driven instability would be the predicted pattern formed . A drawback of this latter approach occurs when comparing predicted patterns from different triangulated meshes . Since each mesh has its own parameterization there is no way of knowing which shape characteristic is responsible for the change in pattern formation . Although changes in the volume of the lateral ventricle in humans during the developmental stages are documented [22] , quantified data on the size and shape of the lateral ventricle during these critical stages is lacking . Further investigations into the size and shape of the lateral ventricle during developmental stages across species are needed . Such parameters could then be incorporated into the GIP model to test its cortical patterning predictions for specific species . Also , further investigations into how the production of IP cells is regulated ( i . e . how the activator and inhibitor interact ) would enhance this model . Several genes ( Pax6 , Ngn2 , and Id4 ) have been shown to modulate the production of IP cells in mice [10] . Further studies into how this modulation occurs , and if this modulation changes evolutionarily , could be incorporated into the reaction kinetics in the GIP model enhancing the cortical patterning predictions . In conclusion , this chemically-based mathematical model ( the global intermediate progenitor ( GIP ) model ) extends the intermediate progenitor model [5] , which describes local phenomena , to encapsulate global characteristics . In doing so , the GIP model shows how the global shape of the lateral ventricle , which drives the shape of the VZ , plays a key role in cortical pattern development . This model is able to capture changes in VZ shape along with the complementary role of domain scaling in only two parameters: 1 ) the focal distance of the prolate spheroid approximating the lateral ventricle , and 2 ) k2 , which is dependent on domain scaling , as given by the formula in Equation 8 . The model also has the ability to predict why the cortex of certain species may have little or no folding , and it accounts for the order and directionality of the sulci formed in different species . We consider this model a first step toward a chemically driven and mathematically predictive explanation of cortical folding development across species . | The size and shape of the cerebral cortex varies across species . The cortical folding pattern also varies from a smooth surface where no pattern is visible , as observed in the common treeshrew ( Tupaia glis ) and Eastern mole ( Scalopus aquaticus ) , to an intricate labyrinthine pattern , as observed in humans . One current model , the intermediate progenitor model , describes the creation of a fold through local interactions in the ventricular zone which surrounds the lateral ventricle . Here we extend the local scenario described in the intermediate progenitor model to include global characteristics that differ between species . We approximate the lateral ventricle with a prolate spheroid and examine how patterns on a spheroidal surface change based on size and eccentricity . Our model reveals a direct correlation between pattern formation and lateral ventricular size and shape . This model's significance is that it elucidates the consistency of cortical patterns among individuals within a species and addresses inter-species variability based on global characteristics , such as size and shape of the lateral ventricle , and provides a critical piece to the puzzle of cortical pattern formation . | [
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion"
]
| [
"developmental",
"biology/morphogenesis",
"and",
"cell",
"biology",
"mathematics",
"developmental",
"biology/developmental",
"evolution",
"developmental",
"biology/pattern",
"formation",
"neuroscience/neurodevelopment",
"developmental",
"biology/neurodevelopment",
"evolutionary",
"biology/pattern",
"formation",
"evolutionary",
"biology/morphogenesis",
"and",
"cell",
"biology",
"neuroscience/theoretical",
"neuroscience",
"evolutionary",
"biology/developmental",
"evolution"
]
| 2009 | Chemically Based Mathematical Model for Development of Cerebral Cortical Folding Patterns |
Massively parallel RNA sequencing ( RNA-seq ) in combination with metabolic labeling has become the de facto standard approach to study alterations in RNA transcription , processing or decay . Regardless of advances in the experimental protocols and techniques , every experimentalist needs to specify the key aspects of experimental design: For example , which protocol should be used ( biochemical separation vs . nucleotide conversion ) and what is the optimal labeling time ? In this work , we provide approximate answers to these questions using the asymptotic theory of optimal design . Specifically , we investigate , how the variance of degradation rate estimates depends on the time and derive the optimal time for any given degradation rate . Subsequently , we show that an increase in sample numbers should be preferred over an increase in sequencing depth . Lastly , we provide some guidance on use cases when laborious biochemical separation outcompetes recent nucleotide conversion based methods ( such as SLAMseq ) and show , how inefficient conversion influences the precision of estimates . Code and documentation can be found at https://github . com/dieterich-lab/DesignMetabolicRNAlabeling .
Changes in gene expression are frequently observed in pathological conditions . In the simplest model [1] , steady state RNA levels are governed by synthesis ( transcription ) and degradation rates ( RNA stability ) . A paradigm is the generation of the hypoxic response in pathological conditions such as heart insufficiency [2] and fast growing tumors [3] . Hypoxia ( <2% O2 ) results in a global decrease of total transcription [4] . However , the transcription of specific target genes is induced under hypoxic conditions by hypoxia inducible factor 1 ( HIF1 ) [5] , which is composed of a stable β-subunit and an oxygen labile α-subunit [6] . Furthermore , different RNA binding proteins such as HuR and TTP as well as miRNAs regulate the stability of their cognate target mRNAs dependent on oxygen availability [7] and contribute to changes in gene expression profiles . Metabolic labeling experiments are a versatile tool to discern dynamic aspects in physiological and pathological processes . These experiments drive our understanding of key processes in molecular systems , such as synthesis and decay of metabolites , DNA , RNA and proteins . Pulse-chase experiments help to determine the kinetic parameters of synthesis and decay in various contexts . In the pulse phase of an experiment , the label is introduced to newly synthesized compounds and unlabeled or pre-existing molecules are only subjected to degradation or some other form of processing . In contrast , during the chase phase , the label in the system is gradually replaced by unlabeled compounds . A typical metabolic labeling experiment may include a pulse , a chase or both phases . The first transcriptome-wide studies by [8] and [9] used 4-thiouridine ( 4sU ) labeling in cell culture experiments to infer kinetic parameters . This approach has become quite popular in RNA biology , which is shown by a vastly increasing number of studies ( see [10] for review ) . Massively parallel RNA sequencing ( RNA-seq ) in combination with metabolic labeling has become the de facto standard approach to study alterations in RNA transcription , processing or decay at the transcriptome-wide level . At the time of writing , the most widely used approach involves metabolic labeling with thiol-labeled nucleoside analogs such as 4sU ( 4sU-tagging ) [11] . Briefly , total cellular RNA is isolated and thiol groups are biotinylated . Subsequently , total cellular RNA can be efficiently separated into newly transcribed ( labeled ) and pre-existing ( unlabeled ) RNA . Very recent innovations are new methods involving the chemical conversion of 4sU residues into cytosine analogs , which is observed as point mutations in RNA-seq data ( T-to-C transitions ) , ( see [12] , [13] and [14] ) . The absence of any biochemical separation method makes metabolic labeling more accessible due to lower input amounts and less laborious protocols . Regardless of all advances in the experimental protocols and techniques , a few important questions remain to be answered by any experimentalist , namely the specific characteristics of experimental design: what should be measured ( i . e . sequenced ) and when ? For example , which approach should I take ( e . g . biochemical separation vs . nucleotide conversion ) , when should I collect my samples ( e . g . time points in a pulse experiment ) and how could this affect my estimates on kinetic parameters . In [15] , the authors proposed guidelines for the design of metabolic labeling experiments , however they provide no kinetic or statistical models for the optimization of such experiments . Within this manuscript , we use kinetic and statistical models to infer the degradation rates from a pulse experiment ( see Fig 1 and Eqs 1 and 2 ) , and derive several aspects on the optimal design of metabolic RNA labeling experiments . We illustrate these implications on a pulse-chase SLAMseq data set [12] and an example for a pulse labeling experiment with biochemical separation .
MCF-7 cells ( ACC-115 ) were obtained from the Leibniz Institute DSMZ German Collection of Microorganisms and Cell Cultures . Cells were routinely tested for mycoplasma contamination with Venor GeM Classic ( Minerva Biolabs ) . MCF-7 cells were cultured at 37°C and 5% CO2 and maintained in DMEM ( Thermo Fisher Scientific ) supplemented with 10% fetal calf serum ( Merck ) , 1xMEM non-essential amino acids ( Thermo Fisher Scientific ) and 1xPenicillin/Streptomycin ( Thermo Fisher Scientific ) . MCF-7 cells were seeded 48 hrs prior to the experiment at a cell density of 0 . 3 × 105cells/cm2 . Cells were labeled with 4-thiouridine ( 4sU ) ( Sigma-Aldrich ) at a final concentration of 200 μM for 2 , 4 or 8 hrs . Cells were scraped in DPBS and the pellet resuspended in Trizol ( Thermo Fisher Scientific ) . Total RNA was isolated using the Trizol method . Briefly , the cell pellet was resuspended in 750 μl Trizol , and incubated 5 min at room temperature before addition of 200 μl chloroform . Samples were centrifuged ( 20 min , 10 . 000g , room temperature ) and the aqueous phase re-extracted with one volume chloroform: isoamylalkohol ( 24:1 ) ( 5 min , 10 . 000g , room temperature ) . The RNA in the aqueous phase was precipitated with one volume isopropanol ( 30 min , 20 . 8000g , 4°C ) , washed twice with 1 ml 80% ethanol in DEPC-H2O and dissolved in 25 μl DEPC-H2O ( 10 min , 55°C , shaking ) . For in vitro transcription of linearized plasmids ( pBSIIKS-Luc-pA-NB [16] and pBSIIKS-Renilla-pA [17] ) , the MEGAscript T7 Transcription Kit ( Thermo Fisher Scientific ) was used according to the manufacturers instructions . Briefly , the reaction was set up in a total volume of 20 μl containing 1 μg linearized plasmid and 2 μl 10x reaction buffer , 3 μl 40 mM m7GppG-cap analogon ( KEDAR ) , 2 μl 15 mM GTP , 2 μl 75 mM CTP , 2 μl 75 mM ATP , 2 μl enzyme mix and 2 μl 75 mM UTP ( for RLuc ) or 2 μl 75 mM 4-S-UTP:UTP in a 1:10 ratio ( for FLuc ) . Reactions were incubated 3 hrs at 37°C . Plasmid-DNA was removed by addition of 1 μl Turbo-DNase ( 15 min , 37°C ) . In vitro transcribed RNA was purified by phenol extraction and Chromaspin-100 ( Clontech ) purification . RNA was precipitated over night after addition of sodium acetate to a final concentration of 0 . 3 M and 2 . 5 volumes 100% ethanol . After centrifugation ( 30 min , 20 . 800g , 4°C ) the pellet was washed with 1 ml 80% ethanol and dissolved in 40 μl DEPC-H2O . Concentration was determined by Nanodrop ( Thermo Fisher Scientific ) measurement and integrity checked by agarose gel electrophoresis . Total RNA was spiked with in vitro transcribed 4sU-labeled FLuc and non-labeled RLuc RNAs and biotinylated using MTSEA biotin-XX ( Biotium ) as described by [18] . Briefly 80 μg total RNA was incubated with 8 ng FLuc and 4 . 8 ng RLuc ( equimolar amounts , 130 amol ) , 10 mM HEPES pH 7 . 5 , 1 mM EDTA and 5 μg MTSEA biotin-XX ( freshly dissolved in DMF ) in a total volume of 250 μl . Reactions were incubated 30 min in the dark at room temperature . Biotinylated RNA was recovered by extraction with one volume phenol: chloroform: isoamylalkohol ( 24:24:1 ) and separated using Phase-Lock-tubes ( 5Prime ) by centrifugation ( 5 min , 20 . 800g , room temperature ) . RNA was precipitated by addition of 350 μl isopropanol , 25 μl 5 M sodium chloride and 1 μl glycogen ( Roche Diagnostics , 20 μg/μl ) to assist precipitation ( 30 min , 20 . 800g , 4°C ) . RNA was washed twice with 500 μl 80% ethanol in DEPC-H2O and dissolved in 25 μl DEPC-H2O ( 10 min , 55°C , shaking ) . For purification of biotinylated RNAs the method described by [1] was adapted . 25 μg biotinylated total RNA was adjusted to 100 μl with DEPC-H2O and filled up with Streptavidin binding buffer ( Strep-BB ) ( 20 mM Tris , pH 7 . 4 , 0 . 5 M sodium chloride , 1 mM EDTA ) to 200 μl . RNA was denatured 10 min at 65°C and subsequently placed on ice . 100 μl magnetic streptavidin beads ( New England Biolabs ) were washed once with 200 μl Strep-BB and resuspended in 100 μl Strep-BB . RNA and beads were incubated 15 min at room temperature on a rotating wheel . Beads were washed three times with 500 μl Strep washing buffer ( 100 mM Tris pH 7 . 4 , 1 M sodium chloride , 10 mM EDTA , 0 . 1% Tween 20 ) prewarmed to 55°C . RNA was eluted three times by de-biotinylation with 100 μl freshly prepared 100 mM DTT and elution fractions pooled for further analysis . RNA was recovered from total RNA , flow through and eluate by phenol: chloroform: isoamylalkohol ( 24:24:1 ) extraction using Phase-Lock-tubes and isopropanol precipitation as described above . The amount of recovered RNA was determined by Nanodrop measurement . 1 μg biotinylated RNA was applied to nylon membrane ( Hybond-N , GE Healthcare ) using a dot blot device ( Carl Roth ) . RNA was crosslinked twice at 254 nm using the “Optimal Crosslink” mode of the Spectroline Select XLE-1000 crosslinker . The membrane was blocked 20 min with PBS + 10% SDS and incubated 2 hrs with Streptavidin-HRP ( Thermo Fisher Scientific , 1:5000 in PBS + 10% SDS ) . Prior to detection with SuperSignal West Pico ( Thermo Fisher Scientific ) the membrane was washed each three times 10 min with PBS + 10% SDS , PBS + 1% SDS and PBS + 0 . 1% SDS . Images were acquired with the LAS4000 system ( GE Healthcare ) . 1 μl RNA from streptavidin purification was reverse transcribed using the Maxima H Minus First Strand cDNA Synthesis Kit ( Thermo Fisher Scientific ) with Random Primers according to the manufacturers protocol . For absolute quantification reverse transcription reactions were set up with different amounts of spike in RNAs , ranging from 1600% to 1 . 56% for FLuc and 400 to 3 . 12% for RLuc in 1:2 dilutions . Briefly , RNA was mixed in a total volume of 15 μl with 1 μl Random Primer and 1 μl dNTP solution and denatured ( 5 min , 65°C ) . Reaction was completed by addition of 4 μl 5xRT buffer and 1 μl Maxima enzyme and incubated 10 min at room temperature followed by 30 min , 50°C and denaturation ( 5 min , 85°C ) . Reverse transcription reactions were diluted 1:10 and used for qPCR analysis on a StepOnePlus instrument ( ThermoFisherScientific ) with Power SYBR Green PCR Master Mix ( Thermo Fisher Scientific ) and primers directed against FLuc ( forward: CCTTCCGCATAGAACTGCCT , reverse: GGTTGGTACTAGCAACGCAC [19] ) and RLuc ( forward: GTTGTGCCACATATTGAGCC , reverse: CCAAACAAGCACCCCAATCATG [20] ) . Total and enriched samples were depleted for ribosomal RNA ( rRNA ) contamination using RiboZeroGold , which is based on the removal of rRNA with biotinylated oligos using streptavidin beads . Thus , also the biotinylated 4sU-labeled molecules were removed from the total samples by the RiboZeroGold procedure and were treated as flow through . Libraries of 2 biological replicate 4sU pulse experiment were sequenced 1x 50bp on an Illumina HiSeq4000 . All relevant details on sequencing depth and mapping rates are listed in S1 Table . Sequencing adapters and low-quality reads were removed from the raw sequencing data with flexbar v3 . 0 . 3 [21] using standard filtering parameters . We excluded all reads with more than 1 uncalled base from the output . All remaining reads ( >18bp ) were then aligned to a custom sequence index including rRNA , tRNA and snoRNA gene loci using bowtie2 with the –very-fast option [22] . Only reads that did not align to any of the contaminant sequences were considered for further analysis . Reads were then aligned to the human genome ( EnsEMBL 85 ) and splice sites from the reference annotation with a splice-aware aligner ( STAR , v2 . 5 . 3a; [23] ) . The BAM files were analyzed with StringTie 1 . 3 . 3b [24] and the final read count matrix was prepared with the supplemented python script prepDE . py .
We describe RNA-seq read counts with the negative binomial distribution , which is widely used in this setting and accounts for overdispersion [25] . For a given gene , the read count follows X ∼ NB ( m ( μ , δ , t ) , k ) , where m is the mean read count , which depends on the time of labeling t , the degradation rate δ and the expression level in the steady-state μ , and k is the overdispersion parameter of the negative binomial distribution NB . In this case , the variance is var ( X ) = m ( m + k ) /k , where low k values correspond to high overdispersion in the data . We describe the RNA amount m in metabolic labeling experiments using simple first order kinetics: d m d t = s - δ m , ( 1 ) where s is the synthesis rate and δ is the degradation rate . In a steady-state , the expression level of a gene is μ = s/δ . The expression level μ can be derived from the total fraction , which ensures identifiability of at least this parameter . For that reason , we use μ and δ to parametrize the model . In this section , we only discuss the case of pulse labeling experiments throughout . However , our considerations extend to chase labeling experiments , where the equations are the same , except that the labeled fraction behaves as the unlabeled one in the pulse experiment and vice versa . For simplicity , we assume that fraction cross-contamination is negligible , in which case , RNA amounts for a given gene are proportional to the means mL , mU and mT derived from the kinetics for labeled , unlabeled and total fractions scaled by sample-specific factors xi ( see Eq 4 in section 2 of Extended Methods ) : m T ( t ) = 1 · μ m L ( t ) = x L μ ( 1 - e - δ t ) m U ( t ) = x U μ e - δ t ( 2 ) Here we treat the mean read count in the total sample as a reference ( coefficient is 1 ) , to make the system identifiable . In the case of labeled and unlabeled fractions , expected read numbers must be scaled by additional coefficients , xU and xL , because the RNA material can be normalized by different degrees during library preparation from chemically separated fractions . A preservation of the ratio of labeled to unlabeled fractions ( see Fig 1 ) yields xU = xL . If the sequencing depth is approximately the same for all samples , we may assume for simplicity xU = xL = 1 , and in this case , mT ( t ) = mL ( t ) + mU ( t ) = μ . In the conventional approach , where labeled and unlabeled molecules are separated , xU ≠ xL , the fraction ratio must be inferred from the data itself or by using an external normalization by spiking in labeled and unlabeled known molecules [26] . In the presence of cross-contamination , the estimations for the rates are biased depending on the relation of the labeling time and the degradation rate: if δt ≪ 1 ( slow rate ) , the bias is towards faster rate values , and , if δt ≫ 1 ( fast rate ) , it is towards slower rate values , for more details see Eqs 13 and 14 , section 2 . 1 in Extended methods . Efficiency of separation procedure may vary between species due to different uridine content , which can be another source of bias , see section 2 . 2 in Extended methods . This phenomenon can be modeled by introducing an additional coefficient to the model , see , for example , [27] and [28] . Although both sources of a bias may potentially affect estimates of certain RNA species , they are beyond the scope of our current work . Here , we concentrate on theoretical results , which are derived from statistical properties of our outlined model . In the following , we discuss pulse labeling experiments with different labeling times t . On the one hand , subtle changes in the RNA level are masked by the measurement noise for short labeling times . On the other hand , estimations at long labeling times are also less informative , because the difference between the steady state level and the RNA levels at time t is negligible and will be masked by the noise as well . To estimate the degradation rate δ from the RNAseq read counts , we use the method of maximum likelihood estimation ( MLE ) . This estimator δ ^ varies from experiment to experiment , and one is interested to minimize its variance , as a large variance results in large confidence intervals and , hence , poor estimates of the true δ . In this paper , we use the asymptotic properties of the MLE , when the number of experiment repetitions n → ∞ , in which case the system can be treated analytically [29 , 30] . Under regularity conditions , the MLE θ ^ is asymptotically normally distributed: n ( θ ^ - θ ) ∼ N ( 0 , I 1 - 1 ( θ ) ) , ( 3 ) where I 1 ( θ ) is the Fisher information matrix ( FIM ) for a single experiment repetition [29 , 30] . The FIM characterizes the curvature of the log-likelihood function L ( θ , X ) near the true parameter values θ and is defined as I ij ( θ ) = - E ∂ 2 log L ( θ , X ) ∂ θ i ∂ θ j . ( 4 ) We assume that the overdispersion parameter k is shared between all genes and neglect the uncertainty in δ propagating from k , i . e . only two parameters , δ and μ , are used to construct the FIM: I ( θ ) = ( I δ δ ( θ ) I δ μ ( θ ) I δ μ ( θ ) I μ μ ( θ ) . ) ( 5 ) The FIM is additive , i . e . if I U ( θ ) and I L ( θ ) correspond to the labeled and unlabeled fractions , the total FIM for the experiment is I ( θ ) = I U ( θ ) + I L ( θ ) , and for n such repetitions , I ( θ ) = n ( I U ( θ ) + I L ( θ ) ) . The diagonal terms of the inverse FIM estimate the variance of θ i ^ var ( θ i ^ ) = ( I - 1 ( θ ) ) ii . ( 6 ) In some cases we use 1 / I ii ( θ ) as a lower bound for ( I - 1 ( θ ) ) ii . Since ( I - 1 ( θ ) ) δ δ = ( I δ δ ( θ ) - I δ μ ( θ ) I μ δ ( θ ) / I μ μ ( θ ) ) - 1 , ( 7 ) and using the fact that I δ μ ( θ ) = I μ δ ( θ ) and I μ μ ( θ ) > 0 , the diagonal term of the inverse matrix is bounded as ( I - 1 ( θ ) ) δ δ ⩾ 1 / I δ δ ( θ ) . ( 8 ) ( I - 1 ( θ ) ) δ δ = 1 / I δ δ ( θ ) if there is no uncertainty , propagating from other parameters , i . e . I δ μ ( θ ) = 0 . Since the FIM I ( θ ) depends on the experiment parameters , such as the labeling time t and the sequencing depth , it is our main interest to reduce the variance of the MLE by selecting the optimal conditions accordingly . Due to additive property of the FIM , it suffices to optimize the FIM of a single experiment repetition . In the case of multiple parameters , it may be not possible to achieve the minimal variance for all parameters at the same time . Different criteria can be constructed as a combination of the elements of the inverse FIM [29 , 31] . We are interested to optimize the estimation of δ only and do not consider variance of the expression level estimator μ ^ in the design criteria . Let us consider first a simpler experimental setup , which preserves the fraction ratio ( e . g . SLAMseq ) . Here we first discuss the case of the Poisson model , which corresponds to the case of no overdispersion ( k → ∞ ) . The derivations for the Poissonian and for more general cases are left to section 3 of the Extended Methods , see Eqs 25 and 26 . Let XL and XU be the read counts corresponding to the labeled and unlabeled molecules for a given gene in a SLAMseq sample , and let t be the time of labeling . In this case , the inverse FIM is diagonal: I slam - 1 ( θ ) = ( I L ( θ ) + I U ( θ ) ) - 1 = ( e δ t - 1 μ t 2 00μ ) ( 9 ) The parameters δ and μ are information orthogonal , because I δ μ ( θ ) = 0 and inference about δ can be done as μ were known exactly . Indeed , for XL ∼ Pois ( mL ( t ) ) , XU ∼ Pois ( mU ( t ) ) , the conditional distributions P ( XL|XU + XL ) and P ( XU|XU + XL ) are binomial with the rates mU ( t ) / ( mU ( t ) + mL ( t ) ) = e−δt and mL ( t ) / ( mU ( t ) + mL ( t ) ) = 1 − e−δt and do not depend on μ . This model was recently discussed in a Bayesian framework for SLAMseq experiments by [32] . For a diagonal I ( θ ) , the inverse term ( I slam − 1 ( θ ) ) δ δ = ( ( I slam ( θ ) ) δ δ ) − 1 = ( ( I U ( θ ) ) δ δ + ( I L ( θ ) ) δ δ ) − 1 . The maximum of the term ( I slam ( θ ) ) δ δ corresponds to the minimal asymptotic variance of δ ^ due to Eq 3 . By optimizing ( I slam ( θ ) ) δ δ with respect to t , we get t slam = 1 . 59 τ , ( 10 ) where τ = 1/δ is the characteristic time of degradation . That means , if one optimizes the SLAMseq experiment and targets the gene with the characteristic time of degradation τ , the measurement at time point 1 . 59τ corresponds to the asymptotically optimal design . For example , if one is interested in an RNA species with half-life time of λ = 1 hr ( i . e . the characteristic time τ = λ/log ( 2 ) ≈ 1 . 44 hr ) , a pulse phase of 1 . 59 × 1 . 44 ≈ 2 . 3 hr corresponds to the asymptotically optimal design . In Fig 2A , we depicted the dependency of ( I slam ( θ ) ) δ δ and corresponding values of ( I U ( θ ) ) δ δ and ( I L ( θ ) ) δ δ as functions of normalized time t/τ for the degradation rate δ = 1 . Interestingly , ( I U ( θ ) ) δ δ and ( I L ( θ ) ) δ δ achieve maximum at tU = 2τ and tL ≈ 0 . 64τ , and the main contribution to the sum ( I slam ( θ ) ) δ δ = ( I U ( θ ) ) δ δ + ( I L ( θ ) ) δ δ comes from the term corresponding to labeled counts at shorter labeling times , and from the term for unlabeled counts at times longer than τ , see Fig 2A . Usually one is interested to measure a rate with a certain relative precision . To reflect this , we normalize the variance of the degradation rate estimator by δ2: var ( δ ^ ) δ 2 ≈ 1 I δ δ ( θ ) δ 2 . ( 11 ) Using a non-dimensional substitute α = t/τ , the corresponding denominator terms are ( I L ( θ ) ) δ δ δ 2 = α 2 μ e 2 α - e α ( I U ( θ ) ) δ δ δ 2 = α 2 e - α μ ( I slam ( θ ) ) δ δ δ 2 = α 2 μ e α - 1 , ( 12 ) see Eqs 50 , section 3 . 5 in Extended Methods . For labeling times much shorter than the characteristic degradation time of a given gene , α ≪ 1 , the normalized FIM terms behave as a power function: ( I slam ( θ ) ) δ δ δ 2 , ( I L ( θ ) ) δ δ δ 2 ∼ α , ( I U ( θ ) ) δ δ δ 2 ∼ α 2 . ( 13 ) However , for labeling times much longer than the characteristic time of degradation τ , α ≫ 1 , the normalized FIM terms vanish exponentially: ( I L ( θ ) ) δ δ δ 2 ∼ e - 2 α , ( I slam ( θ ) ) δ δ δ 2 , ( I U ( θ ) ) δ δ δ 2 ∼ e - α , ( 14 ) see derivations in Extended Methods , section 3 . 5 , Eqs 51 and 52 . In a typical high-throughput experiment , the kinetic parameters are monitored for a large set of genes ( in the order of thousands ) , which may have different degradation rates . In this case , every time point in the experiment will be only optimal for a subset of these genes . To illustrate this effect , we simulated read counts for an ideal SLAMseq experiment ( with no overdispersion ) and fitted the model using various sets of samples . In our in silico experiment , we always included the total fraction ( t = 0 hr ) , and either one additional time point ( labeled and unlabeled fractions ) or all time points ( 2 , 4 , and 8 hr ) . The normalization coefficients were set to 1 to mimic an ideal SLAMseq scheme , as discussed earlier , Eq 2 . We fitted the model using the pulseR package and computed the 95% confidence intervals ( CI ) for δ using the profile likelihood approach [33] . Since we assume no overdispersion ( Poisson distribution ) , for high read counts ( μ = 10000 ) the quadratic approximation of the log-likelihood function applies , and the confidence intervals for the rate estimations may be approximated by the Wald intervals , i . e . ( δ ^ − 1 . 96 ( I − 1 ( θ ) ) δ δ , δ ^ + 1 . 96 ( I − 1 ( θ ) ) δ δ ) , and hence , they reflect the behavior of the FIM term for δ . As expected , the relative CI width is minimal only for a certain subset of the rates , depending on the set of measurements included , see ( Fig 2B ) . If the degradation rate is very fast in comparison to the experiment time scale , the CI width for these fast genes is defined by the earliest time point in the experiment ( see Fig 2B ) . Since every labeling time is optimal only for a single degradation rate , it might be beneficial to focus the design on genes with faster rates δ , if sample size is limited and no other criteria of optimality are given . The justification follows from the faster decay of the FIM term for α ≫ 1 ( i . e . genes with faster kinetics ) , Eqs 13 and 14 . Read count data from RNA-seq experiments exhibit overdispersion ( variance > mean ) , and the negative binomial distribution ( NB ) is the model of choice to account for that [25] . In this section , we explore how overdispersion would affect MLE of δ . The overdispersion parameter k of the NB distribution describes the level of overdispersion in the data , in which case the variance is defined as var ( X ) = m + m2/k for counts X ∼ NB ( m , k ) with mean m . Smaller values of k correspond to higher overdispersion level , and , for k → ∞ , the NB distribution converges to the Poisson distribution , for which var ( X ) = m . For simplicity , we assume that distributions of read counts in all samples share the same value of k . In addition , we do not consider uncertainty in the overdispersion parameter k when we make inference about δ for individual genes , in a way as it is implemented in some packages for differential expression analysis , for example , in DESeq , [25] . A more advanced quasi-likelihood approach , which accounts for uncertainty in the overdispersion parameter , is discussed in [34] . In the case of NB distribution , the FIM is not diagonal for the SLAMseq experiment , see Eqs 29 and 30 in section 3 of the Extended Methods . Hence we need to work with the inverse FIM , and the diagonal term for the SLAMseq design is ( I slam - 1 ( θ ) ) δ δ = e δ t - 1 μ t 2 + 2 ( 1 - e - δ t ) 2 k t 2 . ( 15 ) The presence of overdispersion shifts the optimal time to higher values . But the most important change is that the profile of I - 1 ( θ ) δ δ is more sensitive to the labeling time t near the optimal point . For higher overdispersion values , the variance of the rate estimator δ ^ increases faster in the vicinity of the optimum ( see Fig 2C ) . This imposes stricter conditions on the experimental design . The second term in the Eq 15 vanishes for times t ≫ 1 , and the equation coincides with the case of no overdispersion . The contribution of the second term is higher for smaller values of k ( higher overdispersion ) and for shorter labeling times t , with the maximal value at t → 0: lim t → 0 2 ( 1 - e - δ t ) 2 k t 2 = 2 δ 2 k . ( 16 ) Another limitation , which arises in the over-dispersed model is that an increase of the sequencing depth has a limited effect on the variance . Indeed , only the first term in Eq 15 can be eliminated by an increase of sequencing depth: lim μ → ∞ ( I slam - 1 ( θ ) ) δ δ = 2 ( 1 - e - δ t ) 2 k t 2 . ( 17 ) In contrast , repeating the experiment n times affects both terms in I δ δ - 1 ( θ ) , since for n repetitions , I - 1 ( θ ) = 1 n I 1 - 1 ( θ ) , ( 18 ) where I 1 - 1 ( θ ) is the inverse FIM for one repetition . In the Poissonian case , when k → ∞ and the second term is absent ( see Eq 9 ) , doubling the number of samples or increasing the sequencing depth by two fold results to the same FIM and , consequently , the same approximation of the variance var ( δ ^ ) . Standard deviation of the rate estimate is a linear function of the depth μ on the logarithmic scale and is not bounded below ( Fig 2D , dashed line ) . In contrast , due to Eq 17 , presence of overdispersion imposes a limit , which can not be overcome by arbitrary high sequencing depth ( Fig 2D , solid line with the horizontal asymptote ) . In essence , spreading the sequencing capacity between several biological replicates can be more beneficial than increasing the sequencing depth on a smaller number of samples . A similar phenomenon is discussed by [35] in the context of differential gene expression analysis by RNA-seq . If one is interested in estimating the rates of extreme values by using very short ( e . g . TT-seq , [36] ) or long labeling times , it may be less efficient to use the protocols , which preserve the ratio of labeled and unlabeled molecules ( e . g . SLAMseq ) . Let us consider a study of fast gene kinetics , where very short labeling times are used . In this case , δt ≪ 1 for the majority of the genes , the labeled fraction constitutes only a minor proportion of the input SLAMseq sample , because mL ( t ) = μ ( 1 − e−δt ) ≈ μδt ≪ 1 . After a short labeling time , any SLAMseq sample mainly consists of unlabeled molecules from genes with slower synthesis , which leads to spending sequencing resources on mostly non-informative material . The same idea holds for very long times , when δt ≫ 1 and when most of the unlabeled molecules were already degraded , mU ( t ) = μe−δt ≪ 1 . In contrast , conventional experimental setups with a separation step can be used to focus sequencing capacity on the relevant molecules . However , the conventional approach suffers from the need to normalize sequencing results from different fractions as it does not preserve the ratio of labeled and unlabeled molecules as defined by the input sample . In typical RNA-seq experiments , the normalization coefficients are assumed to be shared between all the genes in a given sample [25] , but nevertheless , it introduces additional uncertainty into rate estimations . As previously mentioned , a whole range of normalization approaches has been discussed in literature [26] . In the following derivations , we neglect the uncertainty in estimating the fraction normalization coefficients xi from Eq 2 . To illustrate the benefit of the conventional approach , let us consider a set of fast turned over genes F , such that there exists labeling time t , when the majority of genes i∉F do not contribute to the labeled fractions , i . e . μ ( 1 − e−δi t ) ≪ 1 for i∉F , but μ ( 1 − e−δit ) ≈ 1 for i ∈ F . If the sequencing depth of the labeled fraction is approximately the same as for the total sample , then the normalization factor is x L = ∑ i μ i ∑ i μ i ( 1 - e - δ i t ) ≈ ∑ i μ i ∑ i ∈ F μ i , ( 19 ) which can be high at short times . Such “zooming” effect can be considered as corresponding increase of the sequencing depth in SLAMseq experiments by the factor of xL for the labeled fraction . The same idea can be applied to the unlabeled fraction and long labeling times , when the sequencing depth is shared out between the most stable set of genes . Since the normalization factor depends on the rate distribution and the expression level in a given system , it is not possible to derive the optimal design criteria analytically without imposing additional assumptions . As in the case of SLAMseq , inference can be improved to a limited extent by increase of sequencing depth , if overdispersion is present in the data , compare to Eq 17: lim μ → ∞ ( I L ( θ ) ) δ δ = t 2 e - 2 δ t k ( 1 - e - δ t ) 2 ⩽ k δ 2 lim μ → ∞ ( I U ( θ ) ) δ δ = t 2 k ( 20 ) For derivations , see Eqs 58 and 59 in section 4 of Extended Methods . It is interesting to note , that for the case of the unlabeled fraction , the bound can be improved by use of longer labeling times ( provided very high sequencing depth ) , which is not the case for the labeled fraction ( with the upper bound I L ( θ ) → k / δ 2 at t → 0 ) . In summary , biochemical separation should be considered for estimation of degradation rates of RNA species with extreme values . Another design choice is to reduce the number of sequencing reactions by using external spike-ins . For slowly turned over RNA species , one may sequence total and unlabeled fractions , and , for fast turned over RNA species , the total and the labeled fractions . The use of external spike-ins ensures identifiability of the normalizing coefficient from only two fractions . In this section , we consider a published SLAMseq pulse-chase experiment from [12] . Here , mESCs were treated for 24 hrs with 100 μM 4sU ( pulse phase ) with samples being collected after 0 , 0 . 5 , 1 , 3 , 6 , 12 and 24 hr of label chase , and subjected to QuantSeq mRNA 3’ end sequencing . While inspecting the data , we noticed that not all the molecules were fully labeled ( i . e . not all reads show T → C conversions ) after a 24hr pulse phase . In this case , the labeled fraction does not reach the total level μ . We adapted our pulse-chase model to reflect this by introducing a parameter describing the background level of the unlabeled fraction μ1 and the maximal level of the labeled fraction μ2 , so μ1 + μ2 = μ: m U ( t ) = μ 1 + μ 2 ( 1 - e - δ t ) m L ( t ) = μ 2 e - δ t . ( 21 ) The equations for the pulse-only experiment and derivations of other results from this section are described in section 3 . 3 of the Extended methods . Inefficient nucleotide conversion or too short pulse times may result in high values for the background level μ1 ≫ μ2 . In this case , the changes due to RNA kinetics , which are proportional to μ2 , constitute only a small part of the read counts , Eq 21 . In the extreme case of μ1/μ2 → ∞ , the unlabeled fraction does not contribute to the FIM term , lim μ 1 / μ 2 → ∞ ( I U ( θ ) ) δ δ = 0 , since it provides information solely on the nuisance parameter for the background μ1 , see Eqs 45 and 47 in Extended Methods . Moreover , if the sequencing depth is fixed to μ = μ1 + μ2 , the amount of labeled molecules is small , μ2 → 0 as μ1/μ2 → ∞ and , hence , ( I L ( θ ) ) δ δ → 0 , see Eq 46 in Extended Methods . It results in high variance of the rate estimate δ ^ , because var ( δ ^ ) = ( I slam - 1 ( θ ) ) δ δ ⩾ 1 / ( I slam ( θ ) ) δ δ , but ( I slam ( θ ) ) δ δ → 0 . Consequently , the sequencing capacity is spent for measuring the background level . Using the inverse FIM to approximate var ( δ ^ ) = ( I - 1 ( θ ) ) δ δ would result in a rather cumbersome expression . But even with our simplified approach , it is possible to see , how inefficient conversion may be detrimental for estimation of δ and no design optimization with respect to time of chase-phase could recover the situation . To illustrate , how the choice of time point affects the confidence intervals of the estimations , we analyzed different subsets of samples from [12] . Since the model includes one more parameter to take the background level into account ( μ1 ) , one needs to use at least two different time points . In our example , we use combinations of different chase-times and always include t = 0 , because these samples directly provide the information on the μ1 and μ2 , since mU ( 0 ) = μ1 and mL ( 0 ) = μ2 . As expected , for a short ( relative to the characteristic time ) chase phase , subtle changes in the levels of the labeled and unlabeled molecules are masked by the noise and the majority of the degradation rates are not identifiable ( Fig 3 , sample sets for [0 , 0 . 5] hr , [0 , 3] hr ) . Using one more early time point ( [0 , 0 . 5 , 1] hr ) did not substantially improve the estimates , S1 Fig . At longer chase phases , the confidence intervals are more narrow ( [0 , 6] hr ) , and for longer time the estimations for fast genes become worse , since most of their labeled RNA molecules are already degraded ( [0 , 12] hr , right side of the x-axis ) . To illustrate , how the FIM term for a single sample depends on the time of the chase phase , we calculated ( I slam ( θ ) ) δ δ δ ^ 2 for a range of different values of t/τ ratio . ( I slam ( θ ) ) δ δ δ ^ 2 depends on other parameters as well ( see Eq . 48 in section 3 . 3 of the Extended Methods ) . In this example , we used parameter values from the model fitted to the full data set , i . e . including 0 , 0 . 5 , 1 , 3 , 6 , 12 and 24 hr chase time points ( overdispersion parameter k ^ = 10 . 4 and medians of μ ^ 1 and μ ^ 2 , 251 and 89 correspondingly ) . Similar to the simpler case in Fig 2A , there is an optimal time , where ( I slam ( θ ) ) δ δ δ ^ 2 is maximal , t ≈ 2 . 9τ ( Fig 4A ) . Genes with a characteristic time τ , which diverge from t/2 . 9 , will have confidence intervals with a large relative width , and , vice versa , the relative interval width will be more narrow for the genes with τ ≈ t/2 . 9 . To be in line with our estimation for a gene with the median μ ^ 2 , we plotted the genes with μ ^ 2 located around the median ( in 40-60% percentile ) for illustration . For 6 and 12 hr points , there is a distinct minimum in relative confidence intervals at 6/2 . 9 ≈ 2 hr and 12/2 . 9 ≈ 4 hr ( Fig 4B ) . The median of the characteristic time estimates is τ ^ = 5 . 4 hr , and the optimal chase time for such “median” gene would be around 15 hr . In agreement with this observation , the degradation rate estimates , calculated using [0 , 12] hr and [0 , 24] hr points , have the highest correlation to the rates , which were derived from the full data set ( S2 Fig ) . Although in majority of cases several different time points are used , the results of this section show that too long or too short times barely contribute to the estimations . Another factor , which influences the quality , is efficiency of the labeling protocol . The presence of non-informative background RNA creates additional noise to the measurements and wastes sequencing capacity . MCF-7 cells were pulse labeled with 200 μM 4sU for 2 , 4 or 8 hrs . 4sU-labeled and unlabeled RNA were separated by streptavidin purification after MTSEA biotin-XX catalyzed biotinylation of 4sU-labeled RNA , which has an efficiency of 95% [18] . The efficiency of purification was monitored in a dot blot assay that detects biotinylated RNA with streptavidin-HRP ( Fig 5A , S3 Fig ) . This analysis revealed a gradual increase in biotinylation with increasing labeling time . Importantly , biotinylated transcripts were efficiently depleted from the flow through . No biotinylation signal could be detected in these samples , which illustrates the high efficiency of the streptavidin purification . Biotin-enriched RNAs are eluted by three rounds of de-biotinylation with DTT . Therefore , we estimated the purification efficiency by the amount of purified RNA determined by A260nm absorption measurement . The amount of purified RNA increased gradually with increasing labeling time ( Fig 5B ) comparable to the biotinylation signal increase in the respective input fractions ( Fig 5A ) . To determine the efficiency and specificity precisely for individual transcripts , we spiked the 4sU-labeled total RNA from MCF-7 with in vitro transcribed 4sU-labeled FLuc and unlabeled RLuc that were followed by RT-qPCR analysis using a standard curve for quantification ( S3 Fig ) . This analysis revealed a purification efficiency of 4sU-labeled FLuc of about 60% ( 58 . 56 ) . The specificity was determined by the cross-contamination of RLuc in the biotin-enriched fractions and FLuc in the flow through fractions , which was about 5% for each transcript ( RLuc in enriched = 5 . 32% , FLuc in flow through = 5 . 01% , see Fig 5C ) . The kinetic model was fitted to the read counts from the sequenced samples for genes with mean read count >50 in the total samples . Two total samples were collected at 0 hr , labeled and unlabeled fractions at other time points ( 2 , 4 and 8 hrs ) in two replicates ( see S2 Table ) . In the model fitting , we assumed no cross-contamination between fractions and shared normalization coefficients for samples originating from the same time point and fraction . Having the estimations for expression levels , degradation rates , overdispersion parameter and normalization coefficients , we calculated the FIM diagonal elements I δ δ ( θ ) for the analyzed genes for different time points and fraction types . In Fig 6A , the value of the diagonal FIM element multiplied by δ ^ 2 , i . e . I δ δ ( θ ) δ ^ 2 ( compare to Eqs 11 and 12 ) , is depicted for both fractions . As mentioned in the previous section , I δ δ ( θ ) can be interpreted as an information gain from the experiment assuming other parameters were known , which represents an upper bound , see Eq 8 . In addition , these terms are bounded due to presence of overdispersion in the data , ( Eq 20 and dashed lines in Fig 6A ) , and increase of sequencing depth can not improve these limits . At short labeling times , the FIM term is higher for the labeled fraction than for the unlabeled one for majority of the genes , ( Fig 6A , 2hr ) , which is a result similar to the SLAMseq case . At longer labeling times , the contribution from the unlabeled fraction increases , and ( I U ( θ ) ) δ δ > ( I L ( θ ) ) δ δ for majority of the genes ( Fig 6A , 8hr ) . However , the proportion of RNA amount from genes with high degradation rates δ in the unlabeled sample exponentially decreases , since lim t → ∞ μ fast e - δ fast t μ slow e - δ slow t = lim t → ∞ μ fast μ slow e - ( δ fast - δ slow ) t = 0 . ( 22 ) It results in very low counts and decrease in the I U ( θ ) for these fast genes , see Fig 6A , 8hr , reduced values at the right tail of the distribution ( blue dots ) . The optimal design for such experiments is complicated by the fact that it depends not only on the degradation rates of some target genes , but on the overall rate distribution in the system being studied . We illustrate a dependency of the ( I ( θ ) ) δ δ δ 2 terms on labeling time for one of the fastest ( 0 . 1% quantile ) and one of the slowest ( 99 . 9% quantile ) genes . The normalization coefficients for the labeled and unlabeled fractions were adjusted in such a way , that at every time point t the sequencing depth equals the sequencing depth of the total sample . In the case of low or no overdispersion , use of labeled fraction and shorter labeling times is preferred for estimation of fast genes , because ( I L ( θ ) ) δ δ > ( I U ( θ ) ) δ δ , see Fig 6B , dashed red line over the blue line . For slow genes , one may benefit from use of unlabeled fraction , since the highest FIM values correspond to ( I U ( θ ) ) δ δ at longer times , see Fig 6B , dashed blue line over red line . At high values of overdispersion ( i . e . low k ) , the FIM term is bounded ( I L ( θ ) ) δ δ δ 2 < k due to Eq 20 . In this case , there may exist values of labeling times at which the terms from the unlabeled fraction ( I U ( θ ) ) δ δ δ 2 is larger than maximal ( I L ( θ ) ) δ δ δ 2 value , Fig 6B , solid lines . As a protection against such situation in the case of fast genes , use of samples from unlabeled fraction may be a solution . Although one may have a prior guess about the range of degradation rates in a system , it is unlikely that there is information about the distribution of the rates and overdispersion level . Hence , such design suggestions are possible only in sequential approach , when an exploratory experiment is done first . It is important to note the “zooming” effect of the conventional design , which we discussed in the previous section Biochemical separation still matters . At a short labeling time , the term ( I L ( θ ) ) δ δ δ 2 decreases as t approches zero in the case of the SLAMseq design , Fig 2A , the red line . In contrast , due to higher sequencing depth of individual fractions in the conventional setting , ( I L ( θ ) ) δ δ δ 2 has a horizontal asymptote , Fig 6B , red lines .
In this study , we discuss some aspects of the optimal design of RNA labeling experiments using the results of the asymptotic theory . First , we show that there exists an optimal time point for which the maximum likelihood estimator possess a minimal variance asymptotically . This first result was developed for the case of experiments , which preserve the fraction ratio and hence do not require normalization between fractions ( e . g . SLAMseq , TUC-seq , TimeLapse-seq ) . In the case of negligible overdispersion , the optimal labeling time for a gene with the characteristic degradation time τ is tslam = 1 . 59τ , and shorter labeling times show better rate estimates in comparison to longer times: the variance increases exponentially for times longer than τ and only by a power law for shorter labeling times . This result is similar to the observations in a simulation study by [32] . Herein , for a given gene with a half-life λ = 2 hr , the most precise estimation were at labeling times 3 hr and 6 hr ( toptimal = 1 . 59 ⋅ 2/log ( 2 ) = 4 . 6 hr ) , and the worst estimations were observed at the longest and the shortest times ( 12 hr and 0 . 5 hr ) . However , the exact ranking of time points is different for the given half-life time , probably due to the influence of prior distribution utilized in the Bayesian framework . We show that at short labeling times ( in comparison to the characteristic time of degradation for a given gene ) , the labeled fraction contributes most to the Fisher information term corresponding to the degradation rate , and , vice versa , at long times the highest contribution is seen for the unlabeled one . In addition , we show that in the presence of overdispersion , the variance of rate estimates is more sensitive to choices of labeling times different from the optimal , which make it more difficult to optimize conditions for a range of rates . The overdispersion imposes a bound on the asymptotic relative standard deviation for the estimator of the rate ( sd ( δ ) /δ , see Fig 2C ) , and , from a certain level , increase in sequencing depth is very inefficient ( Fig 2D ) . We present similar results for SLAMseq data from a published pulse-chase experiment . Herein , we extended our model to reflect incomplete labeling and demonstrate that every chase time is optimal only for genes with a certain ratio of the characteristic degradation time and the chase time ( tslam ≈ 2 . 9τ , see Fig 4 ) . Moreover , we discuss possible benefits of use of the conventional experimental approach , especially for estimation of extreme degradation rates , which deviate highly from the general pool . For nucleotide conversion setups with too short or too long labeling times , the majority of reads in a sample originate from the unlabeled or labeled fractions correspondingly . In contrast , the conventional scheme , which involves biochemical fraction separation , allows to concentrate the experimental costs only on the relevant material . This approach strongly relies on normalization between the samples , as the fraction ratio is not preserved . Besides the use of labeled and unlabeled spike ins additional normalization strategies have been developed to ensure this , see [26] . Obviously , there are certain limitations to our study . First , the method involving FIM calculation describes only the asymptotic behavior of the estimator . Hence , all the conclusions are only approximate , since we do not investigate the behavior of the likelihood function itself , but only the quadratic approximation of its logarithm using the FIM . Secondly , we do not consider uncertainty from the shared parameters , such as the overdispersion parameter of the negative binomial distribution and the normalization coefficients for the fractions . Inference on these parameters is based on the whole pool of the genes , and would involve more complex analytic treatment and assumptions on the distribution of rates . Thirdly , this study is concerned with the statistical aspects , rather than kinetic modeling , and the simplest model of synthesis and degradation is used . More complex models , which describe biochemical networks or RNA maturation can be more relevant depending on the research question . Other phenomena , like dilution due to cell division , may have an effect on the RNA level as well and should be taken into account in the case of the long-lived transcripts [26] . Lastly , cross-contamination between fractions is a highly relevant problem for inference , especially in the absence of external reference molecules ( spike ins ) , which are typically used to assess this phenomenon . However , in section 2 . 1 of the Extended methods , we show that cross-contamination shifts estimations of fast rates to slower values , and slow rates towards faster values . Previously , [28] included a global transcriptome-wide cross-contamination term to presented kinetic model , yet future work is needed to assess possible effect sizes on rate estimations . With regards to our own experimental results , we used unlabeled RLuc and 4sU-labeled FLuc to control the efficiency and specificity of biochemical separation . We reckon that the recovery of only 65% 4sU-labeled FLuc may be caused by inefficient elution or loss during the washing steps . RNA species with a high 4sU content are more likely to be affected by inefficient elution , whereas the loss during the washing steps may be observed for RNAs with very few 4sU . These effects will also introduce a bias in rate estimates , which originate from the biotin-enriched fraction . We hope that our work will encourage further development of the methodology to address the discussed limitations and to improve suggestions on design of metabolic labeling experiments . | Massively parallel RNA sequencing ( RNA-seq ) in combination with metabolic labeling has become the de facto standard approach to study alterations in RNA transcription , processing or decay . In our manuscript , we address several key aspects of experimental design: 1 ) The optimal labeling time , 2 ) the number of replicate samples over sequencing depth and 3 ) the choice of experimental protocol . We provide approximate answers to these questions using asymptotic theory of optimal design . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
]
| [
"sequencing",
"techniques",
"nucleic",
"acid",
"synthesis",
"experimental",
"design",
"rna",
"extraction",
"nucleotides",
"research",
"design",
"molecular",
"biology",
"techniques",
"rna",
"synthesis",
"rna",
"sequencing",
"extraction",
"techniques",
"rna",
"transcription",
"labeling",
"chemical",
"synthesis",
"research",
"and",
"analysis",
"methods",
"cell",
"labeling",
"proteins",
"molecular",
"biology",
"biotinylation",
"biosynthetic",
"techniques",
"metabolic",
"labeling",
"biochemistry",
"rna",
"post-translational",
"modification",
"nucleic",
"acids",
"biology",
"and",
"life",
"sciences",
"nucleic",
"acid",
"labeling"
]
| 2019 | On the optimal design of metabolic RNA labeling experiments |
Continued exposure to malaria-causing parasites in endemic regions of malaria induces significant levels of acquired immunity in adult individuals . A better understanding of the transcriptional basis for this acquired immunological response may provide insight into how the immune system can be boosted during vaccination , and into why infected individuals differ in symptomology . Peripheral blood gene expression profiles of 9 semi-immune volunteers from a Plasmodium vivax malaria prevalent region ( Buenaventura , Colombia ) were compared to those of 7 naïve individuals from a region with no reported transmission of malaria ( Cali , Colombia ) after a controlled infection mosquito bite challenge with P . vivax . A Fluidigm nanoscale quantitative RT-PCR array was used to survey altered expression of 96 blood informative transcripts at 7 timepoints after controlled infection , and RNASeq was used to contrast pre-infection and early parasitemia timepoints . There was no evidence for transcriptional changes prior to the appearance of blood stage parasites at day 12 or 13 , at which time there was a strong interferon response and , unexpectedly , down-regulation of transcripts related to inflammation and innate immunity . This differential expression was confirmed with RNASeq , which also suggested perturbations of aspects of T cell function and erythropoiesis . Despite differences in clinical symptoms between the semi-immune and malaria naïve individuals , only subtle differences in their transcriptomes were observed , although 175 genes showed significantly greater induction or repression in the naïve volunteers from Cali . Gene expression profiling of whole blood reveals the type and duration of the immune response to P . vivax infection , and highlights a subset of genes that may mediate adaptive immunity .
One of the features of Plasmodium species that make them such pernicious parasites is their ability to avoid the host immune system [1 , 2] . While this is achieved in part by virtue of their complex life cycle that includes intra-erythrocyte cycling and periodic sequestration in various tissue compartments [2] , it is also clear that Plasmodium infection causes short- and probably long-term modification of host immune function . Molecular methods are shedding some light on the mechanisms behind these modifications . For example , it is now clear that exposed individuals generally do mount an antigen response to Plasmodium antigens that persists [3 , 4] , and that several biochemical pathways are engaged , including interferon and cytokine signaling , membrane lipid modification , and reactive oxygen species metabolism [5] . Host factors including genetic variation , both within and between populations , play a role in modulating immunity in malaria , as does the microbiome [6–9] . An important factor influencing the clinical course of disease is prior exposure to malaria . Adults and older children tend to experience reduced prevalence of malaria infection and have less severe symptoms [10 , 11] . Nevertheless the mechanisms responsible for host resistance to malaria are still poorly understood . As a prelude to evaluation of vaccine efficacy in a Colombian population , we recently carried out a challenge experiment in which we described the responses of immunologically naïve and semi-immune individuals to deliberate infection with Plasmodium vivax through mosquito bites [12] . All nine volunteers from a malaria endemic region near the town of Buenaventura were weakly positive for IgG antibodies to sporozoites or blood stage proteins prior to the experiment , and after challenge eight of them showed increased antibody titers against blood stages . Similarly , five of seven naïve volunteers from the city of Cali converted to sero-positivity that was generally maintained for at least four months . While there was no significant difference in the time to first appearance of blood stage parasite assessed by thick blood smears ( 12 to 13 days in both groups ) or by polymerase chain reaction ( PCR ) ( around 9 days ) , the naïve volunteers experienced classical early malaria symptoms , whereas the semi-immune volunteers were for the most part nearly asymptomatic , at least at the day of diagnosis when curative prophylaxis was administered [12] . In order to begin to characterize the molecular basis for this difference in clinical course of disease as a function of prior exposure to malaria , we report here two types of transcriptome profiling of peripheral blood samples from the Colombian challenge experiment volunteers . First we used targeted measurement of a set of 96 highly informative transcripts by nanoscale Real Time PCR ( RT-qPCR ) [13] in order to generate a time course of the infection transcriptional response . Second , we used RNASeq [14] on a subset of six volunteers contrasting baseline and incident malaria , to ask whether ( i ) there is a difference in immune profiles between naïve and semi-immune individuals in the absence of infection , and ( ii ) patent infection results in a differential transcriptional response that may hint at the molecular basis of long-term immunity . We also contrasted our findings with those of cross-sectional studies , concluding that history of exposure is just one of many factors mediating host–parasite interactions in malaria .
The experimental design protocol of this research was approved by the Institutional Review Boards ( IRB ) at the Malaria Vaccine and Drug Development Center ( CECIV , Cali ) and Centro Médico Imbanaco ( Cali ) . Volunteers were adults and were extensively informed about the risks of participation . Before signing the written consent , all volunteers had to pass an oral or written exam related to the trial and its risks . Clinical trial was registered under registry number NCT01585077 . It is described in more detail in Arévalo-Herrera et al . [12] , which reports the clinical responses to malaria challenge . Sixteen Duffy-positive ( Fy+ ) male and female volunteers ( 9 semi-immune , previously exposed to malaria , from Buenaventura and 7 immunologically naïve with respect to malaria , from Cali; 10 men and 6 women ) were enrolled . Volunteers where invited to the vaccine center two days ( day -2 ) prior the challenge day ( day 0 ) for physical examination and blood sample collection . Fig 1 summarizes the blood sampling strategy . Blood samples used for the RT-qPCR experiment were collected on day -2 ( pre-challenge ) , day 5 , day 7 , day 9 , on the day of first detection of Plasmodium by thick smear test ( day 12–13 ) , and on month 4 . RNASeq analysis , also approved by the Georgia Tech IRB , was performed for 12 individuals ( six each from Buenaventura and Cali ) for two of the timepoints , namely the diagnosis day and baseline ( pre-challenge day ) . For each sample , approximately 1 mL of blood in 2 mL of buffer was collected into a Tempus tube , which preserves whole blood RNA at 4°C indefinitely . Whole blood mRNA was extracted using Tempus Blood RNA Tube isolation kits provided by the manufacturer Applied Biosystems , and the sample quality was determined based on the Agilent Bioanalyzer 2100 RNA Integrity score ( RIN ) . Two samples had RIN approximatley 4 but these were not outliers in the analysis and there was no indication that RNA degradation influenced the results meaningfully . Reverse Transcription followed by quantitative PCR ( RT-qPCR ) was performed using Fluidigm 96×96 nanofluidic arrays targeting a set of 96 transcripts that are broadly informative of the major axes of variation for peripheral blood gene expression from Preininger et al . [15] at six timepoints ( Pre-challenge , day 5 , day 7 , day 9 , Diagnosis ( Dx ) and month 4 ) . The RT-qPCR was completed in three steps: ( 1 ) Total whole blood RNA was converted to single stranded cDNA using polyT priming of reverse transcription , ( 2 ) the 96 targeted genes were pre-amplified in a single 13-cycle PCR reaction for each sample following conditions outlined in the manufacturer’s protocol by combining cDNA with the pooled primers and EvaGreen Mastermix ( Fluidigm BioMark ) , and ( 3 ) qPCR reactions were performed for each sample and individual gene on each sample on a 96×96 array with 30 amplification cycles . Average Ct value was calculated at a point in which every reaction is in the exponential phase to ensure accuracy and precision of amplification . In order to make the analysis more easily comparable with traditional transcript abundance measures such as those obtained with microarrays or RNASeq , each Ct value was subtracted from 30 , setting missing values to 0 . Since small Ct values correspond to high transcript abundance , this subtraction yields values ranging from 0 ( no expression ) to 30 ( very high abundance ) . All measurements are reported in S1 Table . Library preparation for RNASeq was performed using the Illumina TruSeq Low Throughput ( LT ) RNA Sample Preparation Protocol . Short read sequencing was performed in rapid run mode with eight samples per lane on an Illumina HiSeq 2500 , generating 100 bp paired-end libraries with an average of 15 million paired reads per sample , and then sequencing on an Illumina HiSeq2100 at Georgia Institute of Technology . The raw RNASeq reads ( Fastq files ) for each sample were tested using FastQC software analysis to check the quality of the data ( S2 Table ) and then aligned to the reference human genome ( hg19 / GRCh37 assembly ) using Bowtie as the short read aligner , and splice junctions were identified using TopHat2 in the Tuxedo protocol [16] . After alignment , estimation of transcript abundance measures as fragments per kilobase of exon per million aligned fragments ( FPKM ) values was performed using Cufflinks [16] . As a quality control for high variance associated with low abundance , genes with FPKM greater than 2 . 5 averaged across the 24 samples were retained for downstream analyses , representing 6 , 154 genes . FPKM values were then transformed to logarithm base 2 to guarantee that the data were more normally distributed and to simplify the interpretation of the scale of differential expression ( each unit difference corresponds to a two-fold difference in abundance ) . The supervised normalization of microarray ( SNM ) procedure was then used to normalize the data with the R package from Bioconductor [17] , fitting location and timepoint as the biological variables , and individual as the adjustment variable ( fit but not removed ) . All downstream analyses were performed on this normalized data set provided as S3 Table . Many gene expression profiling experiments start with analysis of Principal Components , but since these are study specific , we also performed an analysis focused on large sets of genes that have been found to consistently covary in peripheral blood , namely blood informative transcript ( BIT ) axes analysis [15] . This analysis focuses on 9 common axes of variation that we detected in all human peripheral blood gene expression datasets that we have examined . They are related to 28 modules of co-expressed genes described by Chaussabel and colleagues [18] and found to be dysregulated in various immune diseases , but which collapse into 9 larger Axes of transcripts that covary in healthy adults . Each axis includes from several hundred to several thousand genes that gene set enrichment analysis suggests are involved in particular immune functions , broadly speaking , T cell signaling ( Axis 1 ) , reticulocyte number ( Axis 2 ) , B cell signaling ( Axis 3 ) , and inflammation ( Axis 5 ) or specific immune or physiological responses ( Interferon signaling , Axis 7 ) . Blood informative transcript ( BIT ) axes analysis was performed by generating the first PC for the 10 genes that are most strongly correlated with each of the 9 Axes reported in [15] . Principal component one ( PC1 ) for each of these 10 sets of BIT provide a summary axis score , to which all of the other genes in the Axis positively correlate , nominally with a Pearson correlation coefficient greater than 0 . 5 as listed in S4 Table . This Axis score is then contrasted with respect to the covariates of interest ( primarily time-point , location , and the interaction between them , but in exploratory analyses gender , parasitemia , and individual ) using standard parametric t-tests or analysis of variance , or with linear regression . Most statistical analyses of both the Fluidigm and RNASeq datasets were performed in JMP Genomics version 5 ( SAS Institute , NC ) , starting with the Basic Expression Workflow , which performs principal components analysis ( PCA ) , and computes a weighted total contribution of the covariates of interest to the axes ( principal variance components analysis , PVCA ) . Linear regression was then used to assess the relationship between the individual covariates and PC , and/or analysis of variance was used to detect differential expression between locations or timepoints . We fit models with timepoint and location as fixed effects , and with individual as a random effect , in order to control for the effect of differential responses among individuals within each sample . No differences in the significance of the fixed effects were observed compared to models without individual , implying that this source of variation is minimal , and we report the significance of the full model . A Benjamini-Hochberg 5% false discovery rate was used to select differentially expressed genes . Volcano plots contrast the significance ( negative log10 of the p-value , NLP ) against the fold difference ( normalized log2 Ct or FPKM units ) between specific conditions . Hierarchical clustering was performed using Ward’s method , and approximately unbiased boostrap support AU values were computed with the R program pvclust [19] . For the RNASeq there were 6 individuals of each gender , but they were asymmetric with respect to location , since 5 females were from Buenaventura and 5 males from Cali . Gender did not however account for a significant proportion of the major PC of gene expression . To confirm this , we fit models with gender , time and location for each Axis , and in each case the gender term was non-significant and the other terms were unaffected . In order to identify whether location influences the axes of variation in another study , we reanalyzed data from Idaghdour et al . [9] who characterized whole blood transcriptomes of infants from the West African Republic of Benin , infected with Plasmodium falciparum . They reported on 61 healthy controls from a hospital in the city of Cotonou , and 92 cases drawn approximately equally and without bias with respect to parasitemia levels from Cotonou and the village of Zinvié , located 36 km from Cotonou ( GEO accession number GSE34404 ) . They identified parasitemia as the major factor influencing transcript abundance overall , but also described a location effect that is considered with respect to the BIT axes here . The RNASeq dataset has been deposited into the Gene Expression Omnibus archive ( GEO ) under accession number GSE67184 and RT-qPCR data accession number GSE67470 .
The first objective of this study was to compare the time course of transcriptional changes during response to infection , between naïve and semi-immune volunteers . There were 16 volunteers in all , 7 from Cali who had not previously been exposed to malaria , and 9 from Buenaventura , a village in an endemic region for the disease , all of whom had experienced between 2 and 5 mild bouts of malaria . Fig 1 summarizes the peripheral blood sampling scheme from 14 volunteers at Day 5 following exposure and again at Day 7 , from 16 volunteers at Day 9 when PCR later confirmed initial appearance of blood-stage parasites , from 11 volunteers on Days 12 or 13 when parasitemia was diagnosed in thick blood smears , and from 14 volunteers four months after the initiation of the experiment . As mentioned above , there were no significant differences between the two groups either in the length of the pre-patent period or the level of parasitemia attained before administration of a curative cocktail of anti-malarial drugs . Whole blood gene expression was monitored in each of the 85 samples using a Fluidigm nanoscale RT-qPCR array targeting 96 genes referred as “blood informative transcripts” ( BIT ) ( S5 Table ) . These BIT consistently capture the covariance of over half of the genes expressed in blood , specifically serving as biomarkers for 10 conserved axes of variation . Across all of the gene expression measurements , 30% of the variance was among individuals , and just 6 . 5% between the timepoints , with very little differentiation between the naïve and pre-immune volunteers ( Fig 2 ) . The remainder of the variance was due to random biological or technical noise , or to the covariance of gene expression along the Axes . We confirmed that most of the genes were co-regulated in this dataset by observing a strong correlation of expression for each of the 10 BIT for each Axis , and then generated Axis scores as the first principal component of the variance of those 10 BIT . Only two of the Axes were differentially expressed among timepoints in the Fluidigm data ( Fig 3A and 3B ) . Axis 5 is related to innate immune signaling and neutrophil number , and seems to decline at Diagnosis , surprisingly , implying a mild reduction in inflammatory gene activity . Axis 7 represents Type 1 interferon induction and is , as expected , elevated at diagnosis , reflecting a transient specific immune response . Both axes had returned to close to baseline levels three months after recovery . No other gene expression differences detected by this targeted RT-qPCR analysis were associated with time or population . These results are consistent with previously observed stable maintenance of peripheral blood gene expression profiles in healthy adults . A caveat to this analysis is that it is possible that other genes not included in the targeted set of probes do change in expression prior to the diagnosis of parasitemia , or alternatively do not return to baseline after recovery . In order to obtain a more comprehensive picture of the changes in gene expression as parasites first appear in the blood , we performed RNASeq on 6 volunteers each from Cali and Buenaventura , both at Baseline and Diagnosis . An average of 15 million paired-end 100bp short read alignments to the human reference genome were obtained for each sample , allowing us to estimate transcript abundance for each of 6 , 154 genes . Analysis of variance was used to contrast gene expression relative to population and timepoint , and to assess the interaction between these two factors . Fig 2B shows that 25% of the total variance was among individuals , similar to the Fluidigm observation , and that very little differentiation was seen between populations . However , just over one third of the variance was between Baseline and Diagnosis samples , implying a much greater response to infection than suggested by the RT-qPCR data , though it should be noted that when only contrasting the two most different timepoints , this contrast was expected to account for more of the variance . The differential expression of Axes 5 and 7 was confirmed by the RNASeq data ( Fig 3D and 3E ) , which also suggested divergence of Axis 2 ( Fig 3C ) . Individual variability in response was minimal , as inclusion of individual as a random effect in the models had no effect on the proportion of variance due to either the time-of-diagnosis or population sampled . Up-regulation of Axis 2 is likely to be a sign of elevated erythropoiesis since they are enriched for erythrocyte-related function [15] and reanalysis of the dataset reported by Whitney et al . ( 2003 ) [20] shows that the genes in this axis are highly correlated with reticuloycte count , suggesting a mild physiological response to loss of red blood cell function even in the early stages of malaria . Interestingly , the increased resolution of RNASeq suggests differential responses of Axes 5 and 7 between the naïve and semi-immune populations . Specifically , the neutrophil and TLR-signaling associated with Axis 5 is much weaker in the naïve individuals ( Fig 3D , solid blue points , p = 0 . 0008 , though the location×timepoint interaction term is not significant , p = 0 . 13 ) , whereas the induction of interferon signaling is variable in semi-immune volunteers ( Fig 3D , open red circles ) , two of whom showed no response . The directional trends were the same in the Fluidigm data , but less apparent . Consistent with timepoint rather than population explaining a large proportion of the variance , gene-specific differential expression analysis revealed more than 250 transcripts up- or down-regulated at the experiment-wide threshold of p<10-5 ( Fig 4A ) , but only two transcripts more highly expressed in Buenaventura and none in Cali ( Fig 4B ) . Approximately 50 genes show more than 2-fold up-regulation at Diagnosis relative to Baseline yet are less significant than many of the orange-colored genes ( Fig 4A , green-colored genes ) . The reason is that these genes are even more highly upregulated in a subset of individuals , namely the naïve ( Cali ) volunteers . In fact , 175 genes show a significant timepoint-by-population interaction effect at p<0 . 05 ( ANOVA , Fig 4C; S6 Table ) . These are represented in the heat-map in Fig 4C , showing two-way hierarchical clustering of transcripts in samples , two-thirds of the genes are actually down regulated at Diagnosis day ( red sample labels , top ) . Interestingly , there was a marked distinction between the two timepoints ( Fig 4C ) in the sense that the Baseline samples were intermingled with respect to whether they were from the naïve or semi-immune populations , whereas the Diagnosis ones showed a near-perfect separation with respect to pre-immune exposure ( bootstrap support 78% ) . In other words , most of the genes showing an interaction effect were more strongly up- or down regulated in the naïve than semi-immune individuals . An exception was a Baseline sample from a Cali volunteer ( number 306 ) , which clustered with the Diagnosis set but still showed a robust response to malaria infection along with moderate thrombocytopenia and leukopenia , as did Cali 310 who was not an outlier . ) Given the importance of cytokines to regulation of the immune response , we specifically analyzed the expression of many of the genes in the RNASeq dataset that are related to Interleukin ( IL ) , interferon ( IFN ) , tumor necrosis factor ( TNF ) , and transforming growth factor ( TGF ) signaling . This analysis revealed three groups of samples , and three clusters of genes ( Fig 5 ) . Once again , the Baseline and Diagnosis samples were separated , excluding the outlier Cali 306 Baseline sample and two others , but in this case there was no clear separation relative to pre-infection malaria status . One cluster of 14 genes , including IL32 and IL8 , was not differentially expressed . Another cluster of 23 genes , including the IL4R , IL6R , and IL7R and IL17R receptors , was upregulated at Baseline , particularly strongly in three volunteers ( Cali 302 and Buenaventura 341 and 375 ) . The third cluster of 19 genes , including TNF , IL1B and IL15 , showed the opposite tendency , namely up-regulation at Diagnosis , particularly strongly in two samples ( 314 from Cali and 324 from Buenaventura ) . These results imply that there is strong co-regulation of the cytokine response and infection , but that this is not mediating the differential response between naïve and semi-immune individuals . This is somewhat surprising , especially given that the experience of fever was significantly different between the two populations , who might have been predicted to differ with respect to the pyrogenic cytokines IL1 , IL6 , IL8 and TNF . Closer examination of the differentially expressed genes between Baseline and Diagnosis suggested a complex pattern of cross-regulatory interactions . The up- and down-regulated cytokines for example both include pro- and anti-inflammatory peptides and their receptors . Similarly , there appear to be counter-balancing signal transduction profiles: JAK1 and RAF1 are both strongly down-regulated in all volunteers at Diagnosis , whereas IL6ST and SOS1 are up regulated . Among the 175 genes showing a significant timepoint-by-population interaction effect , namely a stronger response at diagnosis in the immunologically naïve individuals , there are several types of gene functions of interest ( Table 1 ) . These include lysosomal components ( CTSH , RILP ) , regulators of macrophage activity ( CD163 , MMP25 , SIRPA , TBC1D14 , TNFSF13 ) , splicing factors ( EIF2C4 , SNRPB2 , SNRPG ) , lipid biosynthesis ( DGAT2 , LPPR2 ) , solute carriers ( S100P , SLC6A6 , SLC11A1 , SLC7A7 ) , signal transduction ( G3BP1 , GAB3 , MAPK13 , TLE3 ) and Cell Cycle and DNA damage response ( ATM , PRKDC , ARID4A ) . Some genes with an interaction effect showed stronger down-regulation in Cali ( Fig 6A , ATM ) , or stronger down-regulation in Buenaventura ( Fig 6B , EIF2C4 ) , compared with one that showed a similar up-regulation at both locations ( Fig 6C , ATP1B3 ) . Finally , we reanalyzed an infant malarial gene expression dataset from Benin [9] . All samples were collected within a period of 10 weeks in the Spring of 2010 , and transcript abundance data was generated on Illumina HumanHT-12 BeadChips for 155 individuals ( 61 controls from Cotonou , 24 high parasitemia from the village of Zinvie , 52 low parasitemia from Zinvie , and 18 from the city of Cotonou ) . Critical differences relative to our study include ( i ) comparison with P . falciparum rather than with P . vivax infection , ( ii ) infants versus young adults comparison , and ( iii ) cross-sectional rather than Baseline vs Diagnosis analysis . Nevertheless , a significant correlation ( Fig 7A and 7B ) was observed between parasitemia and two Axes of variation , Axes 1 and 5 . However , in this case there was activation of the innate immunity/inflammation genes as parasite burden increases . Axis 1 , which is enriched for T-cell signaling activity [15] , was strongly reduced as parasitemia increased , but like Axis 5 , not significantly affected in the infants with low parasitemia . From 32 genes showing a significant interaction effect between timepoint and population in our challenge experiment , 12 were nominally differentially expressed between malaria patients in the city of Cotonou and rural village of Zinvie in Benin .
The core result of this study was that gene expression was significantly altered at the time of malaria diagnosis , particularly in the immunologically naïve volunteers . Although the targeted expression profiling is less comprehensive and less sensitive than the RNASeq , it suggests that there is minimal transcriptional change in peripheral blood prior to patent infection , and that individual profiles return to baseline within a few months of parasite clearance . No obvious difference in the transcriptomes of uninfected naïve and semi-immune volunteers was seen , but several hundred genes showed a stronger response in the naïve individuals . We cannot however conclude that prior immune exposure is the only reason for this difference as other lifestyle factors that distinguish the inland city of Cali from the oceanside town of Buenaventura , ( where there is likely a larger proportion of African ancestry ) may also play a role . However , the data is strongly suggestive of a long-term modulation of the malaria immune response involving multiple molecular pathways . Some studies have suggested that clinically immune individuals infected with P . vivax show lower levels of inflammatory and regulatory cytokines , than individuals infected with P . falciparum malaria [21] . Nevertheless , the down-regulation of multiple genes related to innate immunity , inflammation , and neutrophil abundance , all correlated with Axis 5 , observed here was unexpected . A large cross-sectional study of infants with malaria conducted in the West African Republic of Benin [9] documented a strong up-regulation of the same genes , although reanalysis of their data shown in Fig 7A suggests that is only true in the presence of high levels of parasitemia . Even more surprisingly , the reduction in inflammatory gene expression was stronger in the naïve than semi-immune volunteers . One possibility is that there is a transient reduction in relative neutrophil counts and inflammatory gene expression as the parasite first appears in the bloodstream , just as the lymphoid cells begin to amplify their response , and this is corrected as parasite levels increase and neutrophilia occurs a few days into the infection [22 , 23] . An observation that is consistent with published data is the strong induction of an interferon response in association with blood-stage malaria [24 , 25] . It is unclear whether this induction was stronger in Cali or Buenaventura , since a couple of the Cali volunteers had unusually high baseline interferon-related gene expression captured by Axis 7 . It does appear that a few of the semi-immune individuals did not mount an interferon response , consistent with the absence of overt clinical symptoms and implying that their immunological memory was able to deal with at least the early stage of infection without mounting the kind of major immunological response observed in the naïve volunteers . This in turn implies that the presence of blood stage parasites alone is not the only determinant of whether or not an individual mounts an interferon response . The overall cytokine profile shifts reported in Fig 5 did not correlate with the clinical profile differences , which could be explained by the host immunity level that can vary due to the acquired immunity throughout repeated exposure [26] . Presumably larger sample sizes and longitudinal profiling during disease will identify associations between gene expression and physiological response , which is also likely to involve other tissues . On the other hand , multiple classes of gene activity do seem to be differentially activated between naïve and semi-immune volunteers . These include various signal transduction molecules , genes related to macrophage activity , and other cellular processes that are known to influence immune responsiveness including lipid synthesis and lysosomal function concordant with Portugal et al . [27] who suggest that as children develop exposure-dependent immunity to P . falciparum , the responses reduce pathogenic inflammation and boost anti-parasite mechanisms . The aforementioned study in Benin again provides a potential comparison , since it included the contrast between children in the city of Cotonou with the rural village of Zinvié . Differences in human peripheral blood gene expression according to lifestyle are prevalent [28] , but it is nevertheless interesting that , of the 32genes showing a significant interaction effect between timepoint and population in our challenge experiment , 12 were nominally differentially expressed between malaria patients from the two locations in Benin , compared with no more than three expected . Fig 7B shows that Axis 1 ( related to T-cell signaling ) is down-regulated with high parasitemia , and consistently reduced in the village of Zinvié . This Axis was not affected in our study , but collectively these observations of context-dependent alterations in gene expression provide further evidence that immune history is an important mediator of the differential clinical profiles observed among individuals . There is also considerable interest in the use of gene expression profiling to identify genes that may mediate robust vaccine responses . Recent study reports on influenza and yellow fever have highlighted individual genes that are required for vaccine effectiveness , but have also suggested that baseline profiles of immune cell types may provide better predictors of antibody production [8 , 29] . Various properties of Plasmodium suggest that this organism may present a more difficult scenario for dissecting the molecular basis of vaccine responses , but we consider the results reported here to be an encouraging baseline establishing that differential responses to a malaria challenge can be detected by gene expression profiling . It will be interesting to see whether pre-immune exposure influences the molecular basis of vaccination with irradiated sporozoites in the next phase of this study . This study shows that differential gene expression is particularly strong in naïve volunteers in comparison to semi-immune individuals at the time of malaria diagnosis . One way to interpret this result is that it provides a molecular signature of tolerance of , as opposed to resistance to , the pathogen [30] . In the presence of chronic exposure , the host immune system moves toward an equilibrium where pathogen is tolerated by mounting a measured immune response , without requiring complete sterile immunity that would likely have a greater physiological impact on the infected individual . This in turn implies that gene expression profiling of lymphocytes can be used to identify the type and duration of the immune signals that may be biomarkers for vaccine immunogenicity , and to establish how semi-immune exposure modifies their activation . | Plasmodium vivax malaria is a debilitating , occasionally life-threatening , and economically burdensome disease in Central Latin America , where 70%- 80% of the population lives with the risk of infection . We performed a gene expression profiling experiment taking advantage of a previously described sporozoite challenge experiment in Cali , Colombia that reported more severe malaria symptoms in subjects who have never experienced malaria . We show that no major differences are seen in the transcriptomes of uninfected naïve and semi-immune volunteers prior to infection , but differential expression of both neutrophil and interferon-related genes was evident at onset of malaria . Several hundred genes showed a stronger response in the naïve individuals just as parasites appear in the peripheral blood , and these fall into several pathways of interest . These findings show how information from gene expression profiling of whole blood can reveal the type and duration of the immune response to P . vivax infection , and highlights a subset of genes that may mediate adaptive immunity in chronically exposed individuals . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | Transcription Profiling of Malaria-Naïve and Semi-immune Colombian Volunteers in a Plasmodium vivax Sporozoite Challenge |
In individual mammalian cells the expression of some genes such as prolactin is highly variable over time and has been suggested to occur in stochastic pulses . To investigate the origins of this behavior and to understand its functional relevance , we quantitatively analyzed this variability using new mathematical tools that allowed us to reconstruct dynamic transcription rates of different reporter genes controlled by identical promoters in the same living cell . Quantitative microscopic analysis of two reporter genes , firefly luciferase and destabilized EGFP , was used to analyze the dynamics of prolactin promoter-directed gene expression in living individual clonal and primary pituitary cells over periods of up to 25 h . We quantified the time-dependence and cyclicity of the transcription pulses and estimated the length and variation of active and inactive transcription phases . We showed an average cycle period of approximately 11 h and demonstrated that while the measured time distribution of active phases agreed with commonly accepted models of transcription , the inactive phases were differently distributed and showed strong memory , with a refractory period of transcriptional inactivation close to 3 h . Cycles in transcription occurred at two distinct prolactin-promoter controlled reporter genes in the same individual clonal or primary cells . However , the timing of the cycles was independent and out-of-phase . For the first time , we have analyzed transcription dynamics from two equivalent loci in real-time in single cells . In unstimulated conditions , cells showed independent transcription dynamics at each locus . A key result from these analyses was the evidence for a minimum refractory period in the inactive-phase of transcription . The response to acute signals and the result of manipulation of histone acetylation was consistent with the hypothesis that this refractory period corresponded to a phase of chromatin remodeling which significantly increased the cyclicity . Stochastically timed bursts of transcription in an apparently random subset of cells in a tissue may thus produce an overall coordinated but heterogeneous phenotype capable of acute responses to stimuli .
Gene expression in living cells is dynamic and unstable , and fluctuations in transcription may be subject to stochastic regulation of processes including transcription factor and polymerase recruitment , and chromatin remodeling [1]–[5] . Cell-to-cell variation in the amount of protein a gene encodes is generally thought to arise from the typically small number of molecules ( e . g . gene copies ) , which are involved in gene expression . The factors leading to this variation have been defined by studies in prokaryotes and lower eukaryotes as either extrinsic ( deriving from variations in global , cellular factors , such as varying amounts of transcriptional activators ) or intrinsic ( i . e . inherently random molecular events , such as the transcription of mRNA or translation of proteins ) [4] , [6] , [7] . Previous studies addressing the characterization of intrinsic and extrinsic noise have mainly focused on bacteria and yeast models , often using pairs of reporter genes to assess heterogeneity in protein levels as an indirect readout of expression level at a fixed time-point [4] , [6] . One study has reported a similar fixed time-point analysis in single human cells using dual fluorescent protein read-out [7] . Short-term transcriptional pulses ( bursts ) have been observed in both prokaryotes [2] , [3] , [8] and eukaryotes [5] , [6] , [9] , [10] . Chromatin remodeling has been suggested as one possible intrinsic source of variation that may lead to the intense stochastic transcriptional bursts that have been shown to occur in eukaryotic gene expression [5] , [6] , [10] . In mammalian cells , the variation in transcription between cells has been most quantitatively studied by using fluorescent in situ hybridization analysis [10]–[12] . However , these studies do not provide long-term time-course analysis in single cells . One approach to provide real-time semi-quantitative analysis of transcription is the imaging of reporter gene expression—for example , using firefly luciferase [13]–[15] . Such studies support the view that gene expression is very dynamic over long time periods and occurs in transcriptional bursts of varying duration that are not coordinated between different cells . To date , the key aim of understanding real-time dynamics by directly quantifying transcription rates of multiple genes over time in single cells has not been achieved . One gene that displays dynamic transcription and marked heterogeneity between cells is prolactin ( PRL ) [14] , [16]–[18] . Prolactin is a hormone secreted by pituitary lactotrophic cells that is important for reproductive function , lactation , and control of fertility . Pituitary tumors secreting prolactin are common in man , and the hormonal regulation of prolactin secretion and gene expression has therefore been extensively studied [19]–[22] . In the present study , we used single cell reporter gene imaging to explore the pulsatile and cyclical nature of the transcription of PRL . Using measurements of mRNA and protein stability , we were able to quantify transcription rates from separately integrated reporter genes within the same cells and compare the kinetics of transcription over time . Analysis of the response to acute signals and the manipulation of histone acetylation suggested that dynamic chromatin changes control cycle timing .
Human PRL ( hPRL ) promoter-directed transcription was heterogeneous and dynamic in rat pituitary GH3 cells using luciferase reporter genes . Transcriptional pulses were observed in GH3 cell lines stably expressing a 5 kb hPRL-Luciferase ( hPRL-Luc ) reporter gene [14] or a larger 160 kb hPRL genomic locus reporter , a hPRL-Luc Bacterial Artificial Chromosome ( BAC , [16]; Figure 1A , B , and C; Figure S1 ) . Similar patterns were observed in primary cultures of pituitary cells ( taken from transgenic rats [16] ) , where the hPRL-Luc BAC was integrated either into an autosomal ( Figure 1D and E ) or an X-chromosome locus ( Figure 1F ) . In all four model systems distinct transcriptional cycles were discerned ( e . g . Figure 1G ) showing that these responses were not affected by promoter length or integration site . Pulses in gene expression in individual cells could arise from the transcription process itself or from signaling events reflecting the cellular environment . To discriminate between these possibilities , a dual-transgene cell line was constructed expressing separate luciferase and d2EGFP ( destabilized enhanced Green Fluorescent Protein ) reporter genes under the control of identical 5 kb hPRL promoters , integrated as independent gene copies ( GH3-DP1 cell line; Figure 2A; one or at most two copies; Figure S2 , Figure S3 ) . The luciferase and d2EGFP reporter genes were selected due to their reported short protein half-lives . The use of these very different reporter genes ( which have different chemistries for formation of the signal ) was considered an advantage because we could measure them entirely independently . Signal was detected from both reporter genes , but the intensity of the expression of the reporter genes within single cells failed to correlate when measured at a single time-point ( Figure 2A , Figure S4 ) . To measure the profiles of expression from each reporter gene , fluorescence and luminescence intensities were captured from the same field of single cells over several hours ( Figure S5 ) . Due to the different mRNA and protein half-lives of these two reporter genes ( Figure S2 , Figure S10 , Table S1 , Text S1 Section 3 ) direct comparison between the timing of expression could not be made . Therefore , in order to make quantitative comparisons between the timing of expression of these two different reporter genes within the same single cell , a mathematical model was developed ( Figure 2B , [23] ) which used statistical analysis to reconstruct estimates of the time-dependent transcription rate from the reporter imaging data ( Figure S11 , Figure S14 , Text S1 Section 3 ) . Autocorrelation analysis was performed on the reconstructed transcription rates from the hPRL-Luc and the hPRL-d2EGFP reporter genes in the dual-reporter GH3-DP1 cell line . This showed that transcription cycles were occurring at each gene with a dominant period of 11 . 3±3 . 3 h ( Figure 3A and Figure S9 ) . This value was measured from the dual reporter experiments and possibly provided an underestimate due to the limited timeframe of the experiments . Cycles of hPRL transcription were also observed from both luciferase and d2EGFP reporter genes in individual clonal pituitary cells from dual BAC-reporter transgenic rats grown in primary culture ( see Materials and Methods , [16] ) , with a slightly longer period ( 15 . 2±4 . 8 h; Figures 3B , S1 , and S8 ) . The ability to obtain quantitative data for the time-dependent transcription rates from the two reporter genes enabled us to ask whether the transcription cycles observed at individual loci were temporally coordinated or were out-of-phase within a single cell ( Figure 4A , Figure S6 ) . We analyzed the rank correlation coefficient C ( T ) between the transcription time-series for the two reporter genes over a time window of length T for increasing values of T ( see Text S1 Section 3 . 3 ) . In unstimulated conditions there was no significant correlation ( p< . 05 ) in the timing of transcription cycles between the dual reporters in the same single cell ( Figure 4A and B ) . In order to show that this was not an artifact of the genomic integration site we investigated independently derived cell lines: no significant correlation between the two promoters was detected in two different clones of the stably transfected cell lines or in dual-reporter transgenic primary cells ( Figure 4B , Figure S12 ) . These data demonstrate that cycles of hPRL-promoter activity did not depend on promoter length or on transgene integration site . Most importantly , the lack of correlation between the timing of hPRL transcription from promoters within the same cell in time-lapse imaging experiments showed that the expression cycles from distinct loci in a single cell were not synchronized or temporally coordinated . The fact that the cycles at individual loci in unstimulated single cells were uncorrelated ( and that this phenomenon occurred in both cell lines and post-mitotic primary pituitary cells ) suggested that the pulses in PRL gene expression were independent of cell cycle stage . This is in agreement with our previous study that suggested that variation in hPRL-luciferase reporter expression was independent of cell cycle in the GH3 cells ( which have a cell cycle timing of approximately 40 h ) [18] . Overall , these data therefore suggest that the transcriptional pulses were not a reflection of the cellular status , environment , or autocrine cell signaling but rather were due to an intrinsic property of the transcription process itself . In order to further understand and quantify the dynamics of transcriptional switching between “on-” and “off-”phases , we developed a stochastic binary switch model of transcriptional timing and used statistical algorithms to assess the distribution of times of transcriptional switching between the on and off states ( Figure 5A , Text S1 Section 3 ) . Transcription of the mRNA and translation and activation of the corresponding protein were modeled using a stochastic differential equation with binary on-off transcription and were fitted to time-series imaging data using a Markov Chain Monte Carlo ( MCMC ) algorithm ( Figure 5A ) . This produced relatively tight distributions on the levels and the timing of transcription in the on and off periods ( Text S1 Section 3 . 4 , example in Figures S15–S17 ) . This model was used to estimate the average and distribution of the times of luciferase transcriptional switching in the GH3-DP1 stable cell line and gave a dominant overall cycle period of 11 . 0±3 . 3 h ( Figure 5B ) , which was in close agreement with the independent autocorrelation analysis described above ( 11 . 3 h; Figure 3B ) . Furthermore , we estimated that there was an average on-phase duration of 4 . 0±1 h ( which was slightly longer than the timing previously described for transcriptional bursts in mammalian cells [24] ) . The average off-phase duration was 6 . 5±2 h per transcription cycle . No relationship was detected between the duration of the transcription on-phase and the preceding or subsequent off-phase ( Figure 5C and D ) . There was strong evidence for a refractory period of approximately 3 h , in which cells cannot respond to a stimulus with a further transcriptional pulse . For each cell studied , the mean length of the off periods never fell below 3 h ( Figure 5C and D ) . Thus , two distinct types of mathematical analyses indicated a similar duration for transcriptional cycles , and the stochastic binary switch model further suggested that the transcriptional on- and off-phases were independent , with each having defined average and minimum durations that may account for the kinetics of these cycles . The regulation of the transcriptional cycles from the hPRL promoter was then investigated by exposing GH3-DP1 cells to ( 1 ) combined forskolin and BayK-8644 ( FBK ) to activate both cAMP and Ca2+ signaling ( Figure S7 , [25] ) , ( 2 ) Trichostatin A ( TSA , a histone deacetylase inhibitor ) , or ( 3 ) both treatments combined ( TSA+FBK ) ( Table S2 ) . All three experimental treatments resulted in an initial synchronization between the transcription profiles of the two independent hPRL promoter-reporter transgenes: correlation between the profiles of the dual reporters was initially very high ( close to 1 ) , both between the two transgenes within individual cells ( Figure 6 ) and between different cells ( Figure S13 ) . This period of high correlation lasted longer following TSA treatment ( Figure 6C ) , when compared to the very transient synchronizing effect of FBK ( Figure 6B ) , and was most prolonged with combined TSA+FBK exposure ( Figure 6D and Figure S13 ) . Chromatin immunoprecipitation assays showed an increase in acetylated histone H3 DNA binding at the hPRL promoter following all treatments , with the highest level induced by TSA ( Figure 7 ) . Analysis of gene expression kinetics from the hPRL-Luc reporter gene in GH3-DP1 cells showed that the transcriptional cycles persisted following FBK treatment . However , they were only seen in less than 20% of cells treated with TSA ( Figure 8A and B ) . Analysis with the binary switch model showed that when the cells were treated with FBK , the median time to activation was longer and more variable than with TSA or TSA+FBK ( Figure 8C ) . This supports the hypothesis of a refractory period of transcription inactivation in which chromatin remodeling may play an important role . Treatment with TSA increased the duration of the on-phase and the initial rate of transcription ( Figure 8D and E ) . Combined TSA+FBK treatment increased the transcription rate during the on-phase following activation ( Figure 8E ) , resulting in a pronounced increase in maximum reporter gene expression ( Figure 8A ) . The response of cells to treatments that included TSA was more rapid and coordinated , suggesting that histone acetylation has a key role in the coordination of the temporal kinetics of transcription . Transcription of the hPRL gene might therefore require a long period of chromatin remodeling that is the source of the observed refractory phase . The rate at which mRNA is transcribed can be affected by different molecular mechanisms , including binding and dissociation of transcription factors , spatial reorganization , and/or chromatin remodeling . Previous studies [5] , [10] , [26] have considered a model in which the fluctuations of transcription rates are caused by overall dynamics that can be described by a “random telegraph process , ” where the gene switches between an active and an inactive state:The mean residence times for the active and inactive states are tactive = 1/k− and tinactive = 1/k+ and switch on- and off-times are drawn from an exponential distribution with means tactive and tinactive . For the on-times our results fit this hypothesis and the estimated distribution of on-times is exponential ( Figure 9A ) . Such a system would be memoryless in that the time already spent waiting in that state would not affect how much longer one would have to wait until the switch ( Text S1 Section 3 . 5 ) . The distribution of off-times in our data strongly contradicted this and was not distributed exponentially . This is shown in Figure 9B in which the exponential distribution ( black line ) is a poor fit of the data . We found that the system had a definite memory , where the length of time already spent in the inactive state affected the length of time remaining in that state ( Figure 9C ) . Thus the dynamics of these transcription cycles are not compatible with the mathematical models previously derived from analysis of single cell RNA counting [10] . An MCMC algorithm was applied which gave a distribution of estimates for the off-phase durations for single cells . When these estimates were amalgamated into a population distribution , the most likely off duration was at 3 h ( Figure 9B ) , which is consistent with the individual estimates in Figure 5 . Thus , the most likely explanation of the memory effect was the existence of a refractory period of approximately 3 h ( Figure 5C and D ) . If this refractory period is enforced , by removing any chance of an off duration of less than 3 h , then the excess off durations were distributed approximately exponentially ( Figure 9E ) . This refractory period means that the system still has a memory ( Figure 9F ) . Removing the refractory period ( Figure 9H ) meant that the system became memoryless ( Figure 9I ) . One effect of this memory or refractory period is to cause more cyclicity than would be seen in a telegraph process . In a system with a 3 h refractory period where the excess off-time is exponentially distributed ( Figure 9E ) , a higher proportion of the off-times would be just over 3 h and the system would appear more cyclic . To quantify the regularity of the transcription cycles , we simulated 1 , 000 cells with on and off durations drawn from the corresponding distributions . We then performed autocorrelation analysis on this simulated data set and the variance in the timing of the first peak was taken to be our measure of cyclicity . The variance in first peak timing was higher when there was no refractory period ( Figure 9D and J ) than when a refractory period was enforced ( Figure 9G ) . This analysis revealed that the presence of a defined refractory phase increases the regularity of the transcription cycles .
Physiologically important hormones such as PRL may be subject to both acute short-term regulation and long-term seasonal control . This could be achieved at the individual cell level by graded gene expression with feedback control of PRL expression . Such a model would suggest that each cell would express similar levels of PRL . Alternatively individual cells could dynamically switch between on- and off-phases producing a stable population average level of prolactin expression across the whole tissue . Studies of PRL promoter activity in intact pituitary tissue showed the whole tissue response was synchronized , although adjacent cells were not coordinated [27] . We have previously shown that expression from the PRL promoter is heterogeneous over time in individual cells from cell lines [14] and more recently in intact pituitary tissue [27] . The latter study suggested that isolated cells show greater heterogeneity than cells in tissue but that tissue-level cellular heterogeneity is still important . In a different study we recently described data that suggested that cellular heterogeneity may be genetically encoded by the timing of negative feedback loops in the NF-κB signaling system and that this may lead to out-of-phase oscillations in NF-κB signaling between cells [28] . This study raised the hypothesis that cellular heterogeneity may in fact be advantageous , leading to more robust tissue-level responses . ( Studies in other systems are in support of the idea that cellular variability is advantageous; e . g . [29] . ) The present study quantifies the level of heterogeneity in the dynamics of PRL gene expression in single cells . This heterogeneity may ensure stability in gene expression at the tissue level , while ensuring the readiness of the tissue as a whole to respond rapidly to signals . These data suggest that the overall level of hPRL transcription in pituitary tissue may be determined by three variables: ( 1 ) the frequency of transcriptional bursts , ( 2 ) the duration of the on-phase , and ( 3 ) the rate of transcription during the on-phase . Our studies suggest that within a population of cells there is a continuous transition from an activated “on” state to a basal “off” or “low” state , with an overall cycle of around 11 h . This cycle is longer than previously described transcriptional cycles/pulses [10]–[12] , with a different structure due to the presence of a defined refractory period of transcriptional inactivation . Because of the much longer time-scales involved , the source of the stochasticity would not be expected to derive directly from that due to random molecular processes involving small molecule numbers . However , within this cycle , the average transcriptional on-phase is ∼4 h , which is closer to previously defined transcriptional bursts and cycles [24] . The majority of the overall cycle time described here is therefore dominated by the off-phase . This could be due to a repressive chromatin state or alternatively could be regulated by chromosome topology with the timing dependent on the movement of the genes into and out of transcription factories [30]–[32] . Although the transcription cycle maintains relatively defined dynamics , the timing of transcription cycles from two independent promoters within a cell were heterogeneous , indicative of a system where intrinsic noise generated by local chromatin dynamics dominates extrinsic noise . Strong correlation between promoters could only be achieved following disruption of chromatin , suggesting that the cycles of hPRL transcription might involve epigenetic cycles of histone acetylation and deacetylation ( Figure 10 ) . Thus , an independent chromatin-regulated cycle of gene activity may occur at each locus . Cycles in the binding of transcription factors and polymerase at certain genes in the nucleus have been observed following oestrogen stimulation [24] , [33] , [34] . In that system , a refractory period also seems to occur [24] , [34] , as well as cycles of epigenetic chromatin modification [33] . The current study illustrates the importance of new integrated experimental and mathematical approaches for dynamic single cell analyses . Providing high-frequency time-lapse imaging data of the same cells over long time periods enabled the identification of dynamic transcription processes , phenomena which are invisible to the single cell RNA counting snapshots used for previous analyses of transcriptional bursting [35] , [36] . Biologically meaningful model parameters can be directly measured from the imaging data without requiring prior assumptions about the nature of the timings inherent in the system . As such , models fitted to these data accurately represent the underlying processes that lead to the time-course data . The data described in this study and by others [37] suggest that transcription cycles might emerge de novo from the intrinsic kinetics of the processes of transcription initiation , elongation , and termination . In our study we find that a transcriptional refractory period can have a dominant effect that leads to increased regularity in the timing of transcriptional cycles . This is in marked contrast with other cellular oscillatory systems such as NF-κB [38] , [39] , p53 [40] , Erk2 [41] , and the circadian clock [42] where negative feedback loops are believed to lead to the oscillatory dynamics with varying frequencies [43] . In particular in the NF-κB system , where a transcriptional delay ( refractory period ) of 45 min in IkBε activation following TNFα stimulation leads to cell-to-cell heterogeneity through the precise timing of this feedback loop [38] . This transcriptional delay is precisely timed to maximize the effect of IkBε transcriptional noise on ensuring out-of-phase oscillations [28] . A key question is whether the signals that activate hPRL in vivo are themselves graded or pulsatile . hPRL is itself under the control of NF-κB [44] , which responds dynamically to pulsatile cytokine stimulation [38] . Recently , glucocorticoid receptor , which also regulates hPRL expression [44] , has also been shown to cycle in response to pulsatile stimulation [45] . Although the secreted hormone PRL is stored in secretory granules and displays both pulsatile and circadian patterns , is it not yet clear how transcription relates to secretory events in individual cells . This system may be a new paradigm for understanding gene expression dynamics in vivo and may be important for understanding natural cell-to-cell variation in protein levels [46] , [47] . We have previously shown that gene promoter activity in lactotroph cells within pituitary tissue is non-uniform , with varying expression levels from adjacent cells [27] . However , these stochastic patterns together provide tissue-wide long-term coordinated behavior . If we include the new information gained in this present article we can start to build up a picture of a mosaic tissue structure , where , at any one time , a subset of cells are expressing PRL , a subset are in an inactive and refractory state , and a further subset are in an activatable state , ready to respond to a stimulus . A transient hormonal stimulus to the tissue would recruit this last subset of cells immediately , whereas sustained stimulus would progressively recruit additional cells exiting the refractory phase , resulting in a more sustained increase in PRL expression . If these stochastic and cyclical patterns of gene expression occur normally in intact tissue , such a mechanism would facilitate highly flexible transcriptional responses , allowing tissues to mount either acute or chronic responses to environmental cues , while maintaining a controlled average level of gene expression in the resting state .
All animals were handled in strict accordance with U . K . Home Office License regulations and subject to local ethical committee review . Fetal calf serum ( FCS ) was from Harlan Sera-Lab , Crawley Down , U . K . Luciferin was from Bio-Synth , Switzerland . Forskolin , BayK-8644 , and Trichostatin A were all from Sigma , U . K . The GH3-DP stable cell lines were generated by incorporating a 5 kb hPRL-d2EGFP reporter gene into the previously described GH3/hPRL-Luc cell line [14] . This was co-transfected with a hygromycin-selectable plasmid to enable antibiotic clonal selection . Generation of stable BAC-transfected rat pituitary GH3 cells containing the 160 kb hPRL-Luc gene was described previously [16] . GH3-DP1 cells , GH3-BAC cells , and collagenase type I dispersed primary pituitary cells were cultured in DMEM containing 10% FCS and maintained at 37°C 5% CO2 . Generation of BAC-transgenic rats expressing luciferase and d2EGFP under the control of identical hPRL 160 kb genomic fragments was described previously [16] . Transgenic line 37 was a luciferase-BAC transgenic rat line which was found to have two single reporter insertion sites . Further breeding of this line was undertaken to create two separate transgenic lines each with single integration sites . Line 37A had a single autosomal integration site , and line 37B an X-chromosomal integration site . ( These lines were the source of the pituitary cells used in Figure 1 . ) Independently , a destabilized EGFP reporter transgenic rat line was constructed , termed line 455 , which also had a single insertion site . The copy number of the luciferase-BAC transgene in line 37A was measured as ≤2 and that of the d2EGFP line 455 was ≤5 . In order to generate the dual reporter rat cell line ( previously referred to as PRL-Luc/d2eGFP in [16] ) , line 37A was crossed with line 455 . GH3-DP1 cells were seeded in 35 mm glass coverslip-based dishes ( IWAKI , Japan ) 20 h prior to imaging . Luciferin ( 1 mM ) was added at least 10 h before the start of the experiment , and the cells were transferred to the stage of a Zeiss Axiovert 200 equipped with an XL incubator ( maintained at 37°C , 5% CO2 , in humid conditions , carefully monitored to ensure equivalent conditions to a standard cell incubator ) maintained within a dark room . Luminescence images were obtained using a Fluar ×20 , 0 . 75 NA ( Zeiss ) air objective and captured using a photon-counting charge coupled device camera ( Orca II ER , Hamamatsu Photonics , U . K . ) . Bright field images were taken before and after luminescence imaging to allow localization of cells . Sequential images , integrated over 30 min , were taken using 4 by 4 binning and acquired using Kinetic Imaging software AQM6 ( Andor , Belfast , U . K . ) . In the relevant experiments , 5 µM forskolin and 0 . 5 µM BayK-8644 ( FBK ) , 30 ng/ml TSA , or both stimuli were added directly to the dish at the indicated times . Cells were prepared and imaged using the conditions and microscope described above . Excitation of d2EGFP was performed using an argon ion laser at 488 nm . Emitted light was captured through a 505–550 nm bandpass filter from a 545 nm dichroic mirror . Data were captured and analyzed using LSM510 software with consecutive autofocus . Cells were prepared and visualized using confocal microscopy as described above . A single field of cells was selected and five sequential fluorescence images were captured using autofocus . After a 10 min delay , the microscope and surrounding light-emitting devices were turned off or covered and a single luminescence image was captured using a cooled CCD camera ( 30 min integration ) . The equipment was then restarted ( taking 5 min ) , and after a 10 min delay ( to ensure laser stability ) , fluorescence images were taken . This hourly cycle was repeated for up to 21 h . GH3/hPRL-luc ( D44 ) cells ( 3×106 ) were plated in 10 cm2 dishes and left for 40 h . Dishes were treated for 2 h ( unstimulated , FBK , TSA , TSA+FBK ) and then ChIP assays were performed as described previously [38] based on the protocol by Upstate Biotechnology . Immunoprecipitation was carried out using 5 µg of either Anti-Acetylated H3 or Anti-IgG antibodies ( Upstate Biotechnology ) . DNA was extracted and amplified by PCR as described previously [38] . The following primer sequences were used: hPRL Promoter1 left GCAATCTTGAGGAAGAAACTTGA , right AGGCATTCGTTTCCCTTTTC amplifying 347 bp of DNA . PCR products were resolved using agarose gel electrophoresis and were analyzed by AQM Advance 6 . 0 software ( Kinetic Imaging , U . K . ) . Analysis was carried out using Kinetic Imaging AQM6 software ( Andor ) . Regions of interest were drawn around each single cell , and mean intensity data were collected . Data were collected from every single cell within the field . The average instrument dark count ( corrected for the number of pixels being used ) was subtracted from the luminescence signal . In dual reporter experiments , cells dividing within the experiment were eliminated from the analysis . For single reporter experiments , analysis ceased at the point of cell division . For the GH3 cell line , the cell cycle time is approximately 40 h [14] , [18] . We use the following ordinary differential equations model for the reconstruction of transcription profiles from protein data ( see also [23] , Text S1 Section 3 . 2 ) . where M and P denote concentration of reporter mRNA and protein , respectively . The first equation describes the dynamics of mRNA molecules with transcription function τ ( t ) and degradation rate δM . Protein is synthesized at a rate proportional to the abundance of mRNA and is degraded at rate δP . The various parameters will be different for the d2EGFP and Luc reporters . The transcription profile can be reconstructed viawhere the unobserved mRNA profile is expressed as a function of the observed solution path of P ( t ) , i . e . αM ( t ) = dP/dt+δPP ( t ) . Since M is not observed , prior knowledge of the rates δM and δP is necessary for the identification of the transcription profiles . We estimated these from two separate experiments where translation of reporter protein was inhibited by adding cycloheximide and transcription was inhibited by adding Actinomycin D ( see Text S1 Sections 1 and 3 . 2 . 2 ) . The rates estimated for δM and δP associated with d2EGFP and Luc are treated as known parameters . For inference on the transcription profile , the solution path P ( t ) is approximated by a flexible continuous function , here a spline representation , fitted to the observed protein data ( Text S1 Section 3 . 2 . 3 ) . The transcription profile is then reconstructed using the discrete Euler approximation to the differential equations for a small time interval , replacing P ( t ) by the fitted continuous function . In order to study the correlation in transcription of the dual reporter constructs within a cell , we compute the rank correlation coefficient between the two reconstructed transcription profiles of the two reporters d2EGFP and Luc within a cell ( Text S1 Section 3 . 3 ) . As this may vary over time ( in particular for stimulated experiments ) , all correlations are computed as a function of the length of time since a stimulus ( TSA , FBK ) was added starting from 1 . 5 h ( to allow for a reasonable minimal length over which any correlation is computed ) to 8 h . For unstimulated experiments we computed correlations after 2 h into the experiment to avoid any initial bias . The question of estimating and including protein maturation times is addressed in [48] . In calculating the correlations , the relevant quantity is the difference in maturation times between the two reporters . We have therefore included this process with a constant difference of up to 1 h . We have verified that such delays do not change our correlation results . It is clear from this that if we instead assumed an exponentially distributed delay with a similar difference in means , then this would not affect the correlation results . | Timing of biological processes such as gene transcription is crucial to ensure that cells and tissues respond appropriately to their environment . Until recently it was assumed that most cells in a tissue responded in a similar way , and that changes in cellular activity were relatively stable . However , studies of messenger RNA and protein levels in single cells have shown the presence of considerable heterogeneity . This suggested that transcription in single cells may be highly dynamic over time . Using a combined experimental and theoretical approach , with time-lapse imaging of reporter gene expression over 25 h periods , we measured the rate of prolactin gene transcription in single pituitary cells and detected clear cycles of transcriptional activity . Mathematical analysis , using a binary model that assumed transcription was on or off , showed that these cycles were characterized by a minimum refractory period that involved chromatin remodeling . The timing of transcription from two different reporter constructs driven by identical promoters in the same cell was out-of-phase , suggesting that the pulses of gene expression are due to processes intrinsic to expression of a particular gene and not to environmental effects . We further show that the pulses of transcription are independent chromatin cycles at each gene locus . Therefore , heterogeneous patterns of gene expression may facilitate flexible transcriptional responses in cells within intact tissue , while maintaining a well-regulated average level of gene expression in the resting state . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"computational",
"biology/systems",
"biology",
"cell",
"biology/gene",
"expression",
"computational",
"biology/transcriptional",
"regulation"
]
| 2011 | Dynamic Analysis of Stochastic Transcription Cycles |
Cell differentiation status is defined by the gene expression profile , which is coordinately controlled by epigenetic mechanisms . Cell type-specific DNA methylation patterns are established by chromatin modifiers including de novo DNA methyltransferases , such as Dnmt3a and Dnmt3b . Since the discovery of the myogenic master gene MyoD , myogenic differentiation has been utilized as a model system to study tissue differentiation . Although knowledge about myogenic gene networks is accumulating , there is only a limited understanding of how DNA methylation controls the myogenic gene program . With an aim to elucidate the role of DNA methylation in muscle development and regeneration , we investigate the consequences of mutating Dnmt3a in muscle precursor cells in mice . Pax3 promoter-driven Dnmt3a-conditional knockout ( cKO ) mice exhibit decreased organ mass in the skeletal muscles , and attenuated regeneration after cardiotoxin-induced muscle injury . In addition , Dnmt3a-null satellite cells ( SCs ) exhibit a striking loss of proliferation in culture . Transcriptome analysis reveals dysregulated expression of p57Kip2 , a member of the Cip/Kip family of cyclin-dependent kinase inhibitors ( CDKIs ) , in the Dnmt3a-KO SCs . Moreover , RNAi-mediated depletion of p57Kip2 replenishes the proliferation activity of the SCs , thus establishing a role for the Dnmt3a-p57Kip2 axis in the regulation of SC proliferation . Consistent with these findings , Dnmt3a-cKO muscles exhibit fewer Pax7+ SCs , which show increased expression of p57Kip2 protein . Thus , Dnmt3a is found to maintain muscle homeostasis by epigenetically regulating the proliferation of SCs through p57Kip2 .
Myogenic differentiation program has been extensively studied as a model of tissue differentiation since the discovery of MyoD [1] . Although much is known about the gene cascade of myogenesis [2 , 3] , the epigenetic mechanisms that regulate the physiological and pathological condition of skeletal muscles remain unknown [4] . Gene expression is regulated by both genetic and epigenetic mechanisms . DNA methylation is an epigenetic modification , which usually occurs at CpG sites [5]; the cytosine residues at CpG sites are methylated to 5-methyl-cytosine . This DNA methylation is mediated by a group of DNA methyltransferases ( Dnmt ) [6] . Among them , Dnmt3a and Dnmt3b catalyze de novo DNA methylation , and Dnmt1 mediates the maintenance of DNA methylation [7–9] . Accumulating evidence suggests that DNA methylation by Dnmt proteins in the promoter regions is associated with gene silencing , thus linking DNA methylation to gene suppression [6 , 10] . Recent studies have also clarified the roles of DNA methylation in gene bodies and intergenic regions in enhancing gene expression [11–14] . We previously reported that a transcriptional repressor Rp58 , which has been known to bind Dnmt3a [15] , is a direct target of MyoD and has an essential role in skeletal myogenesis [16] , in which DNA methylation at the promoter of myogenic genes is implicated [17] . Dnmt3a-null mice die within 3 to 4 weeks after birth , and deletion of Dnmt1 or Dnmt3b leads to early embryonic lethality [9 , 18 , 19] , indicating that DNA methylation has a critical role in embryogenesis and postnatal homeostasis . The Dnmt1-mediated maintenance of DNA methylation is necessary for self-renewal of the hematopoietic , mammary , mesenchymal and skin stem cells [20–23] . On the other hand , Dnmt3a and Dnmt3b coordinately generate DNA methylation profiles in differentiating stem cells , resulting in determination of distinct cell fates . In embryonic stem cells , concomitant deletion of Dnmt3a and Dnmt3b leads to a loss of differentiation capacity [24] . The precise role of de novo DNA methylation by Dnmt3a and Dnmt3b in muscle SCs , however , remains to be characterized . Hematopoietic stem cells null for Dnmt3a and/or Dnmt3b , progressively lose differentiation potential [25 , 26] and self-renewal capacity [27] . Neural stem cells deficient for Dnmt3a show impaired differentiation and increased cell proliferation [28] , and Nestin-Cre-mediated deletion of Dnmt3a causes motor neuron defects and premature death of the mice [29] . Dnmt3a-deficient osteoclast precursor cells do not differentiate into osteoclasts efficiently [30] . However , little is known about the functions of Dnmt3a in the muscle SCs . Proper muscle development and regeneration require coordinated gene expressions in embryonic muscle precursor cells and adult SCs [2 , 4] . The embryonic muscle precursor cells originate from dermomyotome , a dorsal part of the somite , which gives rise to myotome and dermatome . During embryogenesis , muscle precursor cells expressing Paired box 3 ( Pax3 ) transcription factor appear in dermomyotome . These Pax3+ cells are myogenic progenitor cells and a portion of them also express Pax7 . Most of the Pax3+/Pax7+ cells , and Pax3+/Pax7- cells are defined as myoblasts in later stages and develop into skeletal muscles . A small fraction of the Pax3+/Pax7+ cells becomes quiescent and settle in as SCs in postnatal skeletal muscles [31–33] . The myoblasts express muscle regulatory factors ( MRFs ) such as Myf5 , MyoD , Myogenin ( Myog ) and Mrf4 , and then differentiate and fuse with each other to form myotubes , which mature into myofibers [34] . Pax3-null mice are devoid of all limb muscles [35] . In the muscle tissues , SCs are located on the surface of myofibers , inside the ensheathing basal lamina , and regulated by both extrinsic and intrinsic factors [36–38] . In the steady state , SCs maintain quiescence and express Pax7 [31] . Upon muscle injury , they are activated and proliferate to form muscle fibers for regeneration [39] . Upon activation , Pax7 expression is rapidly lost and the MRFs are induced during regeneration . SCs are also responsible for postnatal muscle growth [40] , and age-related muscle decline is associated with functional impairment of SCs [38] . The number of tissue precursor cells increases during organ development and tissue regeneration . The precise mechanism underlying the proliferation of SCs is not fully understood . Cell cycle is regulated by a set of cell cycle factors , including Cyclins , Cyclin-dependent kinases ( CDKs ) , and CDK inhibitors ( CDKIs ) . CDKIs , the negative regulators of cell cycle , comprise two families , namely the INK4 and the Cip/Kip families . Members of the INK4 family ( p16INK4a , p15INK4b , p18INK4c and p19INK4d ) inhibit CDK4 and CDK6 , whereas Cip/Kip members ( p21Cip1 , p27Kip1 , and p57Kip2 ) mainly inhibit CDK2 and CDK4 [41] . Among them , p57Kip2 ( also called as Cdkn1c ) is reportedly important to maintain the hematopoietic stem cells in a non-proliferative state [42 , 43] . The p57Kip2 is located at an imprinted locus and loss-of-function mutations in p57Kip2 cause Beckwith-Wiedemann syndrome , an overgrowth disorder which is characterized by increased organ sizes including that of muscles [44 , 45] , and gain-of-function mutations cause undergrowth disorders such as Silver-Russell syndrome [46–48] . Here , we show an indispensable role of Dnmt3a in muscle SCs by utilizing muscle precursor cell-specific Dnmt3a deletion in mice , and identify p57Kip2 as a critical target gene of Dnmt3a for the proper proliferation of SCs .
To assess the role of DNA methylation in muscle development , we analyzed muscle precursor cell-specific Dnmt3a cKO mice . We established a mouse line in which Dnmt3a gene was deleted by Cre recombinase driven by a Pax3 promoter ( Fig 1A ) . The efficiency of deletion in tibialis anterior muscles of cKO mice was approximately 70% at the genomic DNA level ( Fig 1B ) , and over 90% at the mRNA level in tibialis anterior , gastrocnemius , paraspinal muscles and diaphragm ( Fig 1C ) ; Dnmt3b expression level was unaffected ( S1A Fig ) . The Dnmt3a-cKO mice exhibited significantly smaller body sizes than WT littermates at 8- to 12-week old ( Fig 1D ) , although they were born at normal Mendelian ratios , and were viable . The Dnmt3a-cKO mice weighed less than WT controls and the difference was more prominent in females ( Fig 1E ) . No apparent skeletal deformity was observed using X-ray whole body imaging ( Fig 1F ) . Muscle tissues were hypoplastic in Dnmt3a-cKO mice ( S1B Fig ) . Computed Tomography ( CT ) scan of distal hindlimbs revealed significantly reduced muscle mass in the Dnmt3a-cKO mice compared to WT controls ( Fig 1G and 1H ) , and the difference was more prominent in females ( Fig 1G and 1H ) . Histological analysis of the gastrocnemius muscles revealed that myofibers in Dnmt3a-cKO muscles were narrower than WT myofibers ( Fig 1I and 1J ) . Median myofiber cross sectional area ( CSA ) of the Dnmt3a-cKO muscles was significantly smaller than that of the WT muscles ( Fig 1K ) . Growth retardation and decreased muscle mass in Dnmt3a-cKO mice persisted at later stages as well and growth did not catch up with WT littermates . These findings indicate that the loss of Dnmt3a in muscles leads to reduced muscle mass . The relatively well- maintained muscle tissue patterns prompted us to investigate the status of muscle differentiation . Gene expression analysis in muscles did not reveal any significant differences in myogenic gene expression between Dnmt3a-cKO and WT muscles ( S1C Fig ) , suggesting that Dnmt3a deletion does not affect myogenic differentiation . These findings suggest that the loss of Dnmt3a in the Pax3+ myogenic precursor cells leads to decreased muscle mass in mice . The finding that Dnmt3a-cKO muscles are hypoplastic implied that the potential of muscle precursor cells to grow organs had reduced . To investigate the role of muscle SCs in recreating muscle tissues , we probed muscle regeneration in the cKO mice ( Fig 2A ) . The tibialis anterior muscles were injected with cardiotoxin ( CTX ) to induce tissue injury . Histological analysis of the muscles 7 days after the CTX treatment revealed smaller regenerated myofibers with a central nucleus , in the Dnmt3a-cKO muscles than in the WT muscles ( Fig 2B and 2C ) . Median regenerative myofiber CSA of Dnmt3a-cKO muscles was significantly smaller than that of WT muscles ( Fig 2D ) . These findings indicate that muscle regenerative capacity is impaired in Dnmt3a-cKO mice . Since the loss of Dnmt3a causes decreased muscle formation in adult mice also , it implies that Dnmt3a loss impairs the function of adult SCs . To gain a mechanistic insight into how loss of Dnmt3a leads to a functional impairment of the SCs , we performed an in vitro analysis of the muscle SCs . We isolated SCs from Pax3-Cre; Dnmt3a-cKO mice and WT littermates and cultured the cells in growth conditions . The proliferation of Dnmt3a-cKO SCs was impaired relative to that of WT SCs , indicating that Dnmt3a is required for SCs to re-enter the cell cycle ( S2A and S2B Fig ) . Because Pax3 is expressed during development , we considered that there may be an effect of Pax3-dependent Dnmt3a deletion during the development of SCs . In our evaluation of the non-muscle effects of the Pax3 promoter-dependent Dnmt3a mutation , we found that Pax7-KO mice , which completely lack SCs , exhibit growth retardation and thin myofibers , indicating that dysfunction in SCs leads to growth retardation [40] . Accordingly , we considered that the Dnmt3a-cKO mouse phenotype was attributable to impaired SC function . To eliminate the possible developmental deficit of SCs and non-muscle effects , we utilized a tamoxifen-inducible Pax7-CreERT2 system and generated Pax7-CreERT2; Dnmt3aflox/flox mice for later analyses . Pax7-Cre; Dnmt3a-KO SCs were isolated from Pax7-CreERT2; Dnmt3aflox/flox mice after tamoxifen injection ( Fig 3A ) . Dnmt3a KO efficiency was over 99% both at the genomic DNA level ( Fig 3B ) and mRNA level ( Fig 3C ) . The morphologies of the isolated Dnmt3a-KO SCs were indistinguishable from those of WT SCs ( Fig 3D , Day 1 ) . However , Dnmt3a-KO SCs showed a striking loss of expansion in culture and their growth rate was significantly lower than that of WT SCs ( Fig 3D and 3E ) . To explore whether the impaired expansion of Dnmt3a-KO SCs was caused by decreased proliferation of the SCs , we performed phosphorylated histone H3 ( PHH3-Ser10 ) immunostaining of the SCs . The frequency of the PHH3-Ser10+ Dnmt3a-KO SCs was significantly lower than that of WT SCs ( Fig 3F and 3G ) . We also performed 5-ethynyl-20-deoxyuridine ( EdU ) incorporation assay . EdU+ cells were significantly less frequent in Dnmt3a-KO SCs than in WT SCs ( S4 Fig ) . These findings suggest that cell proliferation is impaired in Dnmt3a-KO SCs . With regard to apoptosis , we immunostained proliferating Pax7-Cre; Dnmt3a-cKO and WT SCs with a cleaved Caspase-3 antibody . The frequency of cleaved Caspase-3-positivity was very low in Dnmt3a-cKO SCs and not statistically different from that in WT SCs . These results suggest that the loss of expansion observed in Dnmt3a-KO SCs was attributable not to activated apoptosis but to decreased proliferation ( S5 Fig ) . To examine the influence of the Dnmt3a deletion on the differentiation capacity of SCs , myogenic differentiation was induced by serum starvation . The number of cells was strictly adjusted so that differentiation was induced at the same confluency in both Dnmt3a-KO and WT SCs . The Dnmt3a-KO SCs showed no apparent morphological differences from WT SCs ( S6A Fig ) . Also , the expression of myogenic genes was not different significantly , indicative of the unaffected myogenic differentiation capacity of the Dnmt3a-KO SCs , compared to the WT SCs ( S6B Fig ) . Collectively , loss of Dnmt3a leads to decreased proliferation of muscle SCs . To elucidate the mechanism of how Dnmt3a regulates the proliferative capacity of SCs , we performed transcriptome analysis of Dnmt3a-KO SCs . To minimize the potential developmental differences in the SCs of the Dnmt3a-cKO mice , we established a temporal deletion of Dnmt3a by infecting Dnmt3aflox/flox SCs with adenovirus expressing Cre-recombinase ( Ax-Cre ) . The Dnmt3a deletion efficiency was approximately 70% at the mRNA level ( Fig 4A ) . Consistent with the gene expression analysis in the Pax7-dependent deletion of Dnmt3a , the expression of myogenic genes was not significantly altered in the Ax-Cre-mediated Dnmt3a-KO SCs ( S7A Fig ) . Among cell-cycle related genes , the expression of p57Kip2 , a negative regulator of cell cycle , increased in the Ax-Cre Dnmt3a KO SCs without induction of differentiation ( Fig 4B ) . The increased expression of p57Kip2 was also observed in the Pax7-dependent Dnmt3a-KO SCs ( Fig 4C ) , and it continued even after differentiation ( Fig 4C ) . Immunostaining with a p57Kip2 antibody showed significantly higher intensities of fluorescence in Pax7-Cre; Dnmt3a-cKO SCs than in WT SCs , suggesting enhanced expression of p57Kip2 in the Pax7-Cre; Dnmt3a-cKO SCs ( Fig 4D and 4E ) . According to RT-qPCR analysis of Pax7-Cre; Dnmt3a-KO and WT SCs for all of the other CDKIs , the expression level of p16INK4a was only elevated by Dnmt3a loss ( S8 Fig ) . But the difference of p16INK4a expression between Dnmt3a-KO and WT SCs was much smaller than that of p57Kip2 . Therefore , we considered p57Kip2 as a primary candidate of a causative factor of impaired proliferation of Dnmt3a-KO SCs . Collectively , loss of Dnmt3a leads to elevated expression of p57Kip2 in SCs . To determine whether the mis-expression of p57Kip2 in Dnmt3a-KO SCs is attributable to alteration of DNA methylation , we performed a bisulfite sequencing analysis in the Pax7-dependent Dnmt3a-KO and WT SCs . It was found that the p57Kip2 promoter region was extremely hypomethylated in the undifferentiated Dnmt3a-KO SCs ( Fig 5A and 5B ) , suggesting that the extent of DNA methylation in the promoter region underlies p57Kip2 expression . Since we confirmed by lineage tracing that pure Pax7+ cells were isolated by the single myofiber culture method ( S9 Fig ) , the difference in DNA methylation levels between Dnmt3a-KO and WT SCs did not seem to be due to contamination by non-myogenic cells . To examine whether p57Kip2 is a functional target of Dnmt3a in regulating the proliferation of SCs , we tested the effect of p57Kip2 depletion in the Dnmt3a-KO SCs . The cell proliferation defect was partially but significantly rescued by p57Kip2 knockdown ( Fig 5C and 5D ) . In line with these data , the decreased frequency of PHH3+ Dnmt3a-KO SCs was also partly rescued by p57Kip2 knockdown ( Fig 5E ) , indicating that Dnmt3a regulates the proliferation of SCs by controlling the expression of p57Kip2 . Accordingly , our findings suggest that the decreased proliferation of SCs is , at least partly , due to mis-expression of p57Kip2 caused by DNA hypomethylation . DNA hypomethylation of the p57Kip2 promoter in the Dnmt3a-KO SCs prompted us to suppose that it is a methylation target of Dnmt3a . To assess the recruitment of Dnmt3a to the p57Kip2 regulatory region , a ChIP-qPCR analysis was performed with Dnmt3a in undifferentiated proliferating WT SCs . The p57Kip2 regulatory region was enriched with Dnmt3a at a similar level as the H1foo promoter , which is DNA-methylated except in oocytes ( S10A Fig ) . The primers for the ChIP in the H1foo locus were designed on the basis of Dnmt3a2-ChIP-seq data by Baubec et . al [49] ( S10B Fig ) . The housekeeping gene Rps18 promoter , which is consistently DNA hypomethylated , was not enriched with Dnmt3a . These findings suggest that the p57Kip2 regulatory region is a direct methylation target of Dnmt3a in SCs . In contrast to p57Kip2 , the p16INK4a promoter region was not enriched in the Dnmt3a ChIP ( S10A Fig ) , suggesting that this region is not a direct target of Dnmt3a . Taken together , p57Kip2 is a methylation target of Dnmt3a and regulates proliferation of SCs . To extend our in vitro findings to an in vivo context , we checked p57Kip2 expression in the Pax3-Cre; Dnmt3a-cKO muscles . Immunostaining with a p57Kip2 antibody in single myofibers revealed a higher level of p57Kip2 protein expression in Dnmt3a-cKO muscles ( Fig 6A ) . We further performed costaining of Pax7 and p57Kip2 in Dnmt3a-cKO and WT myofibers . The expression of p57Kip2 was very weak in the WT Pax7+ SCs ( Fig 6B ) . In contrast , p57Kip2 was costained with Pax7 in the cKO myofibers , indicating that expression of p57Kip2 is indeed enhanced in the SCs ( Fig 6B ) . Bisulfite sequencing analysis revealed significant hypomethylation at the promoter region of p57Kip2 in the Dnmt3a-cKO muscles ( Fig 6C and 6D ) , corroborating the findings in the Pax7-Cre; Dnmt3a-KO SCs . Since p57Kip2 is also mis-expressed in the Dnmt3a-cKO muscles , this implies that Dnmt3a regulates p57Kip2 expression through epigenetic mechanisms , both in vitro and in vivo . Our findings indicate that Dnmt3a loss impairs muscle regenerative capacity and reduces proliferative capacity of SCs . To determine whether the impaired muscle regeneration was a result of impaired SC proliferation , we assessed the frequency of SCs expressing Pax7 in both the unperturbed and the regenerating muscles . The frequency of Pax7+ cells in all nucleated cells in unperturbed Pax3-Cre; Dnmt3a-cKO muscles was not significantly different from that in WT muscles ( Fig 7A and 7B ) . However , in the regenerating muscles , Pax7+ cells were less frequent in the Dnmt3a-cKO mice than in the WT mice ( Fig 7A and 7B ) . Pax7/Laminin costaining demonstrated that most of these Pax7+ cells were located inside the basal lamina of the regenerated myofibers ( S11 Fig ) . Next , to examine whether the lower frequency of Pax7+ cells in the Dnmt3a-cKO regenerating muscles was caused by decreased proliferation of the SCs , phospho-histone H3 ( Ser10 ) immunostaining was performed in the regenerating tibialis anterior muscles . Immunostaining at 7 days after CTX injection revealed that PHH3+ cells were less frequent in the Dnmt3a-cKO than WT mice ( Fig 7C and 7D ) . These results suggest that the SCs are not wasting in the uninjured muscles of Dnmt3a-cKO mice , but that their ability to proliferate after injury is impaired , leading to defects in their regenerative capacity . Immunostaining with a p57Kip2 antibody showed that p57Kip2+ cells were more frequent in the Dnmt3a-cKO than in the WT regenerating muscles ( S11B and S11C Fig ) . The behavior of SCs was explored by Pax7/MyoD-costaining and Myog immunostaining in regenerating muscles . The ratios of MyoD+Pax7+ cells to MyoD-Pax7+ cells were lower in Dnmt3a-cKO regenerating muscles than in the WT ( S11D and S11E Fig ) , suggesting SC activation is impaired in Dnmt3a-cKO muscles . Myog+ cells were less frequent in Dnmt3a-cKO regenerating muscles compared to those in the WT ( S11F and S11G Fig ) . This lower frequency of Myog+ cells does not necessarily indicate impaired differentiation capacity as a result of the Dnmt3a deletion , because Dnmt3a-cKO reduced the number of proliferating SCs , which produce the differentiating SCs . Taken together , these results suggest that the SCs are not wasting in the uninjured muscles of Dnmt3a-cKO mice but their proliferation after injury is impaired , leading to the defects in the regenerative capacity . In summary , Dnmt3a regulates the proliferation of muscle SCs , thereby influencing the growth of SCs in culture and the regenerative capacity of skeletal muscles . Hence , Dnmt3a maintains muscle homeostasis by regulating the functions of SCs through the epigenetic regulation of p57Kip2 .
In this study , we have shown that loss of Dnmt3a in the Pax3-expressing cell lineage leads to reduced body size and muscle mass in mice . Although Pax3-Cre; Dnmt3a-cKO mice exhibited grossly normal tissue patterns , they had thinner myofibers , unproportionally decreased muscle mass and impaired muscle regeneration , suggesting that Dnmt3a contributes to the function of SCs that are responsible for postnatal muscle growth and regeneration . Pax7-/- mice which completely lack SCs display similar phenotypes to those of Dnmt3a-cKO mice , namely decreased muscle mass and reduced myofiber diameter , although the overall organization of myofibers appears normal [40] . The phenotypes of Pax7-/- mice are attributable to a lack of SC fusion during the postnatal period [40] . We also identified p57Kip2 as an essential downstream target of Dnmt3a for methylation and a causative candidate gene for the functional deficits in Dnmt3a-cKO SCs . This is corroborated by the finding that p57Kip2 knockdown ameliorates the decreased proliferation of the Dnmt3a-cKO SCs . Dnmt3a deletion in SCs impairs proliferation through the mis-expression of p57Kip2 , resulting in quantitative insufficiency of SCs similar to that in Pax7-/- mice ( Fig 8 ) . Roles of p57Kip2 in regulating body and organ sizes have been elucidated in the context of human overgrowth and undergrowth disorders . p57Kip2-deficient mice have phenotypes similar to the manifestations of Beckwith-Wiedemann syndrome ( BWS ) , an overgrowth disorder [50 , 51]; in addition , p57Kip2 activity is lower in BWS patients [44 , 52] . Silver-Russell syndrome ( SRS ) is a heterogeneous disorder characterized by pre- and post-natal growth retardation [53 , 54] . IMAGe syndrome is another undergrowth disorder characterized by intrauterine growth retardation , metaphyseal dysplasia , adrenal hypoplasia and genital anomalies [55] . Loss-of-function mutations of p57Kip2 have been identified in BWS patients [44] , and gain-of-function mutations in the Proliferating cell nuclear antigen ( PCNA ) -binding domain of p57Kip2 have been identified in growth retardation syndromes such as SRS and IMAGe syndrome [46–48] . It is well known that genomic imprinting is controlled by DNA methylation and that p57Kip2 is paternally imprinted . DNA methylation at the imprint center is maintained by Dnmt1 , a maintenance DNA methyltransferase , but Dnmt1 alone is not able to maintain all of the DNA methylation loci , especially in CpG-rich regions [24 , 56] . Therefore , there is a possibility that maintenance DNA methylation deficits besides de novo DNA methylation is caused by Dnmt3a deletion , resulting in the progressive loss of genomic imprinting . However , we think the mis-expression of p57Kip2 in Dnmt3a-KO SCs is not a result of lost genomic imprinting because the imprint center is not located in the p57Kip2 promoter and because p57Kip2 is expressed only from the methylated maternal allele [52] . Considering this regulatory mechanism , the expression of p57Kip2 should be decreased as a result of loss of genomic imprinting . In our Dnmt3a-KO SCs , p57Kip2 expression level was lower than that of the WT , which implies that there was no change in genomic imprinting . If the cell population is perfectly homogeneous , the DNA methylation level of a CpG site should be either 100% or 0% . Isolated SCs in our experiments are all Pax7-positive ( S9 Fig ) , but their differentiation status after in vitro culture is not perfectly homogeneous . We consider some SCs might not get out of quiescence and others might be beginning spontaneous differentiation , and therefore the DNA methylation levels of WT SCs at the p57Kip2 promoter were not 100% . In fact , during culture of isolated myofibers , some SCs divide asymmetrically into two types of cells that are distinctively fated to self-renew or to differentiate [57] . Hence , SCs are considered heterogeneous population composed of stem cells and committed progenitors . A certain proportion of SCs may divide asymmetrically even when cultured on dish . In addition , a DNA methylation level of the p57Kip2 promoter was not 0% even in Dnmt3a-KO SCs . This might be because Dnmt3b incompletely compensates the influences of Dnmt3a deletion . Although our findings reveal an essential role of p57Kip2 in the undifferentiated SCs , p57Kip2 is also known to be a target of MyoD , which promotes muscle differentiation [58] . We also observed a further increase in the expression of p57Kip2 after myogenic differentiation , coincident with the cell cycle deceleration in the differentiating SCs . Our findings suggest that Dnmt3a-KO prematurely triggers the induction of p57Kip2 in the undifferentiated SCs , which results in a reduced number of SCs forming mature myofibers . The decrease in body size and muscle mass of Dnmt3a–cKO mice were more severe in females . We could not identify the reason for this gender difference; previous studies of Dnmt3a deletion in other tissues have not reported such gender-dependent severity of phenotypes . However , female mice show more severe phenotypes of several heart diseases [59 , 60] . In the mdx mouse model of Duchenne cardiomyopathy , aged female mice display more severe cardiomyopathy [61] . Although the detailed reasons for such differences are not clear , it is possible that the female muscular tissues are more susceptible to a specific pathological condition . Another epigenetic regulatory mechanism , histone modification is also known to regulate SC functions . Histone deacetylase inhibitors increase muscle cell size by promoting cell fusion without affecting cell proliferation [62] . On the other hand , conditional ablation of Polycomb-repressive complex ( PRC2 ) subunit EZH2 in Pax7+ cells results in impaired SC proliferation and reduced muscle mass with small myofibers [63] . Taken together , it is suggested that multiple epigenetic mechanisms coordinately regulate SC functions and control the tissue size of skeletal muscles . Thus , the loss of Dnmt3a in muscle progenitor cells leads to premature expression of a CDKI , p57Kip2 , which causes decreased proliferation of the SCs , leading to smaller body size and disproportionately reduced muscle mass in mice . Our findings indicate that there are several potential mechanisms for size regulation . Firstly , DNA methylation , which specifies the sets of genes to be expressed in a certain context , influences body size . Secondly , the number of tissue stem cells , which is balanced between self-renewal and differentiation commitment , might influence body and organ sizes . There is an increased incidence of rhabdomyosarcoma among BWS patients [64 , 65] , which implies that deteriorated size regulation leads to tumorigenesis . Our current understanding of the mechanisms regulating body and organ size is limited; however , further elucidation of the size control machinery may lead to novel therapeutic approaches for cancer that target these mechanisms . In this study , we show that Dnmt3a regulates proliferation of muscle SCs by methylating the p57Kip2 locus and suggest that this Dnmt3a-p57Kip2 axis forms the basis of size-control mechanisms in muscle tissues . Further elucidation of the underlying relation between DNA methylation and body and organ size control , will provide novel insights for developing new therapeutic approaches for some of the incurable human disorders .
We used mice in our research . The mice were anesthetized by intraperitoneal injection of pentobarbital or inhalation of isoflurane . Cervical dislocation was used as a euthanasia method . All animal experiments were approved by the Institutional Animal Care and Use Committee at Tokyo Medical and Dental University ( approval number; 0160127A ) . Dnmt3a-flox mice were kindly provided by Dr . M . Okano . Dnmt3a-floxed allele was previously described [66] . Pax3-Cre mice and Pax7-CreERT2 mice were purchased from the Jackson Laboratory ( Bar Harbor , ME ) . Pax3-Cre allele and Pax7-CreERT2 allele were previously described [67 , 68] . Genomic DNA was isolated from muscle tissues using DNeasy Blood & Tissue Kit ( Qiagen , Hilden , Germany ) according to the manufacturer’s instructions . Gene deletion efficiency was calculated by genomic DNA qPCR . Relative genomic DNA level was determined by the standard curve method . All primer sequences are listed in S1 Table . Computed tomography ( CT ) scan of distal hindlimbs was performed using Latheta LCT-200 ( Hitachi Aloka Medical , Tokyo , Japan ) . Mice were anesthetized by isoflurane inhalation during the scan . The image data were analyzed using Latheta software ( Hitachi Aloka Medical , Tokyo , Japan ) , and muscle and bone cross-sectional volume were calculated . The slice of each limb where the muscle cross-sectional area was the greatest was selected for muscle volume evaluation , for each mouse . Muscle tissues of 8- to 12week-old mice were frozen in isopentane cooled in liquid nitrogen . Frozen tissues were sectioned using a cryostat CM3050S ( Leica , Wetzlar , Germany ) at 10 μm thickness and mounted on MAS-coated slide glasses ( Matsunami Glass , Osaka , Japan ) . The CSA of myofibers were measured in at least five fields of view using ImageJ software ( National Institutes of Health , Bethesda , MD ) . For Hematoxylin-Eosin ( HE ) staining , muscle sections were fixed in 4% paraformaldehyde ( PFA ) in phosphate buffered saline ( PBS ) at room temperature for 10 minutes , then immersed in Mayer’s Hematoxylin Solution ( Wako , Osaka , Japan ) for 5 minutes , followed by washing under running water for 10 minutes . After staining with 1% Eosin Y Solution ( Wako , Osaka , Japan ) for 1 minute , they were sequentially immersed in 70% , 95% and 100% ethanol for 30 seconds , 1 minute and 3 minutes , respectively . Finally , they were washed thrice in xylene for 3 minutes each and embedded in Entellan Neu ( Merck KGaA , Darmstadt , Germany ) . Dnmt3a-KO SCs were harvested from 6- to 8-week-old Pax7-CreERT2; Dnmt3aflox/flox mice . Tamoxifen ( Sigma , St Louis , LA ) was administered to the mice intraperitoneally at the dose of 100 μg/body weight ( g ) for 5 consecutive days . After seven days of the first tamoxifen administration , the mice were sacrificed to harvest gastrocnemius muscles , and SCs were isolated as previously described [69 , 70] . Briefly , single myofibers were obtained by collagenase digestion and cultured in primary cultured myocyte growth medium ( pmGM ) consisting of Dulbecco’s modified Eagle’s medium ( DMEM; Sigma , St Louis , LA ) with 20% fetal bovine serum , 1% penicillin/streptomycin ( Life Technologies , Grand island , NY ) , 2% Ultroser G ( Pall , New York , NY ) , 1000 U/ml mouse leucocyte inhibitory factor ( LIF; AMRAD Biotech , Victoria , Australia ) and 10 ng/ml human basic fibroblast growth factor ( bFGF; PeproTech EC , London , UK ) on type I collagen-coated dishes ( Sumilon , Tokyo , Japan ) at 37°C under 10% CO2 in a humidified chamber . SCs migrated from the myofibers in 4 to 5 days . For analyzing growth of SCs , isolated SCs were cultured in pmGM . To induce myogenic differentiation , SCs were cultured in DMEM with 2% horse serum . Frozen muscles were sectioned at 10 μm thickness and mounted on MAS-coated slide glasses ( Matsunami Glass , Osaka , Japan ) . Single myofibers were isolated by collagenase digestion as previously described [69 , 70] , and plated on MAS-coated slide glasses ( Matsunami Glass , Osaka , Japan ) . Sections or myofibers were dried in the air and then fixed in 4% PFA in PBS at room temperature for 10 min . For immunocytochemistry , cultured cells are fixed in 4% PFA in PBS at room temperature for 10 min . After permeabilization with 0 . 1% Triton X-100 in PBS for 20 min , they were blocked with 1% Bovine serum albumin ( BSA ) in PBS for 1 hour and incubated with primary antibodies at 4°C overnight . The following antibodies were used: anti-Pax7 ( described previously [71] ) , anti-MyoD ( BD Pharmingen 554130 , 1:100 ) , anti-Myog ( Santa Cruz sc-576 , 1:50 ) , anti-Phospho-Histone H3 ( Ser10 ) ( Cell Signaling #9701 , 1:400 ) , anti-active Caspase-3 ( Abcam ab2302 , 1:200 ) , anti-p57Kip2 ( Santa Cruz sc-8298 , 1:100 ) , anti-p57Kip2 ( Cell Signaling #2557 , 1:500 ) and anti-Laminin 2 alpha ( Abcam ab11576 , 1:500 ) . After the primary antibody incubation , sections were incubated with secondary antibodies conjugated with Alexa Fluor 488 or 594 ( Life Technologies , 1:1000 ) . Finally , they were mounted in VectaShield with DAPI ( Vector Laboratories , CA , USA ) . The mean intensity of fluorescence signals in each cell was calculated using ImageJ software ( National Institutes of Health , Bethesda , MD ) . SCs were harvested as described above and cultured in pmGM for about 7 days to expand enough for the assay . One day after a passage to adjust confluency , they were cultured in medium containing 10 μM EdU for 3 hours for EdU labeling . EdU incorporation was assessed using Click-iT Plus EdU Alexa Fluor 488 Imaging Kit ( Life Technologies , Grand island , NY ) . Total RNA was isolated from the homogenized muscle tissues using ISOGEN ( Nippon Gene , Tokyo , Japan ) according to the manufacturer’s instructions . One μg of total RNA was used to synthesize cDNA . Reverse transcription was performed using ReverTra Ace ( Toyobo , Osaka , Japan ) following the manufacturer’s instructions . qPCR was performed by Thermal Cycler Dice Real Time System II ( Takara Bio , Japan ) using Thunderbird SYBR qPCR Mix ( Toyobo , Osaka , Japan ) and the relative expression levels were detected by the ΔΔCt method . All primer sequences are listed in S1 Table . Microarray analysis ( Affymetrix ) was performed with RNA samples derived from the WT- and Dnmt3aflox/flox-SCs infected with Ax-Cre ( MOI 30 ) at 0 , 12 , 24 , 48 , 72 and 96 hours of differentiation in vitro . The data were normalized and z transformed for the hierarchical clustering analysis utilizing Multiple Experiment Viewer [72] . Bisulfite conversion of the isolated genomic DNA was performed by CpGenome Turbo Bisulfite Modification Kit ( Millipore , Billerica , MA ) according to the manufacturer’s instructions . Bisulfite-treated DNA was amplified by PCR using Quick Taq HS DyeMix ( Toyobo , Osaka , Japan ) . All primer sequences are listed in S1 Table . PCR products were cloned into T-Vector pMD20 ( Takara Bio , Shiga , Japan ) and sequenced with the M13 reverse primer from at least 12 clones . Fifty microliters of 0 . 03 mg/ml cardiotoxin ( CTX; Sigma , St Louis , LA ) was injected into the bilateral tibialis anterior muscles of 8- to 12-week-old mice , after making skin incisions to expose the fascia on bilateral hindlimbs under anesthesia . The mice were sacrificed 7 to 14 days after CTX injection , and the injured muscles were harvested for histological analysis and gene expression analysis . p57Kip2 knockdown was achieved by p57Kip2 siRNA transfection . SCs were disseminated on type I collagen-coated dishes at a density of 0 . 1 × 105 cells/ml . After verifying cell adherence to the dishes , siRNA was transfected at a final concentration of 20 nM , using Lipofectamine RNAiMAX Transfection Reagent ( Invitrogen , Carlsbad , CA ) according to the manufacturer’s instructions . SCs were counted daily , starting from day 1 after transfection . MISSION siRNA targeting murine p57Kip2 was supplied by Sigma-Aldrich ( St . Louis , MO ) . p57Kip2 siRNA duplexes of the following RNA sequences were used: 5’-GUGCUGAGCCGGGUGAUGATT-3’; 5’-UCAUCACCCGGCUCAGCACTT-3’ . AllStars Negative Control siRNA ( Qiagen , Hilden , Germany ) was used for the mock transfection control . Approximately 1 . 0 × 107 proliferating SCs for each antibody were fixed with 1% formaldehyde at room temperature for 10 minutes . The cell lysates were sonicated with a Covaris S2 sonicator to shear DNA . Dynabeads Protein A ( Invitrogen , Carlsbad , CA ) conjugated with 10 μg of each primary antibody was added , followed by incubation at 4°C overnight . The beads were washed 5 times with RIPA buffer ( 0 . 2% NP-40 , 0 . 2% Na-deoxycholate , 0 . 16 M LiCl , 10 mM EDTA , 20 mM HEPES-KOH , pH 7 . 6 ) and eluted with elution buffer ( 1% SDS , 50 mM EDTA , 100 mM Tris-HCl , pH 8 . 0 ) . The eluate was incubated at 65°C overnight to reverse the crosslinking , followed by incubation at 55°C for 1 hour in the presence of proteinase K . DNA was purified using a MinElute PCR Purification Kit ( Qiagen , Hilden , Germany ) and quantified by real-time PCR ( Thermal Cycler Dice Real Time System II ( Takara Bio , Japan ) ) . All primer sequences are listed in S1 Table . | How muscle homeostasis is maintained is not completely elucidated yet . Epigenetic disorders such as Beckwith-Wiedemann syndrome , which causes hypergrowth of skeletal muscles and rhabdomyosarcoma , indicate that epigenetic regulations such as DNA methylation , contribute to this homeostasis control . DNA methylation is mediated by DNA methyltransferases , such as Dnmt3a and Dnmt3b , which are de novo DNA methyltransferases . The role of DNA methylation in somatic stem cells is not completely understood , although it has been shown to be indispensable in differentiation of primordial germ cells and embryonic stem cells . In this report , we investigated the role of Dnmt3a in muscle satellite cells by analyzing Dnmt3a-conditional knockout ( cKO ) mice in which Dnmt3a loci are deleted utilizing Cre-recombinase driven by Pax7 or Pax3 promoters that are specifically activated in the muscle precursor lineage . The loss of Dnmt3a in cKO mice causes decreased muscle mass and significantly impaired muscle regeneration . Moreover , Dnmt3a loss also results in a striking loss of proliferation of SCs , which is caused by mis-expression of a cyclin-dependent kinase inhibitor , p57Kip2 . Therefore , our findings suggest that DNA methylation plays an essential role in muscle homeostasis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"muscle",
"tissue",
"skeletal",
"muscles",
"cell",
"differentiation",
"muscle",
"regeneration",
"developmental",
"biology",
"organism",
"development",
"epigenetics",
"dna",
"morphogenesis",
"dna",
"methylation",
"chromatin",
"musculoskeletal",
"system",
"chromosome",
"biology",
"muscle",
"differentiation",
"animal",
"cells",
"gene",
"expression",
"biological",
"tissue",
"chromatin",
"modification",
"dna",
"modification",
"muscles",
"muscle",
"cells",
"biochemistry",
"regeneration",
"cell",
"biology",
"nucleic",
"acids",
"anatomy",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
]
| 2016 | Dnmt3a Regulates Proliferation of Muscle Satellite Cells via p57Kip2 |
Human lymphatic filariasis is a mosquito-vectored disease caused by the nematode parasites Wuchereria bancrofti , Brugia malayi and Brugia timori . These are relatively large roundworms that can cause considerable damage in compatible mosquito vectors . In order to assess how mosquitoes respond to infection in compatible mosquito-filarial worm associations , microarray analysis was used to evaluate transcriptome changes in Aedes aegypti at various times during B . malayi development . Changes in transcript abundance in response to the different stages of B . malayi infection were diverse . At the early stages of midgut and thoracic muscle cell penetration , a greater number of genes were repressed compared to those that were induced ( 20 vs . 8 ) . The non-feeding , intracellular first-stage larvae elicited few differences , with 4 transcripts showing an increased and 9 a decreased abundance relative to controls . Several cecropin transcripts increased in abundance after parasites molted to second-stage larvae . However , the greatest number of transcripts changed in abundance after larvae molted to third-stage larvae and migrated to the head and proboscis ( 120 induced , 38 repressed ) , including a large number of putative , immunity-related genes ( ∼13% of genes with predicted functions ) . To test whether the innate immune system of mosquitoes was capable of modulating permissiveness to the parasite , we activated the Toll and Imd pathway controlled rel family transcription factors Rel1 and Rel2 ( by RNA interference knockdown of the pathway's negative regulators Cactus and Caspar ) during the early stages of infection with B . malayi . The activation of either of these immune signaling pathways , or knockdown of the Toll pathway , did not affect B . malayi in Ae . aegypti . The possibility of LF parasites evading mosquito immune responses during successful development is discussed .
It is estimated that 120 million people are infected with Wuchereria bancrofti , Brugia malayi , or B . timori , the mosquito-transmitted , parasitic nematodes that cause human lymphatic filariasis ( LF ) . In approximately 40% of cases , the disease is manifested by lymphedema of the extremities or hydrocoele . Although human LF does not increase mortality in endemic areas , morbidity causes major economic losses and often leads to psychosocial and psychosexual conditions in infected individuals [1] . Recent efforts by the Global Program for the Elimination of Lymphatic Filariasis ( GPELF ) have decreased the numbers of individuals infected with , and at risk for , this parasitic disease [2] . Several different mosquito species within the genera Culex , Anopheles , Aedes and Mansonia can serve as primary vectors of LF parasites . The geographical location and habitat type influence which mosquito species function as vectors in any particular endemic area . Biological transmission of filarial worms is termed cyclodevelopmental , i . e . , the parasite undergoes development within the vector to become infective to the vertebrate host , but does not multiply . In competent vectors , microfilariae ( mf ) , produced by adult female worms and found circulating in the peripheral blood , are ingested with a blood meal and will quickly ( within 2 hr ) penetrate the midgut epithelium to access the hemocoel [3] . Mf migrate in the mosquito's hemolymph to reach the thoracic musculature and from there penetrate into the indirect flight muscles . This tissue is the site of development , where mf undergo two molts and emerge as infective-stage larvae ( L3s ) . Approximately eight days after exposure , L3s migrate to the head and proboscis from where they escape by penetrating the labellum of the proboscis when the mosquito takes a blood meal . Within the human host , the parasites undergo two additional molts and grow as they migrate to lymphatic vessels where adult male and female worms mate and females give birth to mf . Mf then make their way into the circulating blood from where they can be ingested by another blood feeding mosquito . LF parasites grow nearly seven times in length ( B . malayi grow from ∼200 to ∼1 , 350 µm in length , from mf to L3s respectively ) during the extrinsic developmental period within the mosquito [4] . As parasites develop , the mosquito must tolerate a series of insults due to parasite activities , e . g . , migrating mf damage both midgut [5] and muscle cells as they penetrate through or into them; second-stage larvae ( L2s ) actively ingest mosquito cellular components; L3s are very large and migrate out of the thoracic muscles and through the body cavity to reach the head and proboscis . Ultrastructural studies of Aedes mosquitoes infected with Brugia parasites have revealed that nuclear enlargement ( a sign of a putative repair response ) occurs in both Brugia-infected and neighboring non-infected muscle cells , and that complete degeneration of infected muscle cells occurs once L3s exit the flight muscles [6] . Other studies have shown that mosquito flight muscle cells become devoid of glycogen granules following infection with Brugia parasites [7] , [8] . Considering the amount of tissue damage observed in muscle cells , it is not surprising that Brugia-infected mosquitoes are known to have decreased flight activity and longevity [9] , [10] . But the successful development ( and subsequent transmission ) of LF parasites depends on the ability of competent mosquito vectors to survive infection . Some mosquitoes are able to limit or prevent filarial worm infections with various refractory or resistance mechanisms . For example , mf can be damaged during ingestion by an armed pharyngeal and/or cibarial pump ( often found in Anopheles spp . ) , inhibiting them from penetrating the midgut wall [11] , [12] . Mf that successfully penetrate the midgut and enter the hemocoel come in contact with hemolymph components , including circulating blood cells called hemocytes . Melanotic encapsulation is a hemocyte-mediated , innate immune response that can be very specific and robust , and can limit or prevent parasite development in some mosquito species [13] . In contrast , the involvement of a humoral immune response is not well understood in many compatible filarial worm-mosquito systems . Some parasites are able to evade or suppress a host's immune system in order to survive , but it is unknown if such interactions occur between LF parasites and mosquitoes [14] . Previous studies have investigated the effect of an activated mosquito immune response on filarial worm development , but the results remain inconclusive . When bacteria are inoculated into the mosquito hemocoel , to induce the expression of infection responsive immune factors prior to filarial worm exposure , a reduced B . malayi prevalence and mean intensity in Ae . aegypti was observed as compared to non-inoculated controls [15] . However , when the same bacterial strains and inoculation procedures were used to pre-activate the immune response of Culex pipiens prior to W . bancrofti exposure , there was no difference in prevalence or mean intensity between bacteria-inoculated and control mosquitoes [16] . Further investigation is needed to assess the effects of an activated mosquito immune response on a LF parasite infection . The role of Toll and Imd signaling pathways in the immune recognition , modulation , and response of mosquitoes to LF parasites has yet to be examined . Recently , Xi et al . [17] developed methods to manipulate these immune signaling pathways in Ae . aegypti by ( 1 ) gene silencing of Cactus , a negative regulator of the Toll pathway , ( 2 ) gene silencing of MyD88 , an adaptor required for endogenous Toll pathway signal transduction , and ( 3 ) gene silencing of Caspar , a negative regulator of the Imd pathway . Using these tools , the Toll and Imd pathways are transiently relieved from endogenous suppression ( 1 and 3 above ) or made unresponsive to detected stimuli ( 2 above ) . In this study we assess transcriptome changes associated with the development of B . malayi in Ae . aegypti and investigate the effect of the immune signaling pathways , Toll and Imd , on parasite development .
Aedes aegypti black-eyed , Liverpool ( LVP ) strain used in this study were maintained at the University of Wisconsin-Madison as previously described [18] . Briefly , mosquitoes were maintained on 0 . 3 M sucrose in an environmental chamber at 26 . 5±1°C , 75±10% RH , and with a 16 hr light and 8 hr dark photoperiod with a 90 minute crepuscular period at the beginning and end of each light period . Ae . aegypti LVP was originally selected for susceptibility to Brugia malayi by Macdonald in 1962 . This strain supports the development of mf to L3s . Four- to five-day-old mosquitoes were sucrose starved for 14 to 16 hours prior to blood feeding . Mosquitoes were exposed to B . malayi ( originally obtained from the University of Georgia NIH/NIAD Filariasis Research Reagent Repository Center ) by feeding on ketamine/xylazine anesthetized , dark-clawed Mongolian gerbils , Meriones unguiculatus . The same animals were used for all three biological replicates . Microfilaremias were determined , using blood from orbital punctures , immediately before each feeding and ranged from 50–150 mf per 20 µl of blood . Control mosquitoes were exposed to anesthetized , uninfected gerbils . Mosquitoes that fed to repletion were separated into cartons and maintained on 0 . 3 M sucrose in the laboratory . In early stages of development ( 1 h to 3 d post-infection [PI] ) , individual mosquitoes were separated into midgut , thorax and abdomen ( with midgut removed ) and dissected in Aedes saline [19] . Tissue dissections were cover-slipped and parasites were observed with a compound microscope using phase-contrast optics . The same procedure was used for dissections at 5–6 d PI , except only thoraces were dissected and examined . At 8–9 d PI , the thorax was processed as described above , but the abdomen and head & proboscis were dissected in separate drops of Aedes saline to observe L3s using a dissection microscope . Individual mosquitoes were dissected in a drop of Aedes saline for the recovery of L3s at 13–14 d PI . Images of B . malayi developmental stages were captured and processed as previously described , with the addition of Nomarski optics [20] . Five sample groups were created to study the transcriptional response of mosquitoes to B . malayi development . In each group , 20 mosquitoes were collected for RNA extraction . These sample groups are defined by the time after the blood meal and represent significantly different stages of parasite development . Briefly , Group 1 consisted of mosquitoes collected at 1 , 6 , 12 and 24 h PI . At these early time points , mf are penetrating the mosquito midgut , migrating through the hemocoel and penetrating thoracic muscle cells . Group 2 was collected at 2–3 d PI , a time when mf have differentiated into intracellular first-stage larvae ( L1s ) . At 5–6 d PI , B . malayi complete the molt to second-stage larvae ( L2s ) and actively feed on mosquito muscle tissue ( Group 3 ) . In Group 4 , at 8–9 d PI , parasite development is complete with a second molt to the L3s . Tissue damage continues as L3s break out of the thoracic muscles and migrate to the mosquito's head and proboscis . The final collection ( Group 5 ) , made at 13–14 d PI , occurs when the majority of L3s are located in the head and proboscis ( see [4] , [21] ) . Five mosquitoes , from both B . malayi-infected and uninfected blood meals , were collected at 1 , 6 , 12 , and 24 h PI and ten mosquitoes at 2 , 3 , 5 , 6 , 8 , 9 , 13 and 14 d PI for transcriptional analysis . Mosquitoes were pooled ( 5 mosquitoes/tube ) , flash frozen in liquid nitrogen and stored at −80°C prior to total RNA extraction using RNeasy ( QIAGEN ) . In addition , five B . malayi-infected mosquitoes were dissected to verify filarial worm infection and to determine the stage of parasite development at each time point . Three biological replications were completed . Microarray assays were conducted and analyzed as reported previously [17] , [22] . A full genome microarray platform ( Agilent; 4×44k ) was used with the probe sequences identical to the previous version ( 1×22k ) [22] . In brief , 2–3 µg total RNA was used for probe synthesis of Cy3- and Cy5-labeled dCTP . Hybridizations were conducted with an Agilent Technologies In Situ Hybridization kit at 60°C according to the manufacturer's instructions . Three independent biological replicate assays were performed . Hybridization intensities were determined with an Axon GenePix 4200AL scanner , and images were analyzed with Gene Pix software . To produce the expression data , the background-subtracted median fluorescent values were normalized according to a LOWESS normalization method to reduce dye-specific biases , and Cy5/Cy3 ratios from replicate assays were subjected to t-tests at a significance level of p<0 . 05 using TIGR MIDAS and MeV software [23] . Expression data from all replicate assays were averaged with the GEPAS microarray preprocessing software prior to logarithm ( base 2 ) transformation . Self-self hybridizations have been used to determine the cut-off value for the significance of gene abundance on these microarrays to 0 . 8 in log2 scale , which corresponds to 1 . 74- fold regulation [22] . For genes with P<0 . 05 , the average ratio was used as the final fold change; for genes with P>0 . 05 , the inconsistent probes ( with distance to the median of replicate probe ratios larger than 0 . 8 log2 ) were removed , and only the value from a gene with at least two replicates was further averaged . The robustness of these microarray gene expression assays were validated through qPCR ( Text S1 ) . Real time PCR assays were conducted as previously described to validate gene silencing efficiency and microarray expression data for selected genes [24] . Briefly , RNA samples were treated with Turbo DNAse ( Ambion , Austin , Texas , United States ) and reverse-transcribed using Superscript III ( Invitrogen , Carlsbad , California , United States ) with random hexamers . Transcript relative quantification was performed using the QuantiTect SYBR Green PCR kit ( Qiagen ) and ABI Detection System ABI Prism 7300 ( Applied Biosystems , Foster City , California , United States ) . qRT-PCR reactions were conducted using a 10 minute step at 94°C and 40 cycles of 15 seconds at 94°C , 15 seconds at 55°C and 15 seconds at 72°C . Three independent biological replicates were conducted and all PCR reactions were performed in triplicate . Transcript abundance was normalized against the ribosomal protein S7 gene . All primers used for qPCR assay are presented in Table S1 . The primer sequences used to verify gene knockdown efficiency included: S7 ( AAEL009496-RA ) forward: 5′-GGGACAAATCGGCCAGGCTATC-3′ , reverse: 5′- TCGTGGACGCTTCTGCTTGTTG-3′; Caspar ( AAEL003579-RA ) forward: 5′-GAATCCGAGCGAGCCGATGC-3′ , reverse: 5′-CGTAGTCCAGCGTTGTGAGGTC-3′; Cactus ( AAEL000709-RA ) forward: 5′-AGACAGCCGCACCTTCGATTCC-3′ , reverse: 5′-CGCTTCGGTAGCCTCGTGGATC-3′; MyD88 ( AAEL007768 ) forward: 5′-CATCCCATTCAGTTTCTCAGC-3′ , reverse: 5′-ACCGGTTGGAAGTTCTGATG-3′ . A complete list of PCR primer sequences is presented in Table S1 . RNAi was conducted by intrathoracic injection of dsRNA using described methodology [24] , [25] . Mosquitoes were three to four days old at the time of blood feeding and dsRNA was injected either 48 h before or after parasite exposure , i . e . , dsRNA injections were performed on non-blood fed one- to two-day-old mosquitoes and on blood fed , five- to six-day-old mosquitoes . Approximately 0 . 5 µl of dsRNAs ( 1 . 0 or 0 . 5 µg/µl ) were injected into the thorax of cold-anesthetized mosquitoes . The primers used to synthesize Cactus , Caspar and MyD88 dsRNA have been published previously [17] . To synthesize GFP dsRNA , methods described by Bartholomay et al . [25] were used with minor changes . The following sequences were annealed by heating at 95°C for 5 min and slow cooling: GFP_F 5′-TAGTACAACTACAACAGCCACAACGTCTATATCATGGCCGACAAGCAGA AGAACGGCATCAAGGTGAACTTCAAGATCCGCCACAACA-3′ and GFP_R: 5′-TCGATGTTGTGGCGGATCTTGAAGTTCACCTTGATGCCGTTCTTCTGCTTGTCGGCCATGATATAGACGTTGTGGCTGTTGTAGTTGTA-3′ ( Integrated DNA Technologies , Inc . , Coralville , IA ) . This dsDNA was ligated into pBlueScript KS+ ( Stratagene ) at XbaI ( T7 ) and SalI ( T3 ) sites . A complete list of PCR primer sequences is presented in Table S1 . B . malayi exposures were performed as described above and microfilaremias ranged from 35–162 mf per 20 µl of blood . Mosquitoes injected with GFP dsRNA were blood fed on each infected gerbil and used as a control for parasite infections . Mosquito mortality was observed every 24 h and mosquitoes were dissected at 6 d or 12–13 d PI to observe parasite development . To verify gene knockdown , five mosquitoes were collected 48 h post dsRNA injection , placed in a microcentrifuge tube , flash frozen and stored at −80°C until RNA extraction . Real-time PCR was used to quantify gene silencing efficiency . Silencing of Cactus , Caspar and MyD88 resulted in a reduction of mRNA levels by 60% , 84% and 27% , respectively . For each exposure , the prevalence and mean intensity of infection was calculated . Comparisons of mean intensities and mosquito mortality curves were done with the Mann-Whitney and Log-Rank tests , respectively , using GraphPad Prism 5 ( GraphPad Software , Inc . , La Jolla , CA ) . Results were considered significant at P≤0 . 05 .
The development of B . malayi was observed at 1 h to 14 d PI for each biological replicate and is summarized in Table S2 . Worms were recovered from 166 of the 180 mosquitoes examined for an overall infection prevalence of 92% . Mf were recovered from 1 h to 3 d PI , but after 24 h PI mf were no longer the most abundant developmental stage recovered . At 2 d PI , mf recovery began to decrease and almost all worms had differentiated into intracellular L1s by 3 d PI . Parasites molted to L2s in the thoracic musculature by 5 d PI , and were the only developmental stage identified in transcriptional group 3 . The molt from L2 to L3 occurred at 8–9 d PI . At 8 d PI , L2s and L3s were recovered in the thorax , and only 4% of the worms were located in the head and proboscis . In contrast , at 9 d PI , both L2s and L3s were observed in the thorax region , but the majority were L3s and 47% of all recovered worms were located in the head and proboscis . By 13–14 d PI , all parasites had developed into L3s . Images of B . malayi development , from mf to L3s , are presented in Figure 1 . The prevalence of L3s ( for all three biological replicates at 13–14 d PI ) was 80% ( n = 30 ) and the mean intensity was 6 . 9±5 . 7 . The Ae . aegypti global transcript responses to the successful development of B . malayi were determined using a genome microarray expression approach . These transcriptome infection-response patterns differed significantly , in both the number of regulated genes and their direction of regulation , at the different parasite development stages ( Fig . 2A ) . A general suppression of transcription was evident during the early stages of infection ( group 1 ) ; 20 genes were down-regulated while only 8 genes were up-regulated . However , this transcriptional suppression was reduced when parasites developed into L1s ( group 2 ) , and reversed when they became L2s ( group 3 ) ; 59 genes were up-regulated and only 5 genes were down-regulated at this stage . Parasite infection at later stages of infection mainly caused transcriptional up-regulation ( groups 4 and 5 ) . Strikingly , 120 genes were up-regulated in group 5 that represented mosquitoes in which L3s had migrated to the head and proboscis from where they can be transmitted to a host upon blood feeding . As many as 158 mosquito genes were differentially expressed at this late stage . The second largest number of infection-responsive genes was observed when L2s were present ( group 3 ) . Interestingly , the transcripts that were over represented in the groups that displayed the most prominent transcriptional regulation ( 3 and 5 ) , were highly enriched with putative immune genes ( Fig . 2 ) . Specifically , several antimicrobial peptide effector genes were strongly induced in group 3 and a number of putative pattern recognition receptor and signal modulator genes were up-regulated in group 5 ( Table 1 ) . Group 5 also contained several serine protease cascade components and putative melanization –related factor transcripts in increased abundance ( Tables 1 and 2 ) . To investigate whether the mosquito's two major immune signaling pathways , Toll and Imd , had any effect on B . malayi infection we used an established gene silencing approach to either simulate the activation of the Toll and Imd pathway , by depleting their negative regulators Cactus and Caspar , respectively , or inhibit the Toll pathway through the depletion of the MyD88 factor [17] , [26] . Caspar and MyD88 gene knockdown did not change the mortality rate in either pre- or post-bloodfed dsRNA-injected mosquito groups compared to their respective GFP dsRNA controls ( Log-rank tests , P-values ranging from 0 . 23 to 0 . 41; see Fig . 3 and 4 ) . In contrast , Cactus gene knockdown , that results in the activation of the Rel1 factor , resulted in a significantly increased mosquito mortality in both pre- and post-blood feeding Cactus dsRNA injected groups ( Log-rank test , P<0 . 001; data not shown ) . This increased mortality was not related to infection status ( Fig . 5A and B ) . There was no significant difference in B . malayi mean intensities between Caspar , Cactus or MyD88 silenced mosquitoes compared to the GFP dsRNA injected controls that had fed on the same microfilaremic gerbil . Likewise , there was no difference in prevalence or mean intensity of L3s between Caspar and MyD88 depleted mosquitoes as compared to their respective GFP dsRNA injected controls before ( Fig . 3B and 4B ) or after ( Fig . 3D and 4D ) blood feeding . Although knockdown of Cactus increased the mortality rate of Ae . aegypti , parasites that were recovered from live mosquitoes at six and 12 d PI had developed normally and there was no difference in the infection prevalence or mean intensity compared to the controls ( Figure 5 ) .
In this study , we provide insights into the interactions between filarial worms that cause human LF and their compatible mosquito vectors . As Brugia and Wuchereria parasites develop , the mosquito experiences a series of insults that include: ( 1 ) the penetration of cells and tissues by mf , ( 2 ) the consumption of cellular material by developing larvae , and ( 3 ) the migration of L3s through the body cavity . The infection response of mosquitoes is surprisingly diverse during the course of nematode development , as different gene transcripts and regulatory trends were observed in each of the five different developmental time points examined . By infection response we refer to the overall transcriptional and physiological change that occurs in the mosquito as a result of parasite infection , and it includes a vast array of distinct types of responses ( e . g . , repair , immune , metabolic , reproductive , behavioral , etc . ) . Overall , the response to filarial worm infection in this compatible system is mainly comprised of molecules involved in cellular signaling , proteolysis , stress response , transcriptional regulation , and repair ( see Table S3 ) . And these include several genes that have traditionally been classified as immunity related ( Table 1 ) . Very few transcriptional changes were observed until L2s were present , and the most profound transcriptional changes were observed in mosquitoes that harbored infective-stage parasites for 4–5 days . A large proportion of the regulated transcripts represented genes of unknown function ( 32 . 9% ) , and genes that have multiple or diverse functions ( 40% ) . As expected , the transcriptomic profiles of B . malayi-infected Ae . aegypti are very different than those previously described in B . malayi-infected Armigeres subalbatus ( see Table S4 ) [27] . In this non-compatible relationship , B . malayi development does not occur due to the rapid recognition and melanization of mf in the hemocoel of Ar . subalbatus [28] . The different transcriptional changes following infection of mosquitoes that support parasite development and those that do not can provide clues to the molecular mechanisms that determine compatible versus incompatible mosquito-filarial worm associations . Such comparisons have been made between Cx . pipiens and Ae . aegypti infected with W . bancrofti [29] , but transcriptional responses that occur in these mosquitoes may not represent genes that are used to deter filarial worm infection in an incompatible system , i . e . , it is quite possible that differences in gene transcription of mosquitoes in different genera could represent unique strategies for overcoming damage caused by filarial worms and therefore do not represent anti-filarial worm responses . Similarly , identification of immune-responsive genes activated in response to filarial worm infection does not indicate that the mosquito is/has mounted an immune response against the parasites itself [30] . It is possible that the observed response could be an indirect effect caused by the infection , i . e . , mf midgut penetration or muscle cell damage that occurs later in development in compatible mosquitoes . The current study provides transcriptional data from a strain of Ae . aegypti that is highly compatible with B . malayi , and can help guide the planning of future studies measuring transcriptional changes in a strain of Ae . aegypti that does not support the development of B . malayi . The differences in parasite size ( Fig . 1 ) and behavior among developmental stages provide a foundation for discussing the response of Ae . aegypti to a successful B . malayi infection . In mosquitoes sampled from 1–24 h PI , mf are in the process of penetrating the midgut epithelium and migrating through the hemocoel to penetrate the indirect flight muscle cells . Differentiation to L1s begins as soon as mf become intracellular parasites . At this early stage of infection , transcriptional profiles suggest the presence of B . malayi may alter blood digestion/proteolysis ( four serine proteases; sterol trafficking ) , chitin-related interactions ( two transcripts contain the chitin-binding Peritrophin-A domain; IPR002557 ) , and immune function ( DEF D; AAEL003857 ) . B . malayi-infected mosquitoes sampled at 2–3 d PI harbor L1s , a stage when parasites have a markedly decreased mobility within the indirect flight muscle cells and are in the process of developing a digestive track . The few infection-responsive transcripts ( 13 genes ) during infection with L1s include three down-regulated immunity-related genes: CEC F ( AAEL000625 ) , CEC A ( AAEL000627 ) , and a hypothetical protein ( AAEL003843 ) which is a putative knottin with an interesting genomic location; just upstream and on the opposite strand from DEF A ( AAEL003841 ) . The regulation of cecropin transcripts in response to pathogens is complex [31] , and the interpretation of their decreased transcriptional abundance therefore remains speculative . Mosquitoes sampled at 5–6 d PI contain worms that have molted to L2s . At this stage in development , the parasites are actively ingesting cellular material , have developed a digestive system ( with an open mouth but an anus still closed ) and have grown four times in length compared to mf . Even though filarial worms remain intracellular until the molt to L3s , these internally damaged cells are likely to provide the necessary stimuli for the mosquito's infection response . Genes involved in cellular signaling ( e . g . , G-protein coupled receptors , Spaetzle 5 , DSCAM ) and transcriptional regulation ( i . e . , changing patterns in transcription factors ) were identified as components of the infection response at this time interval ( Table S3 ) . Another component of the infection response to L2s is the increased abundance of six cecropin transcripts ( Table 1 ) . At 8–9 d PI , worms have molted to L3s and begun migrating from the thoracic musculature to the mosquito head and proboscis . In Ae . aegypti , the onset of severe muscle damage occurs when L3s have exited the infected muscle cells [6] . It has been noted that this muscle damage , and the subsequent damage repair response , also occurs in mosquitoes that are mechanically damaged by external thoracic punctures [32] . Considering these ultrastructural observations by Beckett et al . [6] , [32]–[34] , it is interesting to find that the infection response in this group involved only modest changes in transcript abundance . Of the 26 infection responsive transcripts identified during L3 migration , the majority are induced and putatively involved in stress response ( n = 4 ) and transcriptional regulation ( n = 4; see Table S3 ) . The most profound transcriptomic changes , in response to infection , occurred at 13–14 d PI , an infection stage when mosquitoes have harbored L3s for approximately 4–5 days . Many components of intra- and extracellular signaling pathways ( n = 12; see Table S3 ) were differentially transcribed in these infected mosquitoes , again lending support to the fact that detection and communication of stimuli is key to mounting a response and repairing tissues . L3-infected mosquitoes seem to be responding to tissue damage with 15 genes identified with possible functions in proteolysis and seven insect cuticle protein transcripts in increased abundance . There are multiple sources of tissue damage at this point in infection: ( 1 ) the degrading muscle cells that supported the development of the parasites [6] , ( 2 ) L3 migration throughout the mosquito body cavity [35] , and ( 3 ) the ability of L3s to penetrate through the cuticular surface [36] . The data also show that infected mosquitoes are responding to the stressful conditions of harboring L3s ( 9 stress response-related transcripts ) . Studies have shown that mosquito behavior can be modified by filarial worm infection , and occurs in an intensity-dependent manner [9] . Comparisons between an earlier study of spontaneous flight behavior changes in Brugia-infected Ae . aegypti [9] and transcriptional changes seen in the present study could be made for three of our five groups . The estimated mean intensities of Brugia infections in mosquitoes collected for transcriptional analysis fit within the categories of low ( 1–10 parasites ) to moderate ( 11–20 parasites ) intensities created by Berry et al . [9] . Changes in the transcriptome of Brugia-infected mosquitoes are associated with an increase in flight behavior during the time L2s are feeding ( Group 3; 5–6 d PI ) , a marked decrease followed by recovery of flight when L3s emerge and migrate from infected muscle cells ( Group 4; 8–9 d PI ) , and up to a 60% decrease in spontaneous flight activity when mosquitoes harbor Brugia L3s ( Group 5; 13 d PI ) . The data from ultrastructural [6] , [32]–[34] , behavioral [9] , and transcriptional observations ( presented herein ) all support the fact that filarial worm development is not a benign infection to mosquitoes . Certain mosquito-borne pathogens are known to be controlled by vector immune responses , which are regulated by intracellular signaling pathways , such as Toll and Imd . For example , the Toll and Imd pathways in An . gambiae regulates infection with malaria parasites ( Plasmodium berghei and P . falciparum ) and is required for antibacterial defenses [26] , [37] . Innate immune response in tsetse flies has also been implicated in regulating the intensity of trypanosome infection [38] . An immune response is considered a mechanism by which a host attempts to eliminate or reduce an infection . A host's immune response to parasitism may however not always lead to an elimination of parasites because of the latter's capacity to evade the immune defense mechanisms [39] . Previous studies in our laboratory suggest that LF parasites either elicit an immune response , e . g . , melanotic encapsulation , or go undetected and therefore unmolested by an immune response in certain mosquitoes [28] , [40] . Although the interactions between these nematodes and the mosquito immune system are mechanistically undescribed , there is potential for LF parasites to evade and/or suppress the mosquito immune system [41]–[44] . In this study , we manipulated the mosquito immune system in an effort to activate immune response pathways to determine what effects , if any , they might have on parasite infection and development . We used a RNAi–mediated gene knockdown approach to transiently activate the two major immune signaling pathways , Toll and Imd , by targeting their negative regulators , Cactus and Caspar , respectively . Post-transcriptional silencing of these pathway regulators leads to pathway-specific immune responses . Previous studies on the effect of mosquito immune responses on filarial worm development utilized bacterial challenges to activate the immune system [15] , [16] . We selected time points for the activation of immune pathways that would specifically target the parasites early in development , i . e . , when they might be most vulnerable to the mosquito immune system; when microfilariae migrate to the thoracic musculature and when parasites undergo the first molt ( first- to second-stage larvae ) . The activation of these immune pathways had no detectable effect on B . malayi development in Ae . aegypti ( Fig . 3–5 ) . The lack of an anti-parasite effect as a result of activating Toll and Imd pathways suggest that the parasite limiting mechanism , that was observed in bacterial-challenged Ae . aegypti , was not attributed to a Rel1 or Rel2 nuclear translocation . This may imply that the bacteria challenge induced some other defense system , independently of the Toll or Imd pathways , or that the bacteria exerted a direct anti-parasitic effect on the filarial worms . Infection of Ae . aegypti with a compatible filarial parasite , B . malayi , resulted in fairly few changes in the mosquito transcriptome; however , these infection responses were diverse and differed vastly between the different infection stages . The majority of these transcriptional infection responses are most likely a reflection of the mosquito's attempt to repair tissue damage resulting from nematode development . We have also shown that removing the inhibitors of Rel1 and Rel2 activation did not affect the permissiveness of this mosquito to B . malayi infection . This observation may indicate a resistance , immune evasion and/or suppression stategy ( ies ) by the parasite , whereby it remains inert to destruction by the mosquito's immune system . Not all mosquitoes respond the same to LF parasite infection , and differences between natural and artificial systems should be carefully considered . Aedes aegypti is a common laboratory vector that has been genetically selected for susceptibility to many pathogens , including B . malayi [45] , but is not a natural vector of LF parasites [46] . Investigations of flight muscle cell damage caused by developing B . malayi in natural and artificial vectors have concluded that tissue damage is more severe in Ae . aegypti compared to Mansonia uniformis , which is a natural vector [6] , [34] . This increased pathology occurs when L3s migrate out of the flight muscle cells , and is reflected by a spike in Ae . aegypti mortality [47] . It is apparent that Ae . aegypti may utilize different mechanism ( s ) for surviving infection , and future studies comparing the infection response of natural mosquito-LF parasite systems would allow a better assessment of these differences . As advancements are made within the field of lymphatic filariasis parasite-host interactions , it will be interesting to compare the infection responses of both the vertebrate and invertebrate hosts . Mf and L3s are the developmental stages transmitted between hosts , and are known to elicit a vertebrate immune response [48]–[50] . The short time interval between L3s escaping the mosquito and infecting the vertebrate host exemplifies the link between the two host environments . For example , the unknown mechanism ( s ) employed by L3s to suppress the infection response of vertebrates [51] might be functional before L3 escape and may therefore also act on the innate immune response of the mosquito . | Filarial worms that cause human lymphatic filariasis ( LF ) are transmitted by many species of mosquitoes . Within susceptible mosquitoes , Brugia malayi develop from microfilariae ( mf ) to infective-stage larvae ( L3s ) , in approximately eight days . These nematodes develop as intracellular parasites within mosquito flight muscle cells , in which they ingest cellular material and eventually cause cell death when L3s migrate to the mosquito's proboscis . We examined the effects of B . malayi parasitism on Aedes aegypti by analyzing changes in mosquito gene expression at different stages of parasite development . We found that a few genes were differentially expressed at the RNA level relative to non-infected controls . The majority of changes occurred at two time periods , when the filarial worms began feeding and when the L3s were in the head and proboscis . Many transcriptional changes in the later group concur with documented descriptions of tissue damage , clean-up and repair that occurs in mosquitoes infected with filarial worms . In addition , we activated two innate immunity signaling pathways and observed the effects on filarial worm development . B . malayi seems to be capable of evading these immune responses , because its development was not impeded by the activation of either the Toll or Imd signal pathways in Ae . aegypti . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"microbiology/immunity",
"to",
"infections",
"genetics",
"and",
"genomics/gene",
"expression",
"genetics",
"and",
"genomics/functional",
"genomics",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"immunology/innate",
"immunity",
"cell",
"biology/microbial",
"growth",
"and",
"development",
"microbiology/parasitology"
]
| 2009 | Mosquito Infection Responses to Developing Filarial Worms |
Sustained and coordinated vaccination efforts have brought polio eradication within reach . Anticipating the eradication of wild poliovirus ( WPV ) and the subsequent challenges in preventing its re-emergence , we look to the past to identify why polio rose to epidemic levels in the mid-20th century , and how WPV persisted over large geographic scales . We analyzed an extensive epidemiological dataset , spanning the 1930s to the 1950s and spatially replicated across each state in the United States , to glean insight into the drivers of polio’s historical expansion and the ecological mode of its persistence prior to vaccine introduction . We document a latitudinal gradient in polio’s seasonality . Additionally , we fitted and validated mechanistic transmission models to data from each US state independently . The fitted models revealed that: ( 1 ) polio persistence was the product of a dynamic mosaic of source and sink populations; ( 2 ) geographic heterogeneity of seasonal transmission conditions account for the latitudinal structure of polio epidemics; ( 3 ) contrary to the prevailing “disease of development” hypothesis , our analyses demonstrate that polio’s historical expansion was straightforwardly explained by demographic trends rather than improvements in sanitation and hygiene; and ( 4 ) the absence of clinical disease is not a reliable indicator of polio transmission , because widespread polio transmission was likely in the multiyear absence of clinical disease . As the world edges closer to global polio eradication and continues the strategic withdrawal of the Oral Polio Vaccine ( OPV ) , the regular identification of , and rapid response to , these silent chains of transmission is of the utmost importance .
Poliovirus , like other members of Picornaviridae , usually generates mildly symptomatic infection . However , the clinical manifestation of polio , Acute Flaccid Paralysis ( AFP ) , can result when the virus invades the central nervous system [1] . Wild poliovirus ( WPV ) is transmitted fecal–orally and in the Northern Hemisphere exhibits seasonal epidemics in late summer and autumn [1–3] . Polio outbreaks continue today within this narrow seasonal window in Pakistan and Afghanistan [4 , 5] , but the seasonal transmission structure of polio remains unexplored . Propelled by public support , the race for the polio vaccine during the post-World War II era led to the development of the Inactivated Polio Vaccine ( IPV ) and the Oral Polio Vaccine ( OPV ) , which reduced the global incidence to less than 0 . 1% of prevaccine levels [6] . Missing the 2014 goal of globally stopping WPV transmission has left eradication elusive , primarily because of political and social obstacles for effective vaccine distribution , including vaccine hesitancy and mistrust . In light of this—and the call for innovative solutions [7]—an understanding of polio’s ecology can help guide alternative strategies . Looking toward eradication and beyond , a polio-free world requires an understanding of the mode by which polio originally emerged and historically persisted . We contend that a retrospective study of the ecology of WPV in the absence of vaccine interventions can inform future planning and may pinpoint vulnerabilities in WPV’s epidemiology that could be leveraged for eradication . Ironically , because of the success of polio vaccination , critical features of WPV transmission remain obscure . The low global incidence of polio ( due to high vaccine coverage ) , in combination with the relative rarity of symptomatic infections , limits the amount of epidemiological data with which to study transmission . Furthermore , data limitations regarding vaccine coverage in developing countries confound transmission studies , making it difficult to disentangle the effects of the vaccines , demography , and transmission . Therefore , we took advantage of a dataset of unprecedented size and resolution in both space and time to gain insights into the drivers of polio’s historical expansion and the ecological mode of its persistence in the prevaccine period . We present analyses of spatially-replicated incidence reports from the prevaccine era in the United States and built mechanistic transmission models that incorporate these data to reconstruct the unobservable infection dynamics . Our analyses allow us to dissect three axes of polio epidemiology: ( i ) geographical and seasonal variation in transmission , ( ii ) the role of demography in determining incidence , and ( iii ) the mode by which polio persists .
We examined monthly polio case reports ( January 1931–December 1954 ) from the US Public Health Service Morbidity and Mortality Weekly Reports as compiled by [8] and the CDC for each of the 48 contiguous US states and the District of Columbia ( Fig 1A and 1B ) ; data provided in the Supporting Information . Prior to 1945 , the cases in these data were predominantly paralytic [1 , 3]; however , during the later period of this study , nonparalytic cases comprised more than 40% of reported cases in populous cities such as New York , Detroit , Kansas City , and Sacramento [9] . In addition to polio case data , we obtained numbers of births by state from 1931 onward from the US Vital Statistics and state population sizes from the Population Distribution Branch of the US Census Bureau . Data from Vital Statistics are housed in the CDC online repository: National Center for Health Statistics , Products , Vital Statistics . The Census Bureau data were obtained from their Population Estimates Repository , historical data pre-1980; data also provided in the Supporting Information . The polio dataset—with cases detailed weekly—has now been independently digitized and is freely available and maintained online through the University of Pittsburgh Project TYCHO . Birth data were not available for Texas and South Dakota beginning in 1931 but began in 1932 and 1933 , respectively . For exploratory analyses , we quantified the relationship between disease fadeouts and population size . A threshold of 3 mo without a reported infection was chosen to define a fadeout [10] . The portion of fadeout months was taken as the ratio of fadeout months to total months in Fig 1C . To estimate spatial synchrony , we used the nonparametric spatial correlation function [11 , 12] . To measure the relative timing of polio epidemic peaks for each state and each year , the 1 yr wavelet band phase angle was computed [13] and used to rank states earliest to latest based on their epidemic peak timing . We constructed a dynamic stochastic model with components incorporating polio transmission , immunity , seasonality , and symptomatology along with empirical population sizes and birth rates . Birth rates displayed prominent seasonal , secular , and geographical trends ( Fig S4 in S1 Text ) [14] . We utilized Partially Observed Markov Process ( POMP ) models , which are suited for dealing with epidemiological data where the state variables ( susceptible , infected , and recovered individuals ) were not observed in the data; rather , the infected individuals were partially observed through clinical case reports . For our process models , we used seasonally-forced stochastic monthly discrete-time SIR models , where transitions followed a Poisson process . The infectious period was fixed at 1 mo , because multiple studies have found the duration of shedding to be 3–4 wk [15] . Infection-derived immunity was assumed to be lifelong [16 , 17] . The models contained six classes of infants susceptible ( SiB ) to infection . These infant classes contained 0–1-month-olds , 1–2-month-olds , etc . , up to 6-month-olds . Models had a single infected class for infants ( IB ) . The older age class , which contained individuals more than 6 months of age , had its own susceptible ( S0 ) and infected class ( I0 ) . The onset of polio symptoms ranges from 5–35 d postexposure , with a mean of 12 d [18]; therefore , we assumed reporting of symptomatic infections occurred within the 1 mo infectious period . We modeled polio reporting explicitly and , consistent with clinical evidence , assumed that maternal antibodies protected from severe disease and resulted in unreported infant infections [19–23] . Thus , we assumed that infections in individuals under 6 months of age were asymptomatic , and only individuals over 6 months of age could be symptomatic and reported as a clinical case . See model schematic in Fig 2A . The force of infection was modeled as , λt= ( βtItO+ItBNt+ψ ) εt . ( 1 ) The first term of the force of infection , βtItO+ItBNt , represents transmission that occurred locally by individuals infected in the state at time t . The second term , ψ , encompasses WPV infection from external sources that were divorced from the local infection dynamics . ψ placed a lower bound on the force of infection , allowing WPV to rebound in the face of local extinction . We interpret ψ as indicating WPV imported from other geographic regions; however , it could also be interpreted as representing a small number of individuals in the population that shed WPV for an extended period or environmental sources helping WPV persist over the winter . The transmission parameter , βt , was parameterized using a B-spline , providing it the flexibility to have a constant or seasonal transmission rate . There was seasonality , but no interannual variation , in the transmission rate , βt=exp∑i=16qiξit . ( 2 ) Here , each ξit is a periodic B-spline basis with a 1 y period . The process noise , εt , was gamma distributed with mean 1 and variance that scaled to account for both environmental and demographic stochasticity; refer to S1 Text , Equation S6 for further details . We assumed cases were drawn from a rounded , left-censored normal distribution with a mean report rate of ρt and dispersion parameter τ , casest=round ( xt ) , xt~normal ( ρtItO , τItO ) . ( 3 ) For calculating the likelihood , we used a binned-normal probability density . Full model details are found in S1 Text , Section S1 . 3 . We fitted SIR models ( one for each state in the US ) to data independently using Maximization by Iterated particle Filtering ( MIF ) in the R package pomp [24–26] . For each state , we estimated 14–15 parameters . The parameters estimated were: 6 seasonal transmission parameters ( βi ) , 3 parameters accounting for process and measurement noise , 3 initial conditions for the older age class , the external contribution to the force of infection ( ψ ) , and 1–2 report rates ( ρt ) . MIF is a simulation-based likelihood method for parameter estimation . The basis of MIF is particle filtering , which integrates state variables of a stochastic system and estimates the likelihood for fixed parameters . Instead of fixing parameters , MIF varies them throughout the filtering process and selectively propagates particles ( i . e . , parameter sets ) that have the highest likelihoods . By initializing MIF at a variety of points distributed across parameter space , we estimated the shape of the likelihood surface for each US state and identified the Maximum Likelihood parameter Estimates ( MLEs ) . MIF was initialized from 1 million parameter sets for a global search , followed by additional phases of increasingly localized searches , which included profiling . In total , for each US state , MIF was initialized from more than 10 , 000 locations in parameter space to estimate the shape of the likelihood surface and identify the MLEs . Prior to 1945 , nonparalytic polio cases were rarely included in our data , but the reporting of nonparalytic polio became increasingly common [1 , 3] . Thus , we tested an optional parameter to account for increased representation of nonparalytic polio in clinical cases data . We estimated two report rates , one for the pre-baby boom era and another for the baby boom era , and discriminated between models with and without time-varying reporting using Akaike Information Criterion ( AIC ) . Profiles were constructed for the two versions of the model , one in which the report rate was constant through the entire time period and one in which the report rate increased during the baby boom era . For each state , AIC was used to discriminate between constant and time-varying reporting , and the MLEs were drawn from the appropriate two-dimensional profile . Inference was performed using the data from May 1932 to January 1953 , with the exception of South Dakota and Texas , for which the first data used were from May 1934 , and May 1935 , respectively , i . e . , May of the year following the first full year of available data , the lag being needed to estimate initial conditions for the infant classes directly from birth data . For model validation , the last two epidemic years were set aside for forecasting . Full details are provided in S1 Text , Section S1 . 4 . Likelihood profiles were constructed for each US state ( example in Fig 2B , all others in Fig S9–S17 in S1 Text ) . To quantify model–data agreement , we evaluated the accuracy of one-step-ahead predictions for all 49 states , both for data used in model parameterization ( Fig 3B ) and for out-of-fit data ( Fig 3C ) . Because of correlations between states ( which vary significantly in size and mean incidence ) , simple linear regression is not appropriate for assessing model–data agreement; therefore , generalized R2 was calculated to quantify the proportion of the variance explained by the model relative to that explained by state alone . We calculated the generalized R2 for the one-step-ahead predictions and out-of-fit predictions ( See S1 Text , Section S2 . 2 for details ) . For Fig 4A and 4B , infections were reconstructed using particle filtering means , and the reconstruction was limited to data beginning in Jan 1935 , because 1935 is the first full year for which we have the models parameterized for all states . Following model validation and infection reconstruction , the fitted models were used as simulation tools to explore polio infection dynamics . In Fig 4D and 4E , we used 500 simulations for each state from 1935 through 1954 . In Fig 4D , we present the state-specific probability of extinction by examining 500 realizations of the fitted models . Specifically , we calculated the annual probability of polio extirpation during the off-season ( December–May ) and averaged across years . Similarly , in Fig 4E the minimum number of infections during each off-season was based on 500 simulations . For each simulation , the annual minimum number of infections was identified , and the median was taken across the 500 simulations and averaged across years . In order to identify the covariates and epidemiological parameters that influenced the number of trough infections—a measure of WPV persistence—we regressed trough infections with various covariates and parameters; results shown in Fig 5 . In Fig 6B–6D , distributions were generated by characterizing observations across 500 simulations per state . All simulations and data used for producing the figures in this manuscript are available in S1 Data–S14 Data .
In the mid-20th century , polio outbreaks in the US were strongly seasonal . Epidemic peaks typically occurred between August–October ( Fig S1 in S1 Text ) ; but the magnitude was highly variable among states . In the transition from the pre-baby boom era ( 1931–1945 ) to the baby boom ( 1946–1954 ) , epidemics increased in size and became more regular ( Fig 1A and 1B ) . Winter troughs were frequently marked by consecutive months without reported cases . During the baby boom , the frequency of these local fadeouts diminished ( Fig 1C ) , while epidemics became more tightly synchronized ( Fig 1D ) . There was a striking latitudinal gradient in the timing of epidemics across the entire country ( Fig 1E and 1F , Fig S1 in S1 Text ) . Two broad classes of mechanisms can give rise to such a pattern . Seasonal movement of the pathogen from southern populations can generate a traveling wave , which has previously been observed in measles [27] , dengue [28] , influenza [29 , 30] , and pertussis [31] . Alternatively , the pattern may indicate latitudinal gradients in demographics ( e . g . , birth rates [14 , 32] ) and/or environmental factors associated with transmission . Our extensive search of parameter space resulted in the MLEs for each parameter . To quantify the shape of likelihood surface along two parameter dimensions we identified as important ( i . e . , the report rate , ρt , and the external contribution to the force of infection , ψ ) , we constructed two-dimensional likelihood profiles for each US state . Two-dimensional profiles , by definition , have fixed parameter values along two dimensions of parameter space , while the likelihood is maximized along all other parameter dimensions . There were 12 states that had constant reporting ( i . e . , the same report rate during the pre-baby boom and baby boom era ) . Fig 3A illustrates that the fitted models generate epidemic trajectories that display: ( 1 ) the seasonal characteristics of polio and ( 2 ) the large amount of interannual variation in epidemic size . Importantly , the fitted models faithfully reproduce observed dynamics . In particular , the seasonality , epidemic shape , interannual variability in epidemic magnitude , and the increase in incidence during the baby boom are captured by the models . Model fit was formally validated using one-step-ahead predictions for all 49 states ( Fig 3B ) and out-of-fit predictions ( Fig 3C ) , which indicate good agreement between models and data . Furthermore , geographical structure in the timing of observed epidemics is captured by the fitted models ( Fig 3D ) . State-specific examples of one-step-ahead predictions and out-of-fit predictions are shown in Fig S2 and S3 in S1 Text . We hypothesized that the latitudinal gradient in epidemic timing was driven by either: ( 1 ) geographic variation in transmission because of environmental factors that modulated transmission , ( 2 ) the geographic trend in birth seasonality in the US ( detailed in [14] ) , or ( 3 ) the movement of pathogen from south to north . In support of hypothesis 1 ( i . e . , environmental factors ) , we identified a spatial pattern in the phase of seasonal transmission ( Fig 3E–3G , Fig S7 in S1 Text ) . States with earlier epidemics had an earlier peak in the seasonal transmission rate in the fitted models . Interestingly , because of polio’s long infectious period , peaks in transmission preceded incidence peaks by 1–2 mo . States varied geographically not only in the timing of the transmission peak but also in the wintertime transmission trough depth and trough duration ( Fig 3G ) . Epidemiological theory indicates that birth seasonality can have important dynamical consequences for childhood diseases [14 , 33 , 34] . To test hypothesis 2 ( i . e . , birth seasonality ) , we carried out a comparison of the fitted models with and without birth seasonality . Simulations of both models expressed the latitudinal gradient ( Fig S5 in S1 Text ) . Therefore , birth seasonality is not necessary to explain the polio gradient because geographic variation in transmission is sufficient . We attribute the negligible effect of birth seasonality on polio incidence to the low amplitude of birth seasonality , which was approximately 10% in the US at this time . We suggest that hypothesis 3 ( i . e . , pathogen movement ) is an unlikely explanation of the latitudinal gradient . If the latitudinal gradient were a wave of pathogen movement , it would require a high wave speed , which we see as incompatible with transport of the pathogen across the landscape . The pattern in Fig 1E and 1F corresponds to a wave traveling approximately 1 , 200 km/mo . For comparison , waves in pertussis have been estimated to travel 110–320 km/mo [31]; waves in dengue appear to move 150 km/mo [28]; and the measles wave speed in the United Kingdom was estimated at 20 km/mo [27] . A polio wave that is 10-fold faster than pertussis in the US is difficult to justify and unnecessary , because our fitted models support hypothesis 1 . Thus , we have determined that polio’s latitudinal gradient is driven by geographic variation in transmission; we are left with an unidentified seasonal driver that modulates transmission . While geographical variation in birth seasonality was insufficient to explain the latitudinal gradient seen in epidemic timing , birth seasonality had a small but observable effect on the simulated incidence of infant infections . To quantify the influence of birth seasonality on infant infections , we compared simulations of the fitted models to simulations for which seasonal fluctuations in births were removed . In the presence of birth seasonality , infant infection incidence was often higher ( Fig S6 in S1 Text ) ; however , this did not affect the incidence of disease directly , and no indirect effect was observed . It is well known that AFP incidence represents a small fraction of true WPV prevalence [35 , 36] . Reassuringly , our independent estimates from the incidence data agree: our MLEs indicate that typically less than 1% of polio infections were reported . We assumed that infected infants under 6 months of age were asymptomatic , due to protection by polio-specific maternal antibodies . The report rate for individuals not maternally-protected was 0 . 75% ( averaged across states ) in the pre-baby boom era and rose to 1 . 4% in the baby boom era , with considerable variation across states ( Fig S8 in S1 Text ) . Overall , we estimate that there were often over 1 million annual infections in the US; though only 2 , 000–57 , 000 cases were reported every year ( Fig 4A ) . Our results are in line with a 1948 serology-based study in North Carolina , which estimated 62–175 subclinical polio infections per paralytic case [37] . The fitted models revealed vast seasonal and spatial heterogeneity in WPV’s reproductive ratio . Fig 3G shows large seasonal fluctuations in the reproductive ratio within each state . Several states maintained a reproductive ratio above 1 throughout the year . In contrast , 28 states had reproductive ratios that fell below 1 for 4–5 mo from December–April . States in the Northeast and Midwest had extreme seasonal variation in their reproductive ratio . Deep winter troughs in transmission in the Northeast and Midwest often had several consecutive months with a reproductive ratio below 1 . In contrast , at the peak of transmission in June and July , these same states had a reproductive ratio above 20 . Interestingly , each geographic region other than the Midwest had at least one state that maintained a reproductive number above 1 throughout the year . Southern states typically maintained an intermediate transmission rate throughout the year . Our analyses provide a new perspective on polio’s historical emergence . Commonly described as a “disease of development , ” polio’s emergence has been ascribed to improved hygiene that reduced transmission and pushed the burden of infection onto children more susceptible to paralytic polio . This explanation requires that reduced transmission raised the mean age of infection and therefore the risk of AFP [1] . Our results suggest the marked increase in polio incidence from the 1930s to the 1950s was a straightforward consequence of increased birth rates ( Fig 4B and 4C ) , and that hygiene effects on transmission are not required to explain polio’s rise to epidemic levels . Since polio’s epidemic emergence was captured in the models as a consequence of the changing birth rate , we did not explicitly test reductions in the transmission rate as an additional contributor to epidemic size , and we cannot completely rule out trends in transmission as a contributing factor . While the “disease of development” explanation has also been questioned on other grounds [22] , changes in hygiene and sanitation could have contributed to the initial emergence of polio , which occurred from the late 1800s to the early 20th century . Polio cases were consistently observed throughout the US during the period of this study . We hypothesized: ( a ) WPV persisted locally in each state , or alternatively , ( b ) WPV regularly went locally extinct and reinvaded from elsewhere . Due to polio’s high asymptomatic infection ratio , distinguishing between these two mechanisms of persistence cannot be done using reported cases alone , since WPV may be present during the off-season even in the absence of clinical cases . In order to determine which of these two persistence mechanisms was the likely explanation of continued infection , we simulated the fitted models and characterized the dynamics of the process models ( i . e . , the unobserved infection dynamics rather than the observable disease dynamics ) . We focused on determining whether infections persisted during the wintertime off-season or if extinction and reinvasion occurred . In particular , we assessed ( i ) the average annual probability of an extinction event in each state , which results from diminished local transmission and ( ii ) the annual minimum number of infections . Fig 4D and 4E depict the geographic variation in these quantities . Some states experienced frequent local extinction during the off-season , followed by recolonization; we consider these “sink” populations . In contrast to sink states , a few states maintained infections year-round; these we define as “source” populations . The majority of states , however , were neither consistently sources nor sinks , because even sink states had frequent overwintering of WPV . The fitted models suggest that WPV underwent extinction and recolonization in the classic metapopulation sense . We explored characteristics that contributed to states having been WPV sources versus sinks . We used simulated trough infections , shown in Fig 4E , as the indicator of a source versus a sink . States that maintained a high number of trough infections enabled WPV to persist through the off-season; whereas states with a low number of trough infections were likely to have experienced regular WPV extinction . State population size accounted for 65% of the variation in the number of trough infections ( Fig 5A ) . We used multiple regression models to determine whether the ( i ) mean birth rate , ( ii ) amplitude of birth seasonality , ( iii ) immigration rate , ( iv ) seasonal minimum reproductive ratio , and/or ( v ) seasonal amplitude of the reproductive ratio explained the residual variation in trough infections , after controlling for population size . The mean birth rate and amplitude of birth seasonality had a negligible impact on the residual variation in trough infections; therefore , they were removed from the multiple regression model . A multiple regression model with the immigration rate , seasonal minimum reproductive ratio , and the seasonal amplitude of the reproductive ratio explained 56% of the residual variation in trough infections ( Fig 5B ) . Interestingly , even though there were no clear geographic patterns of source-versus-sink localization ( Fig 4D and 4E ) , there was strong geographic clustering in the minimum reproductive ratio ( Fig 5C ) , demonstrating that even though source-sink predictors display geographic clustering , the combination of predictors can generate a source-sink mosaic . We found that after accounting for population size , states with a higher immigration rate had more trough infections ( Fig 5D and 5E ) . States with a higher transmission amplitude , however , had fewer trough infections; we interpret this as being due to susceptible depletion followed by deep infection troughs in states with a high transmission amplitude ( Fig 5D ) . The minimum reproductive ratio had a positive relationship with trough infections; states that maintained a reproductive ratio above 1 during the off-season tended to have more trough infections during the off-season ( Fig 5E ) . Disease eradication programs face the significant challenge of verifying success in the presence of asymptomatic infections . Typically , a criterion for success is the absence of disease for an extended period; however , the utility of this criterion is questioned when the symptomatic cases reported are only the tip of the iceberg in terms of infection . Using our fitted models , we explored the reliability of absence-of-disease as an indicator of WPV extinction . Because of widespread subclinical infections , there was a stark contrast between the simulated number of polio infections and clinical cases ( Fig 4A ) . This contrast ( i . e . , the disconnect between infections and clinical cases ) , can lead to epidemiological scenarios where absence-of-disease is uninformative . In our models , WPV persistence was achieved by one of two mechanisms: ( 1 ) local unbroken chains of transmission , or ( 2 ) local extinction followed by rapid reintroduction . For each of these two mechanisms , we found that clinical case data can be misleading , as outlined in Table 1 . For instance , if WPV circulated at low levels of infection , extended absence of clinical cases could lead to the conclusion that WPV was locally eradicated . Similarly , if local extinction of WPV occurred , and was quickly followed by reintroduction and clinical cases , local extinction could go unrecognized , potentially misdirecting targets for control ( e . g . , to focus on sink populations rather than source populations ) . By simulating our fitted models , we identified extended periods absent of disease and used these periods to quantify the number of silent infections ( Fig 6B and 6C ) . We observed that if infections were maintained at relatively low numbers ( i . e . , under 100 infections per month ) , then WPV could circulate silently for over 30 months ( Fig 6B ) . The silent circulation of WPV can result in thousands of infections before a single reported case is observed ( Fig 6C ) . Our models assumed homogeneous mixing within each US state , and it is important to recognize that different mixing patterns could increase or decrease the lengths of chains of silent transmission . Because of the silent circulation of polio , it is difficult—and perhaps indefensible—to use clinical case data ( i . e . , without fitted models ) to evaluate WPV persistence . We simulated the fitted models to quantify the distribution of cases observed during periods with WPV extinction ( Fig 6D ) . The distribution of cases surrounding WPV extinctions is fairly symmetric because of the reintroduction of WPV following extinction . Therefore , we conclude that , in the face of rapid reintroduction following WPV extinction , case data cannot be used to identify extinction events . Though it is desirable to use fitted models to identify signals of extinction , and apply this knowledge to case data , it would require extensive evaluation of silent circulation .
This work sheds light on the fundamental ecology of WPV . Latitudinal gradients have been identified in several acute viral infections , including influenza , Respiratory Syncytial Virus ( RSV ) , rotavirus , and now polio [38 , 39] . Our results indicate that the observed latitudinal gradient in the timing of polio epidemics is driven by a latitudinal gradient in demographic and/or environmental factors associated with transmission . Determining which mechanism is responsible has implications for control and surveillance efforts . Specifically , knowledge of the seasonal driver could allow for regionally-timed national immunization campaigns or the ability to forecast changes in epidemic seasonality . Our identification of birth rate as a driver of polio’s epidemic emergence during the baby boom of the 1940s and 1950s is yet another demonstration [40 , 41] of the need for full integration of demography into the study of childhood infectious disease epidemiology . The rate of susceptible recruitment has long been known to control the magnitude and frequency of epidemics of fully immunizing childhood diseases [40 , 42] . Today , in an era of human population expansion and emerging infectious diseases , we are reminded of the importance of characterizing changes in host population ecology . As a result of limits of our demographic data , we were unable to address the early emergence phase of polio from the late 1800s through the 1920s . Rather , we focused on the later phase of emergence in the US , as the disease transitioned from small epidemics in the 1930s and early 1940s to large epidemics during the baby boom era . Though there were increases in the report rate , which contributed to the trend in observed cases , we also discovered an increase in the incidence of infection . Importantly , the increase in infection incidence closely tracked birth rates in the mid-1900s . To the extent that our results bear on contemporary polio ecology , the identification of source-sink dynamics in the US suggests that successful local elimination of polio in a sink population is inconsequential in the presence of a source population . This prediction has unfortunately been repeatedly borne out in current epidemics . Regional elimination of polio has been followed by reintroduction from endemic countries , such as the 2013 outbreak in Somalia , Ethiopia , and Kenya , with WPV introduced from Nigeria and repeat reinfection of Afghanistan from Pakistan [43] . Moreover , the metapopulation structure of WPV demonstrates that preventing emigration of WPV from source populations—which may be highly localized—is a requirement for efficient control . We estimate that over 99% of infections were subclinical , with the reporting of total infections regularly below 1% . Importantly , subclinical infections are likely more common today than in the period we studied . This is because , first , both nonparalytic and AFP cases were reported in the US , whilst only AFP cases are currently reported . Second , our models were fit to data during the vaccine-free period of polio endemicity; therefore , infection incidence was elevated each summer , allowing the number of infections to grow sufficiently large to result in a high probability of clinical infections . In contrast , today , as polio’s reproductive number approaches Rt = 1 in highly vaccinated endemic countries , WPV can circulate at levels below the level needed for likely clinical observation . The recovery of environmental WPV isolates in Israel in the complete absence of AFP cases supports this expectation [43] . Furthermore , Fig 6B demonstrates that polio may circulate silently for extended periods ( i . e . , longer than 3 y ) if the number of infections remains below the threshold for likely detection . Two years of silent WPV circulation has been confirmed: The outbreak in Central Africa detected in October 2013 was traced back to WPV circulation in Chad during 2011 [44] . Populations expected to have a small number of monthly infections in the presence of WPV—because of their demography or because they are highly vaccinated—would therefore be desirable targets for intense environmental surveillance . In terms of information gained , environmental surveillance is a powerful tool for identifying silent transmission in locations where polio would otherwise go undetected . In Pakistan , the level of environmental surveillance has increased since 2011 , and WPV has consistently been detected , even in the absence of AFP cases [45] . In the absence of validated transmission models , case data are relied upon to determine whether a pathogen has gone locally extinct and estimate the critical community size required for pathogen persistence . In light of polio’s propensity for silent circulation , we conclude that AFP data can be misleading; this conclusion extends to any communicable disease in which clinical cases represent a small fraction of infections . Extended periods absent of reported cases can mask infections circulating at levels below the threshold for likely reporting . We therefore advocate fitting transmission models to contemporary data to draw inferences regarding extinction . Since infection can persist even in the extended absence of reported cases , knowledge of the local infection dynamics could reveal invaluable epidemiological information . Transmission models fit to endemic countries ( i . e . , Pakistan , Afghanistan , and Nigeria ) could be used to identify how demographic and environmental factors interact with vaccine coverage to determine regional WPV persistence . In addition to coupling case data with transmission models for endemic countries , another useful extension would be to combine genetic data from WPV isolates with transmission models to further distinguish between sustained local transmission and imported infection . Genetic studies have found reductions in WPV genetic diversity in Afghanistan , suggesting local extinction of some WPV strains [6] . Vaccination campaigns might take further advantage of the seasonality and geographic clustering of WPV’s reproductive ratio . Low-transmission-season vaccination campaigns have been utilized by the Global Polio Eradication Initiative ( GPEI ) [6] . We found that the “low season” reproductive ratio can have geographic clusters where the reproductive ratio is greater than 1 , which , if identified in the contemporary setting , might be useful targets for intense low season vaccination campaigns . Additionally , if the: ( 1 ) seasonal reproductive ratio , ( 2 ) birth seasonality , and ( 3 ) vaccine coverage are quantified for endemic countries , vaccination campaigns could use this information to determine the regionally optimal timing for national vaccination days . These three quantities could be used to estimate the seasonal effective reproductive number and evaluate alternative vaccination strategies . For instance , one strategy might be to extend the duration of the wintertime trough ( i . e . , by generating or extending the window during which the effective reproductive number is below 1 ) , which may push WPV to extinction . Alternative strategies might be to vaccinate in the months prior to the seasonal peak in transmission or six months following the peak in births . In the past , mass OPV campaigns held during the low transmission season were deemed “most effective" [46] , but it is unclear to what extent this strategy is used today . Historical data , particularly in pre-vaccine periods , offer a unique glimpse into the ecology of infection , without a high degree of human intervention . Historical data offer several advantages . First , reporting rates from historical eras are informative because they are reflective of ( a ) the symptomatology of infection and ( b ) clinical diagnosis of symptomatic infection . Second , it can be difficult to infer unobserved infection dynamics using data for diseases that are near their eradication or elimination threshold . This is because the parameterization of transmission models with data containing few cases—and lacking recurrent epidemics—can result in ambiguous parameter estimates . The recurrent nature of historical epidemics gives us the unique opportunity to unravel disease-specific transmission ecology . Once the baseline transmission ecology is known , it can be coupled with data from contemporary periods to test hypotheses regarding modern day epidemics and their geographic coupling . Our analyses demonstrate the power of an approach focused on coupling mechanistic transmission models with long-term , spatially replicated longitudinal incidence data . Specifically , we document intriguing continental-scale gradients in polio seasonality , which we suggest are explained by latitudinal gradients in local transmission rates . We also show that the historical emergence of epidemic polio was largely a consequence of demographic trends rather than improvements in hygiene . Importantly , we demonstrate that historical polio persistence in the US was driven by an ever-changing mosaic of source-sink populations . Finally , we found that even protracted AFP-free periods do not reliably indicate WPV extinction . Because of the difficulty in establishing fundamental aspects of WPV transmission in heavily vaccinated populations , it is our hope that these insights will act as a baseline for understanding modern polio transmission and disentangling vaccine effects from the natural ecology of the disease . | Thanks to global vaccination efforts , poliovirus is on the brink of worldwide eradication . However , achieving eradication and preventing re-emergence requires intimate knowledge of how the virus persists . In order to understand a system that is complicated by heavy human intervention , such as vaccination , it is important to establish a baseline by studying that system in the absence of intervention . Historical epidemics that predate the use of vaccines can be used to disentangle the epidemiology of disease from vaccine effects . Using historical polio data from large-scale epidemics in the US , we fitted and simulated mathematical models to track poliovirus and to reconstruct the millions of unobserved subclinical infections that propagated the disease . We identified why polio epidemics are explosive and seasonal , and why they vary geographically . Our analyses show that the historical expansion of polio is straightforwardly explained by the demographic “baby boom” during the postwar period rather than improvements in hygiene . We were also able to demonstrate that poliovirus persisted primarily through symptomless individuals , and that in the event of local virus extinction , infection was reintroduced from other geographic locations . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | Unraveling the Transmission Ecology of Polio |
Mucins are heavily glycosylated proteins that give mucus its gel-like properties . Moreover , the glycans decorating the mucin protein core can alter the protective properties of the mucus barrier . To investigate whether these alterations could be parasite-induced we utilized the Trichuris muris ( T . muris ) infection model , using different infection doses and strains of mice that are resistant ( high dose infection in BALB/c and C57BL6 mice ) or susceptible ( high dose infection in AKR and low dose infection in BALB/c mice ) to chronic infection by T . muris . During chronicity , within the immediate vicinity of the T . muris helminth the goblet cell thecae contained mainly sialylated mucins . In contrast , the goblet cells within the epithelial crypts in the resistant models contained mainly sulphated mucins . Maintained mucin sulphation was promoted by TH2-immune responses , in particular IL-13 , and contributed to the protective properties of the mucus layer , making it less vulnerable to degradation by T . muris excretory secretory products . Mucin sulphation was markedly reduced in the caecal goblet cells in the sulphate anion transporter-1 ( Sat-1 ) deficient mice . We found that Sat-1 deficient mice were susceptible to chronic infection despite a strong TH2-immune response . Lower sulphation levels lead to decreased efficiency of establishment of T . muris infection , independent of egg hatching . This study highlights the complex process by which immune-regulated alterations in mucin glycosylation occur following T . muris infection , which contributes to clearance of parasitic infection .
The intestinal epithelium is lined by a continuous mucus barrier which provides physical protection and chemically protects the epithelial cell layer by sequestering important host defence factors within its complex matrix [1] . Gel-forming mucins ( Muc2 in the intestine ) produced by goblet cells , give mucus its gel-like properties and play a significant role in protection against helminth infections [2] . Mucins are large heavily glycosylated proteins , predominantly consisting of O-glycans which account for at least 70% of their molecular weight [3] . The O-linked sugars are assembled progressively by glycosyltransferases on to a serine or threonine residue , found in the serine-threonine-proline rich tandem repeat regions of the mucin protein core [1] . These glycan chains have well-established fundamental roles in many biological processes including in inflammatory responses [4] . Changes in mucin glycosylation have been previously described in murine parasitic infections . Whether these changes are important in the protective function of the mucus barrier , however , has not yet been established . Several studies have investigated the direct role of glycans as protective elements that attach and clear pathogens from the gut [5–8] , but most of the evidence for the protective role of the glycans present on mucins comes from animal models . A change in the glycans can lead to inflammation because of the alteration in the protective properties of the mucus barrier [9–12] . For example , the terminal sugar fucose has been shown to be crucial in host-pathogen interaction in Helicobacter pylori infection [13] . Moreover , inducing colitis in mice deficient in the N-acetylglucosamine-6-O-sulphotransferase-2 [14] which is expressed highly in the colon , results in a significantly higher leukocyte infiltration and is thought to exacerbate inflammation . Mice lacking either core 1- or core 3-derived O-glycan chains , or both , develop spontaneous colitis with the double knockout mice having the most severe and widespread disease [15–17] . Previous studies have shown that inducing mucin sulphation with reserpine in vivo reduces the establishment of infection with the intestinal helminth Strongyloides venezuelensis [18] . Sulphotransferases were also up-regulated shortly before the rejection of the helminth , Nippostrongylus brasiliensis and have been suggested to play a protective role during worm expulsion [19 , 20] . In addition , the up-regulation of sialomucins observed in other helminth models such as Trichinella spiralis , which have been found to be regulated by T cells , are thought to be protective [9 , 10] . Moreover , gastrointestinal disorders such as ulcerative colitis and Crohn’s disease are associated with a loss of mucin sulphation [21 , 22] . Whether the changes observed in mucin glycosylation in human colitis occur as a consequence of disease or are an active alteration important in resolving the disease has yet to be fully elucidated . Mucin glycosylation is known to be important in maintaining intestinal homeostasis and an absence of mucin glycosylation results in inflammation [17 , 23] . It was shown that a reduction in mucin sulphation , in particular , can lead to an increase in susceptibility to colitis and bacterial infections [24] . In the intestine , a rich commensal flora is maintained within the outer mucus layer and it is thought that the microbiota can influence the level of sulphation in the intestine , which could in turn effect mucin sulphation [25] . Paradoxically , mucin sulphation can also protect the mucin core from bacterial glycosidases [4] . Overall , it is thought that the alterations in glycosylation can lead to reduced effectiveness of the mucus barrier , which in turn may exacerbate inflammation . The intestinal niche occupied by nematodes , in particular the caecum , contains mucins that are more sulphated than in any other site in the body [21] . Using the Trichuris muris ( T . muris ) mouse model , we have shown that intestinal mucins play an essential role in the expulsion of this helminth from the host [2 , 26] . Therefore , the changes in mucin sulphation could alter the protective nature of the mucus barrier and affect helminth establishment and/or expulsion . T . muris helminth infection model provides a robust model of both acute and chronic infection . Using this model , we observed a clear switch in glycosylation from sulphomucins to sialomucins during chronic infection in the T . muris murine model . For the first time , we also demonstrate that mucin sulphation is influenced by IL-13 and mucus with high sulphomucin content is more resistant to degradation by the T . muris excretory secretory products ( ESPs ) . Moreover , underlining the functional importance , Sulphate anion transporter1 ( Sat1 ) -deficient mice with reduced mucin sulphation developed chronic T . muris infection . Interestingly , a preliminary histopathological examination of individuals infected with Enterobius vermicularis , a highly prevalent gastrointestinal dwelling nematode of human cecum showed a decreased mucin sulphation staining supportive of the mouse studies .
To gain an understanding of whether mucin glycosylation plays an important role during helminth infection , we utilised the T . muris model of helminth infection in mice . The degree of mucin sulphation and sialylation were assessed during T . muris infection with High-Iron Diamine-Alcian Blue ( HID-AB ) staining in acute ( high dose ( HD ) infection in BALB/c mice ) and chronic ( HD and low dose ( LD ) infection in AKR and BALB/c mice , respectively ) ; worm burdens are shown in S1a Fig . Using HID-AB staining a distinction can be made between sulphomucins ( black ) and sialomucins ( blue ) , present in the goblet cell thecae ( Fig 1A ) . Normally in the caecum , the niche of the parasite , the mucins are predominantly highly sulphated with little evidence of sialomucin-containing goblet cells ( Fig 1A—naïve levels ) . In the mice with acute infection , as infection progressed , there was goblet cell hyperplasia and the expected increase in Muc2 expression ( Fig 1C ) [2] . Importantly , mucins present within the goblet cells in the caecal crypts were all majorly sulphated ( Fig 1B ) . However , in the chronic models of infection , there was a decrease in Muc2 expression ( Fig 1C ) , accompanied by a loss of sulphated goblet cell thecae and an increase in sialomucins within the goblet cells . This switch in glycosylation was localised to the caecum; no major changes were observed in the colon ( S1B Fig ) . Mucin glycosylation is dependent on the array of glycosyltransferases present within the Golgi apparatus of the goblet cells . Therefore , considering the differences identified , the changes in the expression of major sulpho- and sialotransferases in isolated caecal epithelial cells was determined using qRT-PCR ( Fig 2 ) . Interestingly , mirroring the increase in sulphated mucins present in goblet cells , the galactose-O-sulphotransferases Gal3St1 , Gal3St2 and Gal3St3 ( Fig 2A ) and glucosamine-O-sulphotransferases GlcNAc6ST1 , GlcNAc6ST2 and GlcNAc6ST4 ( Fig 2B ) were highly up-regulated around the time worm expulsion is occurring on day 21 pi . in acute infection ( BALB/c HD ) but not during chronic infection ( BALB/c LD; worm burden data is shown in S1A Fig ) . In acute infection , the level of sulfotransferase expression then reduced after worm clearance paralleling the result for Muc2 expression ( S1B Fig ) . In contrast , in chronic infection in the same strain of mice infected with a low dose of worms , the sulphotransferase genes were not induced but the sialyltransferases ST3Gal1 , ST3Gal2 and ST8GalNAc were upregulated on day 21 , with ST3Gal1 remaining high at day 35 pi ( Fig 2C ) . Thus induction of a spectrum of sulphotransferase genes occurs concomitantly with worm expulsion . We wanted to determine whether the changes in glycosylation were a direct result of infection itself and whether these changes in mucin glycosylation could be altered post infection . Therefore , to address these questions , BALB/c mice were infected with a low dose of T . muris eggs [27]; this resulted in a chronic infection , as adult worms were still present in the caeca of these mice on day 35 pi . ( Fig 3A and 3C ) . As expected , by day 35 pi . no changes in mucin expression were observed ( S2C Fig ) in these mice however , goblet cells predominantly contained sialomucins ( Fig 3B ) . Subsequently , at day 35 pi . , which is considered to be a chronic infection , these mice were re-challenged with a high dose ( >150 ) of T . muris eggs ( Fig 3A ) . A second infection was established in these mice , as approximately 100 worms were present in the caeca after 12 days ( day 47 pi . ) of re-infection ( Fig 3C ) . However , most worms were eradicated 21 days after reinfection ( day 56 pi . ) and no worms were present in the caecum by 35 days post-reinfection ( day 70 pi . ) . Re-challenging these chronically infected mice with a high dose infection resulted in the induction of Muc2 and Il13 expression and a reduction in Ifnγ expression ( S2A and S2B Fig ) . Histological analysis of the caeca with HID-AB staining revealed as expected the change from goblet cells containing predominantly sulphomucins to sialomucins in chronic infection ( Fig 3B ) . These changes were reversible and goblet cells changed from producing sialomucins to sulphomucins after reinfection ( Fig 3B ) , which correlated with T . muris worm expulsion ( Fig 3C ) . It is well-established that the mice resistant to T . muris infection mount strong TH2-type immune responses and those susceptible to chronic infection exhibit TH1-type immune responses ( as can be seen from the relative caecal expression of genes encoding the TH2 and TH1 cytokines Il13 and Ifnγ in S2 Fig ) [28] . To assess the role of the TH2-type immune response and in particular , IL-4 and IL-13 , in mucin sulphation , IL-4 knockout ( KO ) and IL-4R KO mice on the BALB/c background were infected with a high dose of T . muris eggs . The IL-4 KO mice were able to expel worms , however worm expulsion was slightly delayed in these mice ( S3A Fig ) , as reported previously [29] . On day 18 pi . goblet cell hyperplasia was less pronounced in the IL-4 KO mice when compared to the wild-type BALB/c mice , although this was not significantly different ( Fig 3D ) . A significant increase in the number of goblet cells was observed in the IL-4 KO mice by day 32 post infection compared to naïve mice . Furthermore , HID-AB staining showed that goblet cells in the IL-4KO mice predominantly contained sulphated mucins ( Fig 3E and 3F ) . As IL-4 and IL-13 act through a heterodimeric receptor involving IL-4Rα , IL-4R KO mice do not respond to either cytokine and are unable to mount strong TH2-type immune responses , and consequently are susceptible to chronic T . muris infection ( worm burden; S3A Fig ) [29] . Importantly , uninfected control IL-4R KO mice had significantly lower levels of sulphation compared to the naïve wild type and IL-4 KO mice ( Fig 3D–3F ) . This was reflected by very substantial reduction in the sulphotransferase genes , Gal3ST1 , GlcNAcST2 , GlcNAc6ST3 in the naïve IL-4R KO mice ( S3B Fig ) , the other sulphotransferase genes shown in Fig 2 remained unaltered ( S3G Fig ) . As infection progressed , a decrease in the number of goblet cells was also observed in the IL-4R KO mice compared to naïve and IL-4 KO mice . Moreover , the goblet cells lost their ‘goblet like morphology’ , exhibiting very small thecae , which stained blue with HID-AB , indicative of predominantly containing sialylated mucins ( Fig 3F ) . To determine whether the differences in sulphomucin content alters the properties of the mucus barrier , we compared mucus from wild-type mice and NaS1 KO mice , which lack the Na+-sulphate transport 1 ( Slc13a1 ) [30] . Slc13a1 is primarily expressed in the ileum/caecum/colon and kidney where it mediates sulphate absorption and reabsorption , respectively . Deletion of Slc13a1 in mice leads to hyposulfataemia and reduced sulphonation capacity which leads to depleted intestinal sulphomucin content ( S4A Fig [24] ) . Mucus extracted from the uninfected wild-type and NaS1 mice was subjected to agarose gel electrophoresis and western blotted . Staining with HID confirmed that mucin sulphation was significantly reduced in the NaS1 mice compared to wild-type mice , with PAS staining ( reacts with all mucin carbohydrates ) used to confirm that comparable amounts of glycoproteins were isolated and compared ( S4B Fig ) . The excretory/secretory products ( ESPs ) released by T . muris , contains a mixture of enzymes . We have previously demonstrated that serine proteases secreted by the helminth can specifically depolymerise the Muc2 mucin polymers that give mucus its viscoelastic properties [31] . Therefore , we assessed the contribution of the mucin sulphation to the resistance to depolymerisation of Muc2 . Caecal mucus was isolated from wild type and Nas1 KO mice and treated with 50 μg/ml of ESPs from T . muris for 2 or 6 hours at 37°C to determine whether ESPs can alter mucus . Treated or untreated mucus was subsequently subjected to rate zonal centrifugation to assess the change in size and/or shape of mucins by analysing their distribution by PAS-staining . The sedimentation profiles of untreated glycoproteins from wild-type ( high sulphomucin content ) and NaS1 KO ( lower sulphomucin content ) were comparable with broad peaks ( fractions 9–20 ) ( S4C Fig ) . After treatment with ESPs from T . muris , the sedimentation profile of glycoproteins from NaS1 KO mice was altered , whereas the profile of wild type glycoproteins was substantially unchanged . The PAS-positive material from the NaS1 KO mice shifted to the top of the gradient ( fractions 4–17 ) compared to PAS-positive material from the wild-type mice present in fractions 9–20 , consistent with reduced size of mucin polymers ( S4C Fig; quantification of rate zonal data presented as percentage of area under the curve shown in S4C Fig ) . This experiment shows that mucus with lower sulphomucin content is more susceptible to the degradative effects exerted by the T . muris ESPs . High sulphomucin content that accompanies helminth rejection was predominately controlled by IL-13 . This was confirmed in-vitro by treating the intestinal cell lines , LS174T cells , with 50 ng/mL of IL-13 for 24 h . We observed an increase in mucin production , which was accompanied by an increase in sulphotransferase and a decrease in sialyltransferase genes ( S3C and S3D Fig ) . Treatment of LS174T cells with IFN-γ lead to an increase in sialytransferases ( S3E Fig ) , whereas no alterations in sulphotransferases were observed ( S3E Fig ) . We also examined degradation by T . muris ESPs of mucins produced by LS174T cells with and without IL-13 stimulation and found that sulphated mucins induced by IL-13 treatment are more resistant to degradation by parasite ES products . Moreover , sialylated mucins induced by IFNγ treatment were degraded rapidly compared to control mucins ( S3F Fig ) . As IL-13 is an essential mediator in the expulsion of helminths and altered sulphation levels via the regulation of sulphotransferases , we hypothesised that the level of mucin sulphation in the caecum could affect T . muris infection , in particular worm establishment and rejection . To address this possibility , NaS1 KO and their wild-type littermates ( C57BL/6-background ) were infected with a high dose of T . muris eggs . As stated previously , the mucin sulphation in the caecum of the naïve NaS1 KO mice was lower as compared to the wild-type mice [24] . This , however , did not have an effect on the establishment of infection ( S4D Fig ) ; similar number of worms were present , on day 12 pi . , in the NaS1 KO and wild-type mice . Moreover , as infection progressed , the kinetics of worm expulsion were similar in the both wild-type and the NaS1 KO mice ( S4D Fig ) . As we had demonstrated that the NaS1 KO mice expel T . muris over a similar time course as wild-type mice , RT-PCR was used to determine whether these mice mount a similar and concurrent TH2-type immune response to T . muris infection . The levels of Il4 and Il13 were elevated after infection in the NaS1 KO and wild-type mice compared to naïve mice ( S5A and S5B Fig ) . A 2 to 6-fold increase in Ifnγ levels was also observed across the time course of infection ( S5C Fig ) in both NaS1 KO and wild-type mice although this did not significantly affect worm expulsion . The effect of the elevated IL-4/IL-13 levels was reflected in the histological analysis , as goblet cell hyperplasia . We have previously demonstrated that in the absence of the major intestinal mucin Muc2 , results in a delay in the expulsion of the helminth T . muris from the host [2] . Importantly , denovo expression of the mucin Muc5ac in the niche of the helminth is critical for its expulsion [26] . However , in the NaS1 KO mice Muc2 and Muc5ac levels were similar compared to the WT mice ( S5D–S5F Fig ) . HID-AB staining intensity showed that caecal mucin sulphation in the NaS1 knockout naïve mice was significantly less than in wild-type littermates ( S4E and S4F Fig ) . However , surprisingly , as the infection advanced , the levels of mucin sulphation recovered in the NaS1 KO mice , and by day 25 of infection no difference was observed in the sulphation of the goblet cells in the caecal crypts ( S4E and S4F Fig ) . As the T . muris infection progressed in the NaS1 KO mice , the depletion in the sulphomucin content within the goblet cells was reversed , suggesting that free sulphates , although lower systemically due to renal excretion , were transported efficiently through another sulphate transporter expressed in the caecal epithelium to be incorporated into the mucins . In addition to NaS1 , sulphate anion transporter-1 ( Sat1 , Slc26a1 ) is the other major sulphate transporter in the caecum . Sat1 is expressed on the basolateral surface of epithelial cells and functions independently of NaS1 , which is expressed on the apical membrane in the ileum , caecum and colon [32] . Therefore , we next assessed whether Sat1 expression was altered in the NaS1 KO mice following infection . Using immunohistochemistry and RT-PCR , we observed a marked upregulation of Sat1 mRNA in the NaS1 KO mice on day 18 pi . , whereas , the up-regulation of Sat1 occurred on day 25 pi . in the wildtype mice ( Fig 4A and 4B ) . This suggested that Sat1 was compensating in the NaS1 KO mice during T . muris infection to improve sulphate uptake and availability as a substrate for incorporation into mucin oligosaccharides . In light of this , we sought to determine whether Sat1 KO mice also have reduced mucin sulphation in the caecum and whether this affects T . muris infection . HID-AB staining clearly showed that the mucins have low sulphation in Sat1 KO mice ( Fig 4C ) . Interestingly , on day 7 pi . , there was a significant reduction in the worm burden demonstrating a difference in the establishment of infection ( Fig 4D ) . Despite lower establishment of T . muris , as infection progressed it was clear that the Sat1 KO mice were susceptible to chronic infection; these mice harboured infection until day 56 pi . and were unable to clear the infection ( Fig 4E ) . The differences in establishment could be due to a variation in the efficiency of T . muris egg hatching . Therefore T . muris eggs were incubated in-vitro with caecal explants from WT mice and Sat1 heterozygous and KO mice overnight . However , there were no significant differences in the percentage of eggs hatched in WT or Sat1 KO mice under aerobic and anaerobic conditions ( Fig 4F ) . Moreover , no differences in the immune response were observed at day 7 pi . , where IFN-γ and IL-13 cytokine levels from cultured mesenteric lymph nodes were similar between WT and Sat1 KO mice ( Fig 4G ) . We have previously shown that the mucins critical to worm expulsion adversely affect worm metabolism measured by ATP production [26] , and therefore we analysed ATP production in worms exposed to sulphated and non-sulphated mucins . Sulphated mucins ( isolated from the caecum of wild-type mice ) lowered the energy levels of worms , whereas this effect was lost with decreased mucin sulphation ( isolated from the caecum of Sat1 KO mice ) directly implicating sulphated oligosaccharides in the anti-parasitic function of the mucins ( Fig 4H ) . These findings in the Sat1 KO mice show that mucin sulphation is a critical element of effective helminth expulsion . We investigated whether mucin sulphation and sialylation that are associated with chronic T . muris infection were also present in human helminthic infection . To this end in a pilot study we analysed specimens of vermiform appendix with histological evidence of Enterobius vermicularis ( EV ) infection and a control group of non- or mildly inflamed appendices with no histological evidence of infection ( infection is unlikely but cannot be definitively excluded in this group ) . The cases were grouped into EV with or without mild non-erosive appendicitis ( n = 10 and n = 29 , respectively ) and uninfected control appendices ( n = 18 ) . Infection with EV was associated with increased numbers of activated lymphoid follicles ( EV vs control , mean 5 . 9 vs . 4 . 2 , p = 0 . 05 , Mann Whitney U t-test ) . Quantification of HID-AB staining illustrates that the intensity of mucin sulphation staining in EV-infected appendices was decreased compared to appendices lacking evidence of EV infection ( S6 Fig ) .
The mucus barrier overlaying the intestinal epithelium is the first line of defence against enteric parasitic infections . We have previously demonstrated a protective role of mucins , the major component of the mucus barrier , in helminth expulsion [2] . It is thought that the glycans that decorate the mucin protein backbone , contribute to the protective properties of the mucus gel [1] . In this study , we demonstrate that sialylated mucins are mainly present in chronic helminthic infection . We described the changes in glycosylation in chronic and acute T . muris infection and show , in the strong TH2 environment that accompanies worm expulsion , that mucins were highly sulphated driven primarily by IL-13 mediated upregulation of goblet cell sulphotransferases . In contrast , predominantly sialylated mucins were found within the goblet cells during chronic T . muris infection in mice and in human appendices with Enterobius infection . Mice incapable of appropriate mucin sulphation could not expel T . muris demonstrating the importance of mucin sulphation as a key element of immune-driven worm expulsion . Providing a mechanistic explanation for the role of sulphated mucins in worm expulsion , we demonstrated that mucin polymers with high sulphomucin content were inherently more effective in repressing worm metabolism in vitro and were also more resistant to degradation by T . muris excretory secretory products . Using two genetic models of sulphate transporter deficiency we were able to establish the importance of mucin sulphation in both establishment of infection and subsequent immune clearance ( see Fig 5 ) . Changes in mucin glycosylation have been reported to coincide with inflammation in several gastrointestinal diseases such as ulcerative colitis [33] . Moreover , alterations glycosylation of mucins have also been reported to occur during several models of helminth infection; N . brasiliensis , T . spiralis and H . polgyrus [10 , 11 , 34] . Whether these changes occur as a result of an on-going inflammatory response or as a result of an active change in order to resolve infection is not yet known . Our data suggests that distinct mechanisms regulate glycosyltransferases responsible for mucin sialylation and sulphation , corroborating previously published work [19] and definitively showing that IL-4/IL-13 receptor signalling preferentially drives expression of sulphotransferases . Histological analysis revealed that the changes in glycosylation with chronic T . muris infection were localised only within the caecum , consistent with the localised mucosal immune response . Interestingly , similar to goblet cell hyperplasia which is restricted to the caecum [2] , no major changes in glycosylation were observed in the colon post T . muris infection . The changes in glycosylation were restricted to the helminths niche , it suggests these could be due to or a local immune response to/by the helminth itself . As human EV infection was also associated with an increased number of activated lymphoid follicles , this may represent an altered inflammatory milieu , which could contribute to the changes in mucin glycosylation . Mirroring the changes observed with the histology , an increase in the gene expression of sulphotransferases was observed in resistance , and sialotransferases were up-regulated in susceptibility . In chronic infection by T . muris , along with the down-regulation of goblet cell differentiation transcription factors , glycosylation within the goblet cells could possibly be perturbed . Indeed , there is evidence of a general loss of mature glycosylation on mucins in chronic infection [35] . The loss of sulphation was in naïve uninfected IL-4R KO mice when compared to the IL-4 KO mice , suggesting that in health the basal level of mucin sulphation is in part due to IL-13 . In-vitro data showed that IL-13-treatment of colonic cell lines leads to an increase in sulphotransferase expression and subsequently protects the mucins from degradation by T . muris ESPs . Alterations in glycosylation could affect the hydration of the mucus gel and the viscosity of the barrier , which we have previously shown to be altered during worm expulsion [26] . It is important to acknowledge that in endemic areas , people are often infected with low doses of helminths , repeatedly . Therefore , the gradual change in glycosylation that occurs as a result of infection may eventually contribute to ‘acquired immunity’ and worm expulsion . Parasitic excretory secretory products have the ability to degrade purified mucins [36] . The excretory secretory products ( ESPs ) of the T . muris helminth , that lives within the blanket of mucus , have been shown to be highly immunogenic [37] . However , whether ESPs affect crude mucus has not been previously investigated . Comparable amounts of mucus extracted from naïve NaS1 and wild-type mice were used to compare the effect of ESPs on low and highly sulphated mucins , respectively . After treatment with T . muris ESPs , the mucus with low sulphomucin content ( from NaS1 KO mice ) was reduced in size , which is evidence of degradation of the polymeric macromolecule responsible for the viscoelastic properties of the mucus and its ability to retain other host factors . This difference in the effect of ESPs was apparent on mucus with low sulphomucin content after only 2 hours of treatment , whereas ESPs had little effect on the mucus from wild-type naïve mice even after 6 hours . As we have demonstrated ESPs may be released as part of the helminths regime to promote its own survival , as the degradation of mucins would lead to diminishing the gel-like consistency of the mucus surrounding the worm and the delivery of anti-helminthic host factors [31] . Our findings here suggest that glycosylation plays an important role in preventing degradation , and that highly sulphated mucins protect the mucin protein core from the degradative effects of the helminths ESPs . Indeed studies of von Willebrand factor ( vWF ) , which is a large multimeric glycoprotein homologous to gel-forming mucins , revealed that sialylated glycans increase susceptibility to its proteolysis [38] . We demonstrate that mucin sulphation patterns seen during T . muris infection influence both the establishment and expulsion of the parasite . Increased sulphation on intestinal mucins has been shown to reduce the establishment of the helminth Strongyloides venezuelensis [18] . We noted that depleted levels of sulphation using the Sat1 KO had a major effect on the establishment of T . muris infection . Of note , the caecum , which is the chosen niche of the T . muris helminth , is the highly sulphated . There is a possibility that sulphation is the cue for the T . muris eggs to hone in to the caecum as with reduced sulphation we did observe a decrease in establishment , no worms were found in the other parts of the intestine ( small intestine , proximal colon or distal colon ) . The differences in establishment were not due to differences in the hatching of T . muris eggs . Together with our previous studies [2 , 26] we now know that mucins are critical to expulsion , MUC5AC is more important than MUC2 , oligomerisation of the mucins is not required for the adverse effects on metabolism , and sulphated oligosaccharides are required . How the long intact sulphated mucin molecules transmit their adverse effects on worm metabolism remain to be demonstrated . No effect on T . muris establishment was noted in the NaS1 KO mice , possibly because the depletion in sulphation was not sufficient to have an effect on worm establishment . In the absence of the apical sulphate transporter , NaS1 , these hyposulphataemic mice up-regulated the expression of Sat1 transporter , which is involved in the uptake of sulphates from blood [32] . Interestingly , Sat1 is elevated with the increasing demand for sulphates during the rejection of T . muris , earlier in the NaS1 KO mice compared to the wild-type mice . Sat1 levels return almost to baseline in the NaS1 KO mice by day 25 pi . , which is when the goblet cell hyperplasia seemed to subside . Such evidence suggests that Sat1 expression is up-regulated in response to the increased cellular requirement of sulphates , rather than the immune response and is effectively scavenging sufficient sulphate to effectively sulphate the mucins and ensure worm clearance . Importantly , the sustained sulphation observed in the acute T . muris model is essential for the expulsion of the helminth from the host . In summary , we have demonstrated that maintained mucin sulphation , influenced by the TH2 type immune response , is clearly a feature of resistance to T . muris infection . This has likely evolved as a part of the immune response against T . muris as increased sulphation changes the properties of the mucus barrier making the mucins more resistant to degradation . Depleted levels of mucin sulphation , have a significant effect on the establishment of the T . muris infection that appear independent of changes in the microbiome . With the high demand of cellular sulphate during T . muris infection , Sat1 was up-regulated in the caecal epithelium and is essential for efficient mucin sulphation with deficiency leading to inability to clear infection . Given the essential role of mucins in clearing helminth infections , this study highlights the complex process by which alterations in mucin glycosylation occur following infection and contribute to the establishment and clearance of infection .
AKR , BALB/c ( Harlan Olac ) , IL-4 KO and IL-4R KO ( BALB/c-background ) mice were maintained in the Biological Services Unit at the University of Manchester . All mice used were at 6–10 week old male mice . The protocols employed were in accordance with guidelines by the Home Office Scientific Procedures Act ( 1986 ) . NaS1 knockout ( KO ) and Sat1 KO mice ( 4–12 weeks old male C57BL/6 ) and their wild-type littermates originally produced by gene mutation [24 , 32] were housed at the Mater Medical Research Institute and experiments were approved by the University of Queensland Animal Experimentation Ethics Committee . All mice were kept in sterilized , filter-topped cages , and fed autoclaved food . Cases of enterobius vermicularis infestation of the vermiform appendix over the period 2011–2014 were identified in the formalin fixed paraffin embedded archive of Mater Pathology Services . All cases were reviewed by an anatomical pathologist ( RL ) and anatomical pathology trainee ( PH ) to confirm the diagnosis . Cases with more than mild lamina propria acute inflammation were discarded to exclude confounding effects from other sources of inflammation . The cases were grouped into EV with or without mild non-erosive appendicitis ( n = 10 and n = 29 , respectively ) and uninfected control appendices ( n = 18 ) . 5μm thick sections were cut onto Superfrost coated slides ( Thermo Scientific , Braunschweig ) and submitted for staining as per below . Use of human specimens was approved by the Mater Health Services Human Research Ethics Committee ( reference 15MHS69 ) and Research Governance Office ( reference RG-15-147 ) . The techniques used for T . muris maintenance and infection were described previously [39] . Mice were orally infected with approximately 100–300 eggs for a high dose infection and >15 eggs for a low dose infection . Worm burdens were assessed by counting the number of worms present in the caecum as described previously [39] . For ATP measurements , live worms were extracted from the caecum of Rag1-/- mice , which was longitudinally cut and segmented before incubation with 0 . 1M NaCl for 2 hours at 37◦C with frequent shaking . Worms were counted after separation from debris and epithelial cells using a 0 . 7μm filter and kept in RPMI-1640 supplemented with 10% FCS . Live worms were incubated with either RPMI or mucus isolated from 6 wk old WT , NaS1 KO or Sat1 KO mice for 48 h . Alive worms were subsequently homogenised . The CellTiter-Glo luminescent cell viability assay was carried out according to manufacturer’s instructions ( Promega Corp . , USA ) . Relative light units ( RLUs ) were calculated per worm: RLU = ( sample light units − blank light units ) /number of worms . Substrate only was used as a blank control . A 1cm segment or the whole tissue ( rolled ) was fixed in 10% neutral buffered formalin or 95% ethanol and processed using standard histological techniques . Sections were treated with 0 . 1M potassium hydroxide for 30 minutes prior to staining with periodic acid Schiff’s reagent ( PAS ) . Slides were counterstained with either haematoxylin and eosin or 1% fast-green . To assess mucin sulphation , sections were stained with High Iron Diamine-Alcian Blue ( HID-AB ) as previously described [40] . Standard immunohistochemical staining methods [41] were used for immunohistochemistry with monoclonal Sat1 ( Sulphate anion transporter 1 ) antibody [32] . RNA from epithelial cells was isolated using the previously described method [42] . cDNA was generated using an IMPROM-RT kit ( Promega ) or Superscript III ( Invitrogen ) . Absolute QPCR SYBR Green ( ABgene ) was used for quantitative PCR . Primer efficiencies was determined using cDNA dilutions and genes of interest ( S1 Table ) were normalised against the housekeeping gene , β-actin , and expressed as a fold difference to uninfected naïve message levels . RT-PCR products were directly sequenced to verify the identity of the amplified genes . In brief , products were digested with Exonuclease I and Calf Intestinal Phosphatase and subsequently sequenced using the ABIPRISM Big-Dye Terminator cycle sequencing reaction at the Sequencing Facility in the University of Manchester . The data was analysed using Chromas Pro v1 . 34 and the sequences obtained were compared against the GenBank database ( http://www . ncbi . nlm . nih . gov/BLAST ) . LS174T cells ( originally obtained from ATCC ) were cultured as previously described [31] . In brief , cells were cultured in high glucose DMEM with 2mM L-glutamine , 100 U/mL penicillin , 100 g/mL streptomycin and 10% FBS until confluent ( 70–80% ) . Cells were then treated with 50 ng/mL of recombinant human IL-13 or IFNγ for 24 h and samples were taken for protein/RNA or rate zonal centrifugation . For hatching experiments , sections of mouse caeca or caecal extracts were isolated and kept overnight with T . muris eggs at 37°C and 5% CO2 or at 10% CO2 , 10% H2 , 80% N2 . Caeca from NAS1 KO mice and their wild-type littermates were gently flushed with PBS to remove the faecal matter . The mucus was lightly scraped and lyophilized and subsequently solubilised in 6M urea , reduced using 50mM dithiothreitol ( DTT ) and carboxylmethylated using 0 . 125M iodoacetamide . Samples were electrophoresed on 1% ( w/v ) agarose gels in TAE buffer ( 40mM Tris acetate and 1mM EDTA , pH8 ) and 0 . 1% SDS at 30 volts for 15 hours . After electrophoresis , mucins were transferred to a nitrocellulose membrane by vacuum blotting in 0 . 6M sodium chloride and 60mM sodium citrate at a pressure of 45–50 mbar for 2 hours detected using PAS , HID staining or probed with the MUC2 antibody [43] . The excretory secretory products ( ESPs ) were collected using methods previously described [44] . Aliquots of crude mucus scrapes ( in equal volumes of PBS ) were incubated at 37°C with the ESPs at 50 μg/ml for various time points ( as specified ) . Control samples were not treated with the ESPs , but were incubated at 37°C for the maximum time point as previously described [31] . Mesenteric lymph nodels ( mLNs ) were removed , cells were isolated and resuspended at 5 x 106 cells/mL in RPMI 1640 with 10% FBS , 2 mM L-glutamine , 100 U/mL penicillin , 100 μg/mL Streptomycin . Cultures were stimulated with 50 μg/mL of ESPs for 24 h at 30°C and 5% CO2 . Cell free supernatants were stored at -80°C . IFNγ and IL-13 levels were determined using ELISAs as per manufacturer’s instructions ( R&D ) . 6–8 M guanidinium chloride gradients were formed in centrifuge tubes using an MSE gradient maker connected to a Gilson Minipuls 2 peristaltic pump . Mucin samples ( in 4 M guanidinium chloride ) were loaded onto the gradient and centrifuged in a Beckman Optima L-90K Ultracentrifuge using a Beckman SW40 rotor at 40 , 000 rpm for 2 . 45 hours at 15°C . Fractions were taken from the top of the tubes , analysed by slot blotting and PAS-staining [45] . The refractive index of each fraction was measured using a refractometer to determine the guanidinium chloride concentration; the gradients were comparable . All histological analysis was done blinded . Sulphomucin containing crypts within the caecum ( identified as blackish goblet cells ) were quantified and compared to the total number of crypts . The numbers of goblet cells expressed per crypt were counted in 20–50 longitudinally sectioned crypt units . The area stained ( pixel/mm2 ) per 20–50 crypts was determined by using the ImageJ software version 1 . 39a . The goblet cell staining intensity was measured using the BioRad GS-800 densitometer in 250 goblet cells . All results are expressed as the mean ± SEM . Statistical analysis was performed using SPSS version 16 . 0 or GraphPad Prism version 6 . 0e . Statistical significance of different groups was assessed by using non-parametric tests ( all figure legends describe the analysis used ) . P<0 . 05 was considered statistically significant . | Approximately 2 billion people are infected with worms every year , causing physical , nutritional and cognitive impairment particularly in children . Mucins are large sugar-coated ( glycosylated ) proteins that form the intestinal mucus layer . This mucus layer protects our ‘insides’ from external insults and plays an important role during worm infection . We discovered that there is a difference in the glycosylation of mucins in people infected with worms compared to uninfected individuals . Therefore , using different mouse models we investigated the role of glycosylation , and in particular sulphation of mucins in infection . We found that mucin glycosylation is controlled by the immune response and increased sulphation correlated with the expulsion of the worm from the host . Highly sulphated mucins were protected from degradation by the worm . Moreover , mice lacking a sulphate transporter had significantly lower sulphation levels on mucins , which resulted in a reduction in the establishment of the worms and chronic infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"chemical",
"compounds",
"helminths",
"salts",
"parasitic",
"diseases",
"animals",
"animal",
"models",
"model",
"organisms",
"experimental",
"organism",
"systems",
"glycosylation",
"digestive",
"system",
"research",
"and",
"analysis",
"methods",
"mucus",
"proteins",
"sulfates",
"chemistry",
"mouse",
"models",
"gastrointestinal",
"tract",
"biochemistry",
"mucin",
"helminth",
"infections",
"anatomy",
"post-translational",
"modification",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"glycobiology",
"organisms"
]
| 2017 | Immune-driven alterations in mucin sulphation is an important mediator of Trichuris muris helminth expulsion |
The Huntington's disease gene ( HTT ) CAG repeat mutation undergoes somatic expansion that correlates with pathogenesis . Modifiers of somatic expansion may therefore provide routes for therapies targeting the underlying mutation , an approach that is likely applicable to other trinucleotide repeat diseases . Huntington's disease HdhQ111 mice exhibit higher levels of somatic HTT CAG expansion on a C57BL/6 genetic background ( B6 . HdhQ111 ) than on a 129 background ( 129 . HdhQ111 ) . Linkage mapping in ( B6x129 ) . HdhQ111 F2 intercross animals identified a single quantitative trait locus underlying the strain-specific difference in expansion in the striatum , implicating mismatch repair ( MMR ) gene Mlh1 as the most likely candidate modifier . Crossing B6 . HdhQ111 mice onto an Mlh1 null background demonstrated that Mlh1 is essential for somatic CAG expansions and that it is an enhancer of nuclear huntingtin accumulation in striatal neurons . HdhQ111 somatic expansion was also abolished in mice deficient in the Mlh3 gene , implicating MutLγ ( MLH1–MLH3 ) complex as a key driver of somatic expansion . Strikingly , Mlh1 and Mlh3 genes encoding MMR effector proteins were as critical to somatic expansion as Msh2 and Msh3 genes encoding DNA mismatch recognition complex MutSβ ( MSH2–MSH3 ) . The Mlh1 locus is highly polymorphic between B6 and 129 strains . While we were unable to detect any difference in base-base mismatch or short slipped-repeat repair activity between B6 and 129 MLH1 variants , repair efficiency was MLH1 dose-dependent . MLH1 mRNA and protein levels were significantly decreased in 129 mice compared to B6 mice , consistent with a dose-sensitive MLH1-dependent DNA repair mechanism underlying the somatic expansion difference between these strains . Together , these data identify Mlh1 and Mlh3 as novel critical genetic modifiers of HTT CAG instability , point to Mlh1 genetic variation as the likely source of the instability difference in B6 and 129 strains and suggest that MLH1 protein levels play an important role in driving of the efficiency of somatic expansions .
Huntington's disease ( HD ) is a fatal , dominantly inherited neurodegenerative disease , which is caused by the expansion of a CAG repeat within exon 1 of the HTT gene , resulting in an extended glutamine tract in the huntingtin protein ( HTT ) [1] . The length of the longer CAG repeat tract is the primary determinant of age of disease onset [2] . However , precise disease expression and timing are clearly modifiable by other factors , with strong evidence supporting the contribution of genetic factors [3] , [4] . The identification of such factors could lead to the development of novel therapeutic interventions that modify the nature and/or pace of the HD-associated pathogenic process , and is being pursued via a number of candidate and global genetic approaches [5] . The expanded HTT CAG repeat is highly unstable both in the germline and in somatic tissues [6]–[13] . In somatic tissues instability is expansion-biased and prevalent in brain regions that are most susceptible to neurodegeneration [7] . Approximately 10% of expanded HTT CAG alleles are further expanded by at least 10 repeats in human HD postmortem brain , with dramatic increases of up to 1 , 000 repeats also occurring , albeit at a lower frequency [7] , [11] . Given the strong CAG length-dependence of disease onset and severity , somatic expansion is predicted to accelerate the disease process . Mathematical modeling has suggested a mechanism by which somatic expansion beyond a threshold repeat length is required for clinical onset [14] . Whether in fact somatic expansion beyond a typically inherited repeat length of 40–50 CAGs is required for disease onset is unclear . Nevertheless the hypothesis that somatic expansion is at least a disease modifier is supported by the finding that longer somatic HTT CAG expansions are associated with an earlier residual disease onset ( onset unexplained by inherited CAG length ) in HD patients [11] . These data suggest that factors that modify somatic instability will also modify disease and could be targeted to delay onset or progression of HD . Identification of modifier genes in the mouse has the potential to provide insight into disease pathways at the earliest stages of the pathogenic process . To study mechanisms of HTT CAG instability and pathogenesis in the mouse we have developed a series of accurate genetic Huntington's disease homologue ( Hdh or Htt ) CAG knock-in mice [15]–[17] that provide powerful tools to uncover genetic modifiers of early dominant , HTT CAG length-dependent events . Using candidate gene knockout approaches we have found that Msh2 and Msh3 genes , encoding a key mismatch recognition complex designated MutSβ ( MSH2–MSH3 heterodimer ) , are essential for somatic HTT CAG expansion in HdhQ111 knock-in mice [18]–[20] . Similar studies using various mouse models of HD and other trinucleotide repeat diseases support a central role for the mismatch repair ( MMR ) pathway in somatic instability [21]–[28] . While the effects of MMR proteins on instability can vary according to the repeat sequence and its context [21]–[28] , it is notable that Msh2 and Msh3 enhance CAG/CTG expansion both in HD and DM1 mouse models [18]–[23] , [25]–[27] , and Pms2 , encoding a subunit of the MutLα ( MLH1-PMS2 ) complex that acts downstream of mismatch recognition by MutSα ( MSH2–MSH6 heterodimer ) or MutSβ , was identified as a genetic enhancer of CTG expansion in a DM1 mouse model [24] . These observations highlight underlying similarities of the CAG/CTG expansion process across disease loci . Importantly , in HdhQ111 mice Msh2 and Msh3 promote HTT CAG-dependent mutant huntingtin diffuse nuclear localization and nuclear inclusion phenotypes . While the relationship between instability and nuclear huntingtin localization/inclusion phenotypes is correlative , these data support the hypothesis that somatic expansions contribute to an ongoing HTT CAG-dependent process [18]–[20] . An alternative approach for identifying modifiers in the mouse is to take advantage of naturally occurring strain-specific phenotypic variation . Interestingly , mouse strain-specific differences in trinucleotide repeat instability [17] , [22] , [29] and various HD mouse model phenotypes [17] , [30] , [31] have been identified . Notably , strain-specific differences in the instability of the HTT CAG repeat in R6/1 transgenic mice were recently found to be associated with polymorphisms in the Msh3 gene [29] . With the aim of performing unbiased genetic screens for HTT CAG-dependent phenotypes we have generated congenic HdhQ111 mice on several different genetic backgrounds [17] . In a comparison of congenic B6 . HdhQ111 , FVB . HdhQ111 and 129 . HdhQ111 strains we previously showed that intergenerational HTT CAG instability , somatic HTT CAG instability , diffusely immunostaining nuclear huntingtin and intranuclear inclusions in striatal neurons were modified by genetic background [17] , providing the opportunity to perform unbiased searches for genetic modifiers of HTT CAG-dependent events . Here , we set out to perform a genetic linkage study with the aim of mapping genetic modifier ( s ) of somatic HTT CAG instability in HdhQ111 mice , in order to gain further insight into factors underlying somatic instability with the potential to uncover novel targets for slowing somatic instability and/or early events in the HD pathogenic process .
Our previous qualitative analyses revealed high and low levels of HTT CAG instability in striata from B6 . HdhQ111/+ and 129 . HdhQ111/+ mice , respectively , at both 10 and 20 weeks of age [17] . At 10 weeks of age B6 . HdhQ111/+ striata display a broadened and expansion-biased CAG length distribution , in contrast to 129 . HdhQ111/+ mice that display very low levels of somatic expansion ( Figure 1A and [17] ) . By 20 weeks of age a bimodal CAG length distribution is apparent in B6 . HdhQ111/+ striata , while 129 . HdhQ111/+ show a broadened CAG distribution similar to that in B6 . HdhQ111 striata at 10 weeks of age ( Figure S1 and [17] ) . We were interested in identifying early-acting modifiers of instability , and therefore we determined whether the difference in instability in B6 and 129 strains at 10 weeks of age could be captured as a quantitative trait for genetic mapping experiments . We thus quantified a somatic “instability index” from GeneMapper traces of PCR-amplified HTT CAG repeats from B6 . HdhQ111/+ and 129 . HdhQ111/+ striata using a previously described method [32] . In addition , given the observation of high levels of HTT CAG instability in the liver of CD1 . HdhQ111/+ mice [33] , we also quantified instability indices in B6 . HdhQ111/+ and 129 . HdhQ111/+ livers . In concordance with our previous qualitative assessment [17] , the quantification of instability in striatum and liver of 10-week-old mice revealed significantly higher levels in B6 . HdhQ111/+ mice compared to 129 . HdhQ111/+ mice ( 2-tailed unpaired t-test: p<0 . 0001 for both striatum and liver; Figure 1B ) . Note that there was a significant difference in the constitutive CAG repeat size between these B6 and 129 mice ( 2-tailed unpaired t-test: p<0 . 0001; Figure S2 ) . While CAG length could , in principle , account for at least some of the difference in instability between strains , our previous analyses demonstrated a strain-specific difference in instability that was unaccounted for by CAG size alone [17] , strongly indicating that identification of additional instability modifiers would be plausible . Striatal instability indices from the two strains were quite distinct ( Figure 1B and Figure 2A ) , indicating that the instability index was likely to provide a sensitive quantitative trait for mapping genetic modifiers . Liver instability indices were less well separated between the two strains ( Figure 1B ) , predicting less power in the ability to identify genetic modifiers of liver instability than striatal instability . Based on the findings above we used striatal instability index , which showed very good separation between B6 and 129 strains , as a quantitative phenotype for linkage mapping . Analyses of HTT CAG instability in striata from ( B6x129 ) . HdhQ111/+ F1 mice showed comparable instability indices to those in B6 . HdhQ111/+ mice ( 2-tailed unpaired t-test: p = 0 . 11 ) , and significantly higher instability indices than in 129 . HdhQ111/+ mice ( 2-tailed unpaired t-test: p<0 . 0001 ) ( Figure 2A ) , suggesting the presence of a B6 genetic locus or loci that dominantly enhance HTT CAG expansion . While these data were consistent with a dominant B6 modifier ( s ) we established an F2 intercross in order to search in an unbiased manner for both dominant and recessive modifier loci [34] . Instability indices were quantified from the striata of 69 10-week-old ( B6x129 ) . HdhQ111/+ F2 animals ( Figure 2A ) . We observed no correlation between constitutive CAG size and striatal CAG instability in the F2 intercross mice ( Pearson correlation: R2 = 0 . 011 , p = 0 . 39 ) , implying the contribution of other genetic factors to the difference in HTT CAG instability between the two strains . Note that the genetic background of the region surrounding the HdhQ111 allele in both strains is 129 due to the etiology of the targeted ES cells , ruling out the possibility of identifying cis-acting modifiers . The F2 intercross mice were genotyped using an initial panel of 117 SNPs that distinguishes B6 and 129 strains ( Figure S3 and Table S1 ) . Linkage analysis identified a single quantitative trait locus ( QTL ) on chromosome 9 associated with striatal HTT CAG instability with a peak LOD score of approximately 11 ( Figure S4 ) . Notably , the MMR gene Mlh1 is located within this interval ( Figure S5 ) . As MMR genes Msh2 and Msh3 had been previously established as modifiers of somatic CAG repeat expansion in HdhQ111 mice [18]–[20] , additional members of this pathway would be strongly indicated as potential modifiers . In an attempt to primarily enhance resolution at this QTL , but also to specifically investigate the Mlh1 gene , we genotyped the F2 animals for 10 additional markers distributed across the QTL region , including two markers located within the Mlh1 gene ( Figure S3 and Table S1 ) . We also genotyped additional markers to improve overall genome coverage and specifically the coverage of the Msh2 and Msh3 genes . Subsequent linkage analysis that included these additional markers ( total 147 SNPs ) not only confirmed the mapping of a single QTL on chromosome 9 ( Figure 3 ) , but also significantly narrowed down the implicated genomic region to an interval of approximately 5 Mb ( chr9:107 , 982 , 655–113 , 057 , 967; GRCm38/mm10 ) ( Figure S6 ) . This genomic region , which represents a 95% confidence interval , is defined by the markers encompassing a 2-LOD drop-off from the peak LOD score [35] . Interestingly , the markers at the Mlh1 locus defined the QTL peak , which was significantly increased to a LOD score of approximately 14 ( Figure 3 and Figure S6 ) . We did not find any evidence for linkage to the Msh2 or Msh3 genes on chromosomes 17 and 13 , respectively ( Figure 2B and Figure 3 ) . Note that constitutive CAG repeat lengths in the F2 mice did not cluster with genotype at the Mlh1 locus ( Figure S2 ) , consistent with the lack of correlation between constitutive CAG length and instability index in these mice . The chromosome 9 QTL explains approximately 60% of the variance in striatal instability , with the remaining 40% of the variance being attributable to differences within the parental strains , strongly supporting this locus as the single major modifier of instability between the two strains . Further , the effect of the QTL was consistent with the B6 allele acting in a dominant fashion ( Figure 2 ) . In addition to Mlh1 , the implicated genomic region contains numerous genes ( Figure S6 ) , none of which we are able to objectively exclude as a modifier based on our genetic data . However , none of these genes has been shown or is suspected to be involved in repeat instability . Past observations that the MMR pathway plays a major role in modulating somatic HTT CAG instability , together with the highest LOD scores observed with two markers that were located within the Mlh1 gene , strongly suggest that this MMR gene is a likely candidate modifier underlying the chromosome 9 QTL . Based on our above findings we hypothesized that Mlh1 was a modifier of somatic HTT CAG expansion . Therefore , to investigate the role of the Mlh1 gene in somatic HTT CAG expansion we crossed B6 . HdhQ111 and Mlh1 null mice ( B6 ) [36] and quantified CAG repeat size distributions in tail , striatum and liver of 22-week-old B6 . HdhQ111/+ animals on Mlh1+/+ , Mlh1+/− and Mlh1−/− genetic backgrounds ( Figure 4 ) . By 22 weeks a bimodal repeat size distribution was apparent both in striata and liver of Mlh1+/+ mice , as previously shown [33] . Mlh1+/− mice exhibited similar levels of instability in striatum and liver to those in Mlh1+/+ mice ( 2-tailed unpaired t-tests: striatum , p = 0 . 30; liver , p = 0 . 47 ) . However , no instability was present in either of these tissues in Mlh1−/− mice ( 2-tailed unpaired t-test: p<0 . 0001 compared to Mlh1+/+ ) . These findings demonstrate that Mlh1 is absolutely required for somatic HTT CAG expansions in B6 . HdhQ111 mice , and provide compelling evidence that genetic differences between B6 and 129 strains at the Mlh1 gene are likely to underlie the difference in somatic instability between these two strains . Note that the effect of the Mlh1 knockout is to eliminate somatic HTT expansion at 22 weeks of age , while the 129 genetic background results in reduced somatic expansion at the same age ( Figure S1 ) . Therefore , if Mlh1 genetic variants do indeed underlie the difference in striatal instability between B6 and 129 strains , such variants are likely to confer a moderate effect on MLH1 . We have previously shown that deletion of mismatch repair genes Msh2 or Msh3 is sufficient to delay the accumulation/epitope accessibility of diffusely immunostained mutant huntingtin in the nuclei of striatal neurons [18]–[20] . This early phenotype , which is both dominant and CAG length-dependent [16] , is a sensitive marker of the ongoing pathogenic process in these mice . To determine whether Mlh1 also modified this phenotype we quantified diffusely-immunostained nuclear huntingtin in striatal sections of 22-week-old B6 . HdhQ111/+ animals on Mlh1+/+ , Mlh1+/− and Mlh1−/− genetic backgrounds ( Figure 5 ) . Nuclear huntingtin immunostaining intensity was reduced in Mlh1+/− striata to approximately 60% of Mlh1+/+ levels , although this difference did not reach statistical significance ( 2-tailed unpaired t-test: p = 0 . 06 ) . In Mlh1−/− striata nuclear huntingtin immunostaining intensity was dramatically reduced to approximately 18% of Mlh1+/+ levels ( 2-tailed unpaired t-test: p = 0 . 0018 ) . Together , these findings reveal Mlh1 as a genetic enhancer both of somatic expansion and of an early CAG length-dependent phenotype in B6 . HdhQ111/+ mice , supporting the hypothesis that somatic expansion accelerates HTT CAG-dependent events . Given the critical role of MLH1 in somatic HTT CAG expansion we were interested in investigating further this MLH1-mediated pathway . It is known that MLH1 is an obligate subunit of three MutL complexes: MutLα ( MLH1-PMS2 ) , MutLβ ( MLH1-PMS1 ) and MutLγ ( MLH1–MLH3 ) ( reviewed in [37] , [38] ) . These MutL heterodimers are essential downstream factors in MMR and are recruited to the MMR reaction following the binding of mismatched DNA by MutSα ( MSH2–MSH6 ) or MutSβ ( MSH2–MSH3 ) . Outside of its role in meiotic recombination [39] , MutLγ appears to function predominantly with MutSβ both in somatic and germ cells [40] , [41] . Given the specific requirement for MutSβ in somatic CAG expansion in HdhQ111 mice [19] and other mouse models of CAG/CTG disease [22] , [25] , [26] , we hypothesized that MLH3 may also play a major role in somatic expansion . A role for MLH3 had also been suggested from findings in a mouse model of myotonic dystrophy type 1 in which knockout of Pms2 , encoding MLH1's major binding partner , reduced the rate of somatic CTG expansion by ∼50% , but did not eliminate somatic expansions [24] . We therefore crossed B6 . HdhQ111 with Mlh3 null mice ( B6 ) [39] and quantified HTT CAG repeat size distributions in the tail , striatum and liver of 24-week-old B6 . HdhQ111/+ animals on Mlh3+/+ , Mlh3+/− and Mlh3−/− genetic backgrounds ( Figure 6 ) . Slightly reduced striatum- and liver-specific CAG instability was observed in Mlh3+/− mice when compared to Mlh3+/+ animals ( 2-tailed unpaired t-tests: striatum , p = 0 . 06; liver , p = 0 . 03 ) . Interestingly , no instability was present in Mlh3−/− striatum or liver ( 2 tailed unpaired t-tests: p<0 . 0001 for both tissues compared to Mlh3+/+ ) , demonstrating , as for MLH1 , that MLH3 is absolutely required for somatic HTT CAG instability in B6 . HdhQ111 mice , and implying that MutLγ dimers act in this process . The slight reduction of instability in Mlh3+/− mice ( Figure 6 ) , not apparent in Mlh1+/− mice ( Figure 4 ) suggests that Mlh3 may be a limiting factor in somatic expansion , as previously reported for Msh3 [19] , [26] . The relatively strong impacts of heterozygous loss of Mlh3 and Msh3 compared to heterozygous loss of the Mlh1 and Msh2 genes encoding their respective binding partners may be explained in part by the lower levels of MSH3 compared to MSH2 and of MLH3 compared to MLH1 [42] , [43] . While our linkage peak contained many genes , given the finding that Mlh1 is necessary for somatic HTT CAG expansion , we focused on this gene as the most likely candidate modifier at the linked chromosome 9 locus . We initially investigated polymorphisms at the Mlh1 locus between C57BL/6NCrl and 129S2/SvPasCrlf strains ( in which the QTL mapping was carried out ) by sequencing all Mlh1 exons as well as the immediate 5′ and 3′ flanking regions ( 2 . 6 kb and 2 kb respectively ) . A relatively high frequency of SNPs was identified in the 5′UTR of Mlh1 ( 8 SNPs in an 84 bp region ) , and a single SNP was detected in the 3′UTR ( Table 1 ) . We also identified 14 exonic SNPs , 4 of which result in an amino acid change: F192I , E390D , G404V and M528I ( Figure 7 ) . A subsequent investigation of the Mlh1 locus in the highly related C57BL/6NJ and 129S1/SvImJ strains using whole genome sequencing data from the Mouse Genomes Project [44] , [45] confirmed all of the B6-129 polymorphisms initially identified by us by Sanger sequencing . It also resulted in the identification of a large number of additional polymorphims between B6 and 129 strains , dispersed throughout the entire Mlh1 locus ( Table 1 and Figure S7 ) . In total , 642 polymorphisms were identified in a 64 kb region encompassing the Mlh1 gene , averaging approximately 10 polymorphims per kb . In comparison to the average genome wide variation between B6 and 129 strains of 2 . 4 polymorphisms per kb the Mlh1 gene exhibits a high degree of variation , with only 5 . 9% of the genome displaying a relative density greater than or equal to 10 polymorphism per kb ( see Materials and Methods and [44] ) . It is noteworthy that the haplotype across this 64 kb region in FVB/N and DBA/2J strains that display similar high somatic HTT CAG instability to B6 strains is highly similar to the B6 haplotype ( Figure S7 and Figure S8 ) . While this finding was consistent with a B6-like haplotype at the Mlh1 locus underlying high instability , the relatedness of the B6 , FVB/N and DBA/2J haplotypes did not provide the means to further refine the putative instability-associated region ( s ) . All 4 nonsynonymous SNPs are suspected to be in key protein domains: F192I falls within the putative ATP binding domain , though outside conserved ATP binding motifs [46]; E390D and G404V are within a domain thought to be necessary for interaction with MSH3 [41] , and M528I is in a region implicated in interaction with MSH3 , EXOI , MLH3 , PMS1 and PMS2 [41] ( Figure 7A ) . Note that none of these variants has been identified in human MLH1 [47] . Cross-species alignment of MLH1 proteins ( Figure 7B ) shows that the Phe residue at aa192 of the B6 MLH1 protein was fully conserved across the organisms investigated , with an Ile residue at this position present in 129 strains . At positions 390 and 528 the B6-like amino acid is highly conserved , mainly in higher organisms , while the 129-like amino acid at position 528 is also well represented , particularly among lower organisms . In contrast , aa404 is poorly conserved . While none of the SNP variants alters the general chemical similarity of the amino acids , the conservation data indicate that the F192I substitution may have a functional impact . This is supported by PolyPhen-2 analysis [48] predicting E390D , G404V and M528I to be “benign” , but predicting the F192I mutation to be “probably damaging” with a maximum confidence score . The highly polymorphic nature of the Mlh1 gene indicated that delineation of the functional polymorphism ( s ) that drives the difference in instability between B6 and 129 mice may well be complex . However , based on the above prediction that at least the F192I substitution may have a functional impact we tested the simplest hypothesis that the B6 and 129 versions of MLH1 have different levels of activity . As there is currently no good assay for MutLγ function , we performed cell-free assays using MutLα ( MLH1-PMS2 complexes ) , known to be required to repair G-T mismatches and single repeat slip-outs of CAG/CTG tracts [49] , [50] , in order to provide the most sensitive test of B6 and 129 MLH1 function . We thus cloned and co-purified B6-like ( mMLH1 . B6-hPMS2 ) and 129-like ( mMLH1 . 129-hPMS2 ) MutLα proteins ( Figure S11 ) and assessed the ability of these proteins ( containing all 4 amino acid differences; Figure 6A ) to repair various DNA substrates using cell-free assays . The results revealed that B6 and 129 MLH1 proteins displayed no overt difference in their abilities to repair a G-T mismatch ( Figure S12 ) . In addition , the human MLH1 protein carrying the F192I mutation showed MMR activity comparable to that of wild-type human MLH1 ( Figure S12 ) . We then tested the ability of B6 and 129 MLH1 proteins to repair a single CTG slip-out ( CAG ) 47• ( CTG ) 48 [50] , [51] , a potential intermediate in the expansion process , as requirements for processing of slipped-DNAs formed by trinucleotide repeats may more closely resemble those that ultimately result in CAG expansion in mice . As shown previously [50] , complementation of MLH1- and PMS2-deficient HEK293T cells with wild-type human MutLα restored repair activity ( Figure 8A ) . Complementation with mMLH1 . B6-hPMS2 or mMLH1 . 129-hPMS2 MutLα complexes also restored repair to similar efficiencies ( Figure 8A ) . Titration of concentration of the B6-like and 129-like MutLα complexes confirmed similar repair efficiencies between the MLH1 protein from the two mouse strains at each concentration ( 2-tailed unpaired t-tests: 5 ng , p = 0 . 477; 25 ng , p = 0 . 885; 100 ng , p = 0 . 736 ) , but also demonstrated a statistically significant MutLα dose dependency of CTG slip-out repair ( linear regression: R2 = 0 . 557 , p = 0 . 0004; Figure 8B ) . Together , these results demonstrate that B6 and 129 MLH1 proteins , in the context of the mixed-species MutLα complex , do not differ substantially in their G-T mismatch or CTG slip-out repair activities and that the F192I mutation in the human protein does not have a significant functional impact . This suggests that if Mlh1 gene variations are in fact the source of the CAG repeat instability differences between the B6 and 129 mouse strains in vivo , this is unlikely to be due to major differences in MLH1 protein activity within the context of the MutLα complex . However , the dose-dependence of the MutLα complex in the CTG slip-out repair assay indicated that differential MLH1 protein levels between the two strains may be relevant to their different levels of instability in vivo . The cell-free CTG slip-out repair assays suggested that levels of MLH1 may impact the ability of MutL complexes to execute a repair process that results in CAG expansion in vivo . We therefore assessed whether Mlh1 expression levels differed between the B6 and 129 strains that exhibit comparatively high and low HTT CAG instability , respectively . Striatal Mlh1 mRNA amount was significantly reduced in 129 mice to 54% of that in B6 mice ( 2-tailed unpaired t-test: p = 0 . 017 ) , reaching approximately the same mRNA level as that in B6 . Mlh1+/− mice ( Figure 9A ) . Striatal Mlh1 mRNA levels were consistently reduced in 129 mice across 3 distinct regions of the primary Mlh1 transcript ( exons 4–5 , 11–12 , and 18–19 ) , and in various other tissues ( cerebellum , liver , jejunum and ileum ) to between 25% and 50% of B6 levels ( Figure S13 ) . Analysis of MLH1 protein by western blot showed similarly reduced protein levels in 129 compared to B6 striata ( Figure 9B , C ) . In contrast to the mRNA , however , the MLH1 protein level in B6 . Mlh1+/− mice was intermediate between that in B6 ( Mlh1+/+ ) and 129 striata ( Figure 9B , C ) . We were unable to detect any evidence for novel isoforms or truncation products in the 129 mice ( Figure S14 ) . Given the difference in Mlh1 mRNA levels between B6 and 129 strains we investigated possible polymorphisms that might underlie this difference . As we had identified polymorphisms in both 5′ and 3′ regulatory regions of Mlh1 ( Table 1 and Figure S7 ) we tested whether either the immediate 5′- or 3′-flanking regions ( 2 . 4 kb and 1 . 7 kb , respectively ) of either the B6 or 129 Mlh1 gene were able to drive differential steady state levels of a luciferase reporter gene ( Figure 10 ) . As shown in Figure 10A there was no significant difference in firefly luciferase activity when either the B6 5′ region or the 129 5′ region was used to drive firefly luciferase expression ( 2-tailed unpaired t-test: p = 0 . 18 ) . In contrast , when the 3′ region was cloned downstream of the firefly luciferase gene ( Figure 10B , panel i ) , whose expression was driven from the SV40 promoter , the 129 3′ region resulted in a ∼2-fold reduction in firefly luciferase activity compared to the B6 3′ region ( 2-tailed unpaired t-test: p = 0 . 012 ) . These results suggest that polymorphisms in this 3′ genomic region may be relevant to the ∼2-fold reduction of Mlh1 mRNA seen in vivo in 129 mice compared to B6 mice ( Figure 9 ) . In an effort to narrow down the polymorphisms within this region that contributed to the differential luciferase expression we performed further luciferase reporter assays in which the 3′ genomic region from either strain was either successively deleted ( Figure 10B , panels ii–iv ) or in which the original 1 . 7 kb 3′ region from the B6 Mlh1 gene was substituted with different subdomains of 129 genomic sequence ( Figure 10B , panel v ) . The deletion experiments ( panels ii , iii , iv ) indicated that neither the single polymorphism within the 3′UTR ( Figure 10B , panel iv ) , nor the 3′ most 4 polymorphisms ( Figure 10 , panel ii ) contributed to the differential firefly luciferase expression . The data indicated that polymorphisms both in the 129 3′ genomic region from 205 bp to 591 bp ( panel iii ) and in the genomic region from 591 bp to 1 , 280 bp ( panels ii and iii ) contributed to the 2-fold reduction in firefly luciferase activity . The domain “swap” experiments ( panel v ) showed partial reduction of firefly luciferase activity when each of three B6 genomic regions was substituted with 129 sequence , confirming the contribution of multiple 3′ polymorphisms to the differential firefly luciferase activity . Taken together , the results of our expression analyses indicate that genetic differences between B6 and 129 strains result in lower steady state Mlh1 mRNA levels in 129 compared to B6 mice . Luciferase reporter assays suggest that this may , at least in part , be driven by a combination of polymorphisms 3′ to the Mlh1 coding region . In addition , the lower relative level of MLH1 protein in 129 versus B6 . Mlh+/− striata despite similar Mlh1 mRNA levels ( Figure 9 ) further suggests that genetic differences between these strains also act post-transcriptionally . While we currently have no good evidence for altered protein isoforms/truncation products in 129 versus B6 mice , the high degree of variation at the Mlh1 locus suggests that mechanisms that might impact the levels of full-length protein in 129 mice , including altered mRNA splicing , warrant further investigation . Our data indicate , therefore , that the low HTT CAG instability in 129 versus B6 mice may be in part driven by reduced levels of MLH1 protein . These findings are consistent with the strong genetic linkage of an instability modifier to the Mlh1 gene and indicate that B6 versus 129 variants may act in multiple ways to ultimately determine the different MLH1 protein levels in these strains .
Here we report the first unbiased QTL mapping study in a mouse model of Huntington's disease , in which we have mapped a locus that modifies the somatic expansion of the HTT CAG repeat . Using a quantitative measure of striatal HTT CAG instability we were able to detect a single modifier locus of large effect using as few as 69 F2 intercross mice . These results indicate that , depending on the number and effect size of the modifier loci , an intercross mapping strategy in congenic HdhQ111 strains is a potentially powerful approach that could be applied to identify modifiers of a variety of HTT CAG-dependent phenotypes . While our genetic data do not exclude a role for other gene ( s ) within the linked locus as instability modifiers , the high LOD score observed with markers positioned over the Mlh1 gene and the knowledge that this gene is essential for somatic HTT CAG instability provide compelling evidence that Mlh1 is the likely genetic modifier underlying the difference in striatal HTT CAG instability between the B6 and 129 HdhQ111 mice . Further experiments would be needed to determine whether the same QTL contributes to the difference in liver instability between B6 and 129 strains , and/or whether other genetic loci might play a role . Two additional genes , Trex1 and Atrip , located within the 2 LOD drop-off interval , are involved in DNA repair [52] , [53] . However , in a comparison with two additional unstable strains , FVB . HdhQ111 and DBA . HdhQ111 ( Figure S8 ) , we note that Trex1 and Atrip polymorphisms do not correlate with the instability phenotype ( Figure S9A , B ) . Further , Trex1 and Atrip striatal mRNA levels are not significantly different in 129 and B6 strains ( 2-tailed unpaired t-test: p = 0 . 73 and p = 0 . 43 , respectively ) ( Figure S9C ) . While these data do not rule out a role for these genes , these observations make them less compelling candidates as the likely modifiers of strain-specific instability . In contrast , the observation that a “B6-like” haplotype at the Mlh1 locus is also shared in unstable FVB . HdhQ111 and DBA . HdhQ111 strains ( Figure S7 and Figure S8 ) is consistent with the hypothesis that genetic variation at the Mlh1 locus underlies the difference in striatal HTT CAG instability between B6 and 129 strains . This hypothesis also predicts that strains with a “129-like” Mlh1 haplotype might be more likely to exhibit low HTT CAG instability . It is important to note , however , that somatic instability in any particular strain background is likely to be influenced by other genetic variation . Notably , the Mlh3 gene ( chromosome 12 ) , found to be a modifier of CAG instability in this study , does not show genotype differences between B6J and 129S1 strains [44] , which are closely related to the B6N and 129S2 strains used here . Therefore , linkage to the Mlh3 gene would not be expected in our genetic cross . Interestingly , Msh3 gene variants were recently found to correlate with HTT CAG instability in some strains of R6/1 transgenic mice [29] . However , at least for the B6N and 129S2 strains in which we have performed genome-wide QTL mapping , it is clear from the genetic data that any polymorphisms in the Msh3 gene do not play a significant role in driving these strain-specific differences in somatic expansion of the HdhQ111 CAG repeat ( Figure 2B ) . To understand this further we compared non-synonymous Msh3 SNPs , proposed to underlie the difference in CAG instability between B6 ( high instability ) and BALB/cJ ( low instability ) R6/1 mice [29] , in strains ( B6 , 129 , FVB and DBA ) for which we had quantitative measures of HdhQ111 striatal instability ( Figure S8 ) . Notably B6-BALB/cJ SNPs that are present in 129 and that might be predicted to contribute to low instability in HdhQ111 mice ( those in exons 2 , 3 and 7 ) are also present in unstable FVB and DBA strains ( Figure S10A ) . This suggests that these SNPs are unlikely to contribute to the differences in HdhQ111 CAG instability between B6 and 129 striata . We also note a very high degree of B6 versus BALB/cJ genetic variation relative to B6 versus 129 genetic variation at the Msh3 locus ( Figure S10B ) , suggesting the possibility that the apparently complete CAG repeat stabilization in BALB/cJ . R6/1 mice [29] is driven by a Msh3 polymorphism ( s ) present in BALB/cJ but not in 129 . It is also noteworthy that a single 129 allele increases the instability of the R6/1 CAG repeat in BALB/129 heterozygotes , consistent with higher levels of MSH3 in 129 mice than in BALB/cJ mice [29] . Despite possible locus-specific ( HdhQ111 versus R6/1 mice ) and sub-strain differences , the data presented here and previously [29] suggest that the combination of genetic variants in Mlh1 , Msh3 , and potentially other MMR genes that are present in any particular mouse strain may determine the rate of CAG expansion in certain tissues . Given that MLH1 protein levels correlate with striatal expansion in B6 and 129 strains and that the activity of MLH1-dependent DNA repair in cell-free assays is dose-dependent , it is more than plausible to hypothesize that the reduced levels of Mlh1 expression in 129 mice play an important role in determining the reduced somatic CAG instability observed in HdhQ111 mice in this genetic background . Given the finding that Mlh1 is an enhancer of nuclear huntingtin immunostaining , it is also possible that the lower levels of MLH1 in 129 mice contribute to the slowed nuclear huntingtin and inclusion phenotypes previously identified in 129 . HdhQ111/+ mice compared to B6 . HdhQ111/+ mice [17] . Further unbiased genetic studies would be needed to identify the modifier gene ( s ) that contribute to these phenotypes . It is worth noting that a number of other studies support a role for the levels or stoichiometries of DNA repair proteins in trinucleotide repeat instability [43] , [54]–[57] . Expression analyses of MLH1 mRNA and protein in B6 and 129 strains ( Figure 9 and Figure 10 ) indicate that strain-specific polymorphisms may act at both transcriptional and post-transcriptional levels . Assuming that B6 . Mlh1+/− and B6 . Mlh1+/+ striata display comparable levels of instability at 10 weeks of age , as seems likely from the similar levels of instability in B6 . Mlh1+/− and B6 . Mlh1+/+ mice at 22 weeks of age ( Figure 4 ) , a comparison of somatic instability and MLH1 protein in B6 . Mlh1+/+ , 129 . Mlh1+/+ and B6 . Mlh1+/− striata ( Figure 1 , Figure 4 , Figure S1 , and Figure 9 ) suggests that there may be a threshold level of MLH1 protein below which MLH1-dependent process ( es ) that mediate expansion are compromised . In this scenario , MLH1 protein in B6 . Mlh1+/− mice , although reduced compared to that in B6 . Mlh1+/+ mice , exceeds this threshold , with the result that the HTT CAG repeat remains unstable . In 129 mice , the MLH1 protein level falls below the threshold and the HTT CAG repeat is consequently stabilized . Alternatively , it is possible that reduced MLH1 protein alone is insufficient to explain the HTT CAG repeat stabilization in 129 mice , but that a functional alteration of the 129 protein acts in concert with the reduced expression level to decrease HTT CAG expansion efficiency . Although we were unable to demonstrate any difference in activity between B6 and 129 recombinant MLH1 proteins in cell-free MMR assays ( Figure 8 and Figure S12 ) , these assays may not be sufficiently sensitive to detect subtle alterations in function . It is also important to note that the MMR ability of MLH1 was only investigated in the context of MutLα-mediated repair . Therefore , taking into account our finding that MLH3 is essential for somatic HTT CAG instability in vivo , we cannot rule out the hypothesis that B6 and 129 MLH1 proteins may have dissimilar MutLγ-mediated repair potential . It is also possible that MLH1 function may differ between B6 and 129 strains in other ways in vivo that cannot be captured in the cell-free systems , e . g . altered interaction with binding partners . Thus , while our data indicate that MLH1 protein levels are likely to be a driving force in determining the differential HTT CAG somatic expansion potential in B6 and 129 strains , phenotypic comparisons between strains at the level of MLH1 mRNA , protein and HTT CAG instability , together with the highly polymorphic nature of the Mlh1 locus , suggest that the genetic architecture underlying the strain-specific differences in instability may be complex . MLH1 has been found to play a role in CAG repeat instability in a selectable cell-based system [58] . A functional form of MLH1 , with an intact ATPase domain , is also required to repair slipped CAG/CTG structures in vitro [50] ( Figure 8 ) . To our knowledge no role for MLH3 in trinucleotide repeat instability has been previously demonstrated . Here , we show for the first time that both Mlh1 and Mlh3 genes enhance HTT CAG expansion in a trinucleotide repeat disease mouse model . Our data further consolidate the critical role of MMR genes as enhancers of HTT CAG-dependent events [18]–[27] , [29] in HdhQ111 mice . We were unable to determine the effect of loss of Mlh1 or Mlh3 on intergenerational instability of the HTT CAG repeat in HdhQ111 mice as Mlh1 and Mlh3 null mice are sterile [36] , [39] . Interestingly , as with somatic instability , B6 . HdhQ111 mice show a greater degree of intergenerational CAG repeat instability than 129 . HdhQ111 mice [17] . Given evidence suggesting a role for MMR pathways in both somatic and intergenerational repeat instability [18] , [23] , [59] , it is plausible that genetic variation at the Mlh1 locus also underlies the difference in intergenerational instability between the two strains . The mechanism ( s ) by which MMR proteins mediate somatic CAG/CTG expansion is unclear . Importantly , we find that the MutLγ components , MLH1 and MLH3 , are as critical to somatic HdhQ111 CAG expansion as the MutSβ components MSH2 and MSH3 [18] , [19] , suggesting that MutLγ and MutSβ are involved in the same pathway that promotes CAG/CTG expansion . While a role for proteins downstream of MutL complexes in somatic CAG/CTG expansion has not been demonstrated to date , the requirements for MLH1 and MLH3 indicate that the generation of somatic HdhQ111 CAG expansions requires active engagement of the MMR machinery , in contrast to a model whereby expansions occur due to the inability of MutSβ-CAG/CTG repeat binding to execute coupling to downstream effector functions [25] , [60] . Our findings also argue against MutSβ-mediated expansion arising via other pathways that are MutL-independent , such as single strand annealing [23] , [61] . Our results support previously published studies in mouse models of DM1 in which somatic expansion of the CTG repeat was reduced in Pms2 null mice [24] or inhibited in mice deficient in MSH2's ATPase function , which is required for MutL complexes recruitment [27] . Recruitment of MutL complexes is a required step for subsequent enzymatic processing of the DNA mismatch [37] , [38] . An essential function of MutLα is the activation of the latent endonuclease activity of PMS2 [62] , which , interestingly , is activated by extrahelical CAG/CTG repeats in vitro [63] . It would therefore be of interest to determine whether MLH3's putative endonuclease domain [62] is required for CAG expansion in vivo . The MMR pathway , as traditionally described , is employed to repair errors that are incurred during DNA replication . However , there is increasing evidence that MMR proteins play various roles in the absence of DNA replication and participate in a variety of other pathways , distinct from MMR [64]–[69] . Recently , a promutagenic noncanonical MMR pathway has been described , which occurs in multiple cell types , is independent of DNA replication and is activated by DNA lesions rather than mismatches [70] . The findings that MMR proteins are required for , rather than protect against somatic CAG/CTG instability , that repeat expansions occur in postmitotic cells [10] , [33] , [71] and that expansions in neurons require MSH2 [20] , suggest that CAG/CTG repeat expansion may arise via a noncanonical MMR pathway ( s ) . With regard to potential mechanisms of CAG expansion it is of interest that MSH3 and MLH3 appear to play relatively minor roles in classical MMR inasmuch as Msh3 and Mlh3 deficiencies result in weak mutator phenotypes and relatively low cancer predisposition phenotypes [42] , [72]–[75] . In strong contrast , loss of either of these two proteins has a major impact on CAG/CTG expansion . Conversely , MSH6 and PMS2 play prominent roles in classical MMR [72]–[74] . However , MSH6 is either unnecessary for , or plays a very minimal role in mediating somatic CAG/CTG expansions [19] , [22] , [25] , and knockout of Pms2 had a moderate effect of CTG expansion in DM1 mice [24] , implicating a role for different MLH1 partners . In the present study the complete absence of HTT CAG expansion in HdhQ111/+ Mlh3 null mice argues against a role for PMS2 in generating expansions in these mice . Further genetic crosses in both DM1 and HdhQ111 mice would be needed to determine whether the relative contributions of Pms2 and Mlh3 genes in the two mouse models depends on the genomic locus of the repeat and/or strain background . While we do not expect PMS2 levels to be altered in Mlh3 knockout mice [74] , additional experiments are needed in Mlh3 and Pms2 knockout mouse tissues to determine whether any compensatory changes in PMS2 or MLH3 proteins , respectively , occur . However , overall , the data thus far indicate that MLH3 is a more significant player than PMS2 in CAG/CTG expansion and suggest that CAG/CTG repeats may preferentially engage a pathway ( s ) involving MutSβ and MutLγ complexes , as illustrated in Figure 11 . Given the overlapping roles of MMR proteins in both DM1 and HD mouse models [18]–[27] , [29] , the findings in the present study are predicted to be directly relevant both to DM1 and likely other CAG/CTG repeat expansion diseases . However , subtle qualitative and quantitative differences in the effects of MMR genes in the various mouse models suggest a potential modulatory role for the cis-sequence surrounding the repeat . In addition , proteins in base excision repair and nucleotide excision repair pathways have also been found to play role in mouse models of CAG/CTG expansion disorders [76]–[78] . Further studies will be needed to determine how the various DNA repair proteins might intersect to mediate CAG/CTG expansion and the extent to which their effects might depend on genomic context . In summary , we have taken both unbiased and candidate gene approaches towards understanding the factors that underlie the instability of the HTT CAG repeat . Unbiased linkage mapping in congenic HdhQ111 mice indicated Mlh1 as a potential genetic modifier of strain-specific HTT CAG instability . Subsequent candidate gene approaches demonstrated both Mlh1 and Mlh3 as critical novel modifiers of HTT CAG instability . The identification of Mlh1 and Mlh3 as modifiers of CAG instability in HdhQ111 mice suggests that variation in the human MLH1 and MLH3 genes may contribute to differences in somatic HTT CAG expansion that occurs between HD patients [9] , [11] . Further , given their minor roles in human tumorigenesis , both MLH3 and MSH3 currently stand as the most promising targets of the MMR proteins that have been identified as modifiers of the HTT CAG pathogenic process to date . Further delineation of the factors involved in somatic instability and the pathway ( s ) involved are likely to increase the ability to specifically intervene in the process of CAG/CTG expansion in HD as well as other trinucleotide repeat disorders .
Ethics statement: All animal procedures were carried out to minimize pain and discomfort , under approved IACUC protocols of the Massachusetts General Hospital and Cornell University . Congenic HdhQ111 strains on C57BL/6NCrl ( B6N ) , 129S2/SvPasCrlf ( 129 ) and FVB/NCrl ( FVB ) genetic backgrounds have been previously described [17] . In addition we generated HdhQ111 strains on DBA/2J ( DBA ) and C57BL/6J ( B6J ) backgrounds by repeated backcrossing of CD1 . HdhQ111/+ mice [15] for at least 10 generations . To map genetic modifiers of somatic HTT CAG instability we generated ( B6Nx129 ) . HdhQ111/+ and ( B6Nx129 ) . Hdh+/+ F1 mice which were subsequently intercrossed to generate ( B6Nx129 ) . HdhQ111/+ F2 progeny . B6 . Mlh1 knockout mice ( B6N ) [36] were crossed with B6N . HdhQ111 mice , and B6 . Mlh3 knockout mice ( B6J ) [39] were crossed with B6J . HdhQ111 mice to generate B6 . HdhQ111/+ mice heterozygous for the respective DNA repair mutation . These mice were then intercrossed to generate B6 . HdhQ111/+ littermates that were wild-type ( +/+ ) , heterozygous ( +/− ) or homozygous mutant ( −/− ) for the respective DNA repair gene . For reasons of simplicity , both B6N and B6J will be referred to as B6 unless otherwise specified . Mlh1 knockout mice were also generated on the 129 background by repeated backcrossing of B6 . Mlh1+/− mice for 4 generations . These mice were then intercrossed to generate 129 . Mlh1+/+ , 129 . Mlh1+/− and 129 . Mlh1−/− littermates . Animal husbandry was performed under controlled temperature and light/dark cycles . Genomic DNA was isolated from tail biopsies at weaning for routine genotyping analysis or from adult tissues ( fresh frozen or fixed as below ) for somatic instability analysis , using the PureGene DNA isolation kit ( Qiagen ) . Routine genotyping was carried out as previously described [19] , [36] , [39] . The size of the HTT CAG repeat was determined using a human-specific PCR assay that amplifies the HTT CAG repeat from the knock-in allele but does not amplify the mouse sequence [79] . The forward primer was fluorescently labeled with 6-FAM ( Applied Biosystems ) and products were resolved using the ABI 3730xl DNA analyzer ( Applied Biosystems ) with GeneScan 500 LIZ as internal size standard ( Applied Biosystems ) . GeneMapper v3 . 7 ( Applied Biosystems ) was used to generate CAG repeat size distribution traces . Repeat size was determined from the peak with the greatest intensity in the GeneMapper trace from the tail biopsy ( “main allele” ) . CAG repeat instability index was calculated as previously described [32] . Briefly , the highest peak in each trace was used to determine a relative threshold of 20% and peaks falling below this threshold were excluded from analysis . Peak heights normalized to the sum of all peak heights were multiplied by the change in CAG length of each peak relative to the main allele size in tail . These values were summed to generate an instability index , which represents the mean CAG repeat length change in the population of cells being analyzed . Statistical comparisons of instability indices were carried out using 2-tailed unpaired t-tests . Somatic CAG instability indices were determined in the striatum of 69 10-week-old ( B6x129 ) . HdhQ111/+ F2 mice , as described above . These F2 intercross mice were originally genotyped using a panel of 117 SNPs that distinguishes between C57BL/6J and 129S1/SvImJ strains ( Figure S3 and Table S1 ) [80] . An additional set of 30 SNPs was subsequently used to add resolution to the analysis ( Figure S3 and Table S1 ) , particularly at the chromosome 9 QTL , including two markers inside the Mlh1 gene ( dbSNP rs30131926 and rs30174694 ) ; as well as to specifically investigate the Msh2 ( dbSNP rs33609112 and rs49012398 ) and Msh3 ( dbSNP rs29551174 ) genes . Linkage analysis was performed using Mapmaker/QTL [81]–[84] , with striatal HTT CAG instability indices as quantitative traits . A threshold LOD-score of 4 . 3 was considered for the identification of significant QTLs [85] . A QTL 95% confidence interval was determined by using the 2-LOD-dropoff method [35] , [86] . Polymorphisms at the Mlh1 locus were investigated between C57BL/6NCrl ( B6N ) , 129S2/SvPasCrlf ( 129S2 ) , FVB/NCrl ( FVB ) and DBA/2J ( DBA ) genetic strains by standard DNA Sanger sequencing . PCR products were generated using Taq DNA polymerase ( Qiagen ) with DNA extracted from tail as template . A combination of primer pairs ( Table S2 ) was used to screen the complete coding sequence of Mlh1 as well as its immediate 5′ and 3′ flanking regions ( 2 . 6 kb and 2 kb respectively ) by sequencing both sense and antisense strands . Polymorphisms were validated in two animals from each genetic strain . We also utilized an online database for the Mouse Genomes Project ( http://www . sanger . ac . uk/resources/mouse/genomes ) , provided by the Wellcome Trust Sanger Institute . This database was derived from whole genome sequencing of 17 different genetic mouse strains [44] , [45] , including C57BL/6NJ ( B6NJ ) , 129S1/SvImJ ( 129S1 ) , DBA/2J ( DBA ) and FVB/NJ ( FVB ) . We used this database to retrieve information on SNPs , short indels and structural variants over a 64 kb region encompassing the Mlh1 gene ( chr9:111 , 223 , 496–111 , 287 , 496 ) , as well as at the Mlh3 ( chr12:85 , 234 , 529–85 , 270 , 591 ) , Msh3 ( chr13:92 , 201 , 881–92 , 365 , 003 ) , and Trex1/Atrip loci ( chr9:109 , 057 , 933–109 , 074 , 124; GRCm38/mm10 assembly ) . The average genome-wide variation between B6 and 129 was determined using the total number of SNPs and indels reported in this database ( B6J versus 129S1 ) relative to the GRCm38/mm10 genome size ( chromosomes 1–19 and X ) . The relative density of polymorphisms between B6 and 129 was determined by binning genome-wide SNPs and indels into 64 kb regions ( the same size as the Mlh1 genomic region analyzed ) and the mean density of polymorphisms/kb determined over each of the 64 kb bins . For reasons of simplicity , both B6N and B6NJ are referred to as B6 , 129S1 and 129S2 are referred to as 129 , and FVB/NCrl and FVB/NJ are referred to as FVB , unless otherwise specified . Immunostaining was carried out with polyclonal anti-huntingtin antibody EM48 [87] on 7 µm paraffin-embedded coronal sections of periodate-lysine-paraformaldehyde ( PLP ) -perfused mouse brains , as previously described [17] . Diffuse EM48 immunostaining was quantified as a “staining index” that captures both the nuclear staining intensity and the number of immunostained nuclei , as described previously [17] . Statistical comparisons of staining indices were carried out using 2-tailed unpaired t-tests . Total RNA was isolated from the striatum of wild-type B6 and 129 mice using Trizol ( Life Technologies ) by mechanical grinding with disposable pestle and cDNA was then prepared using the SuperScript III First-Strand Synthesis SuperMix for qRT-PCR ( Invitrogen ) . Full-length Mlh1 cDNAs were amplified by PCR ( for primers used see Table S2 ) using Phusion High-Fidelity DNA polymerase ( New England Biolabs ) , and were subsequently cloned between the unique NcoI and XhoI sites of a modified pFastBac1 baculovirus expression vector [88] , so that the resulting recombinant MLH1 proteins would carry N-terminal FLAG and 6xHis epitope tags . Mlh1 cDNA pFastBac1 constructs were fully verified by DNA sequence analysis confirming the presence of all B6-129 SNPs ( for primers used see Table S2 ) . The wild-type human MLH1 cDNA ( hMLH1-WT ) baculovirus expression vector [49] was used to generate a mutant hMLH1 cDNA construct carrying the 129-like Ile residue at aa192 ( hMLH1-F192I ) by site directed mutagenesis . Mouse and human recombinant MLH1 proteins were independently co-expressed with human PMS2 and purified using a baculovirus expression system to near homogeneity ( Figure S11 ) , as previously described [49] . Protein concentrations were determined spectrophotometrically and confirmed by polyacrylamide gel electrophoresis ( PAGE ) . Repair of a single base mismatch by MLH1 was investigated as previously described [49] . In essence , repair of single base mismatch ( G-T ) substrate containing a 5′ nick was assessed using HeLa or MutLα-deficient HCT116 [89] nuclear protein extracts ( 100 ng ) complemented with equal amounts of purified MutLα protein complexes: hMLH1 . WT-hPMS2 , hMLH1 . F192I-hPMS2 , mMLH1 . B6-hPMS2 or mMLH1 . 129-hPMS2 ( 100 ng ) . Note that as mMLH1-hPMS2 was functional in this well-established human-based assay , consistent with previous mixed yeast-human MMR assays [90]–[92] , we compared B6 and 129 MLH1 proteins in a mixed mouse-human MutLα complex , avoiding the need to introduce mouse PMS2 as another assay variable . Single base mismatch repair was analyzed by agarose gel electrophoresis followed by ethidium bromide staining [49] . Repair of a single trinucleotide repeat slip-out by MLH1 was investigated as previously described [50] . In summary , repair of single CTG slip-out substrates ( CAG ) 47• ( CTG ) 48 containing a 5′ nick was assessed using HeLa or MutLα-deficient HEK293T [42] , [93] whole cell extracts ( 120–180 ng ) complemented with equal amounts of purified hMLH1 . WT-hPMS2 , mMLH1 . B6-hPMS2 or mMLH1 . 129-hPMS2 complexes ( 100 ng ) , or with increasing amounts of mMLH1 . B6-hPMS2 or mMLH1 . 129-hPMS2 complexes ( 5 , 25 and 100 ng ) . This experiment with increasing concentrations was reproduced three times . Repair of CTG slip-outs was analyzed by Southern blotting . For both MMR assays , intensity of fragments was determined by densitometry and repair activity was determined as the intensity of repair fragments in proportion to the total intensity of all fragments [49] , [50] . Statistical comparison between mMLH1 . B6-hPMS2 and mMLH1 . 129-hPMS2 repair efficiency was carried out using 2-tailed unpaired t-tests . MutLα dose-dependency of CTG slip-outs repair was determined by linear regression . The HEK293T cell line was a gift from Dr . G . Plotz . HeLa cells were from the National Cell Culture Center , National Center for Research Resources , National Institutes of Health . mRNA and protein expression was investigated in frozen striatum samples from 10-week-old mice ( B6 . Mlh1+/+ , n = 3; 129 . Mlh1+/+ , n = 3; B6 . Mlh1+/− , n = 3; and B6 . Mlh1−/− , n = 1 ) , with the striatum from one hemisphere being used for mRNA analysis by qRT-PCR and the other being used for protein analysis by western blotting . Total RNA extraction and first-strand cDNA synthesis were performed as described above . Relative qRT-PCR was performed on a LightCycler 480 Real-Time PCR System ( Roche ) using TaqMan Gene Expression Master Mix ( Applied Biosystems ) and TaqMan Gene Expression Assays ( Applied Biosystems ) for: Mlh1 ( exons 4–5 , Mm01248478_m1; exons 11–12 , Mm00503449_m1; exons 18–19 , Mm00503455_m1 ) , Trex1 ( Mm00810120_s1 ) , and Atrip ( Mm00555350_m1 ) . Relative mRNA expression levels were determined using the 2−ΔΔCp method [94] by normalization to the housekeeping gene Actb ( Mm00607939_s1 ) . Each sample was run in triplicates and a total of 2 runs were performed . Protein lysates were prepared in RIPA buffer supplemented with 5 mM EDTA and protease inhibitors ( Halt Protease Inhibitor Cocktail , Thermo Scientific ) by mechanical grinding with disposable pestle and two 10-second sonication pulses ( Branson sonifier , power level 3 . 5 ) , on ice . The homogenates were kept on ice for 30 min and then clarified by centrifugation at 4°C for 30 minutes at 14000 rpm . Protein concentration was determined using the DC protein assay kit ( Bio-Rad ) . Western blot analysis was carried out by resolving protein extracts ( 50 µg ) on 4–12% Bis-Tris polyacrylamide gels ( NuPAGE , Life Technologies ) . All samples were run in the same gel and a total of 2 gels were run . Rabbit polyclonal antibody against the C-terminal end of MLH1 ( 1∶200; sc-582 , Santa Cruz Biotechnology ) and mouse monoclonal antibody against α-tubulin ( 1∶1 , 000; #3873 , Cell Signaling Technologies ) were used as primary antibodies and horseradish peroxidase-conjugated goat anti-rabbit and anti-mouse ( 1∶10 , 000; NA934VS and NA931VS respectively , Amersham ) were used as secondary antibodies . Signals were visualized using enhanced chemiluminescence ( ECL ) detection system ( Thermo Scientific ) . Densitometric analysis of protein levels was performed using UN-SCAN-IT software ( Silk Scientific Corp . ) . Following background subtraction , MLH1 protein levels were normalized to α-tubulin , and determined relative to B6 . Mlh1+/+ levels . Statistical comparisons of mRNA and protein levels were carried out using 2-tailed unpaired t-tests . The immediate 5′- and 3′-flanking regions of Mlh1 were amplified by PCR from both B6 and 129 genomic DNA ( for primers used see Table S2 ) using Phusion High-Fidelity DNA polymerase ( New England Biolabs ) . The immediate 5′-flanking region of Mlh1 ( 2 , 441 bp ) was cloned upstream of the firefly luciferase reporter in pGL4 . 20 ( Promega ) between the unique KpnI and NheI sites . Progressively smaller segments of the immediate 3′-flanking region of Mlh1 ( 1 , 676 bp , 1 , 280 bp , 591 bp and 205 bp ) were cloned downstream of the firefly luciferase reporter in pGL3-Promoter ( Promega ) between the unique XbaI and BamHI sites . Additional “swap” constructs were also generated for the immediate 3′-flanking region of Mlh1 ( 1 , 676 bp ) by dividing this region into 3 distinct subdomains ( 5′-3′: 530 bp , 438 bp and 708 bp; using PacI and KpnI ) and replacing individual subdomains from the B6 3′-flanking region of Mlh1 with the corresponding 129 subdomain . “Swap” constructs were cloned downstream of the firefly luciferase reporter in pGL3-Promoter ( Promega ) at the unique XbaI site . Mlh1–luciferase reporter constructs were fully verified by DNA sequence analysis , confirming the presence of all B6-129 SNPs ( for primers used see Table S2 ) . Individual Mlh1–firefly luciferase reporter constructs were co-transfected ( Lipofectamine LTX , Invitrogen ) with the Renilla luciferase reporter control pGL4 . 74 ( Promega ) into wild-type mouse immortalized striatal cells [95] . The transfected cells were cultured for 36–48 hours and luciferase expression was subsequently quantified using the Dual-Luciferase Reporter Assay System ( Promega ) on a microplate luminometer ( MicroLumat Plus LB96V , Berthold Technologies ) . Analogous B6 and 129 Mlh1–luciferase constructs were investigated in the same experiment in triplicate . The relative luciferase activity was calculated by normalizing firefly luminescence to the internal Renilla signal and determined relative to the corresponding B6 construct . Statistical comparison of relative luciferase activity between analogous B6 and 129 Mlh1–luciferase constructs was carried out using 2-tailed unpaired t-tests . | The expansion of a CAG repeat underlies Huntington's disease ( HD ) , with longer CAG tracts giving rise to earlier onset and more severe disease . In individuals harboring a CAG expansion the repeat undergoes further somatic expansion over time , particularly in brain cells most susceptible to disease pathogenesis . Preventing this repeat lengthening may delay disease onset and/or slow progression . We are using mouse models of HD to identify the factors that modify the somatic expansion of the HD CAG repeat , as these may provide novel targets for therapeutic intervention . To identify genetic modifiers of somatic expansion in HD mouse models we have used both an unbiased genetic mapping approach in inbred mouse strains that exhibit different levels of somatic expansion , as well as targeted gene knockout approaches . Our results demonstrate that: 1 ) Mlh1 and Mlh3 genes , encoding components of the DNA mismatch repair pathway , are critical for somatic CAG expansion; 2 ) in the absence of somatic expansion the pathogenic process in the mouse is slowed; 3 ) MLH1 protein levels are likely to be a driver of the efficiency of somatic expansion . Together , our data provide new insight into the factors underlying the process of somatic expansion of the HD CAG repeat . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2013 | Mismatch Repair Genes Mlh1 and Mlh3 Modify CAG Instability in Huntington's Disease Mice: Genome-Wide and Candidate Approaches |
The autonomous parvovirus Minute Virus of Mice ( MVM ) induces specific changes in the cytoskeleton filaments of infected permissive cells , causing in particular the degradation of actin fibers and the generation of “actin patches . ” This is attributed to a virus-induced imbalance between the polymerization factor N-WASP ( Wiscott-Aldrich syndrome protein ) and gelsolin , a multifunctional protein cleaving actin filaments . Here , the focus is on the involvement of gelsolin in parvovirus propagation and virus-induced actin processing . Gelsolin activity was knocked-down , and consequences thereof were determined for virus replication and egress and for actin network integrity . Though not required for virus replication or progeny particle assembly , gelsolin was found to control MVM ( and related H1-PV ) transport from the nucleus to the cell periphery and release into the culture medium . Gelsolin-dependent actin degradation and progeny virus release were both controlled by ( NS1 ) /CKIIα , a recently identified complex between a cellular protein kinase and a MVM non-structural protein . Furthermore , the export of newly synthesized virions through the cytoplasm appeared to be mediated by ( virus-modified ) lysomal/late endosomal vesicles . By showing that MVM release , like entry , is guided by the cytoskeleton and mediated by vesicles , these results challenge the current view that egress of non-enveloped lytic viruses is a passive process .
The genus parvovirus ( PV ) consists of small icosahedral non-enveloped particles with a 5 . 1-kb linear single-stranded DNA genome . During productive infection , PVs induce dramatic morphological and physiological changes in their host cells , culminating in cell death and lysis . PV cytotoxicity is attributed mainly to the large non-structural viral protein NS1 , an 83-kDa multifunctional protein endowed with enzymatic and non-enzymatic properties enabling it to control various processes necessary for progeny particle production and spread ( reviewed in [1] ) . To function in a concerted way , NS1 is regulated by specific phosphorylations driven mainly by members of the PKC family [2] , [3] . In addition to its direct involvement in particle production , NS1 acts specifically to jeopardize the integrity and survival of infected cells [4] , [5] , [6] . It has been shown to control the activity and properties of selected cell components through physical interaction [7] , [8] and/or induction of post-translational modifications [9] , [10] . Such targets might be modified either directly by NS1/CKIIα , a recently described complex formed by NS1 with the catalytic domain of cellular CKII [8] , or indirectly through activation/modulation of the PDK-1/PKC signaling cascade [11] . PV infection leads to characteristic alterations of host-cell morphology that might facilitate virus replication or the release of progeny particles . Subnuclear APAR-bodies acting as replication centers for parvoviral DNA amplification are formed early in infection [12] , [13] . Later , PVs induce cytoskeletal changes evidenced by rounding-up and detachment from the culture dish prior to cytolysis [14] , [15] . In MVM-infected mouse A9 cells , these morphological alterations have been attributed to the activity of NS1 [4] and shown to result from changes in micro- and intermediate filaments [10] . While tropomyosin is a direct target of NS1/CKIIα , MVM-induced actin-filament alterations appear to result from an imbalance between the polymerizing factor N-WASP ( Wiscott-Aldrich syndrome protein ) and gelsolin [10] , a multifunctional protein known mainly for its actin-filament-severing and capping activities and its participation in processes requiring rapid actin remodeling [16] . Roles in apoptosis and lipid signaling are also reported [17] . By altering the availability of PIP2 , gelsolin activity might interfere with PIP2-dependent signaling cascades affecting phospholipase C [16] . Little is known about the impact of cytoskeletal rearrangements on virus replication and spread . Cytoplasmic collapse is thought to be part of a process leading to virus release upon cytolysis [7] , but there is also indirect evidence of PV release in the absence of cell disruption [18] . The aim of the present study was to assess the role of gelsolin activity and actin reorganization in PV replication and spread . We show that gelsolin-induced modulation of actin filaments is essential to virus egress and provides strong evidence that progeny virions move to the cell periphery through vesicular transport and start to be released into the medium before cell collapse at the end of infection .
MVM-induced cytopathic effects include actin-fiber degradation and subsequent formation of actin patches at late stages of infection . The proposed cause is a virus-induced imbalance between actin polymerization and severing [10] . To investigate the impact of gelsolin on actin modulation and parvovirus replication we used confocal laser scanning microscopy . A9 cells infected ( or not ) with MVM were examined for gelsolin's subcellular distribution and its association with actin structures . Gelsolin was found to colocalize with phalloidin-stained actin structures ( Figure 1A ) whatever the infection status and time . In non-infected cells it accumulated abundantly along the rigid actin network and in the actin-rich region beneath the plasma membrane . Upon infection , concomitantly with destruction of the actin network , it became redistributed to the plasma membrane and perinuclear regions , later becoming associated with the above-mentioned cytoplasmic patches . The identical distribution of gelsolin and disorganized actin in infected cells suggests a link between the former and the state of the latter . Cell fractionation experiments confirmed the above findings ( Figure 1B ) . In non-infected cells , actin was found predominantly in the scaffold-containing fractions ( iS , sS ) , but after infection it was found in all subcellular fractions , including the cytosol ( C ) and the membrane-associated fractions ( nM , pM ) . In agreement with its association with remodeled actin structures , MVM-induced gelsolin was similarly found in all actin-positive fractions . As additional proof of gelsolin/actin interaction during MVM-induced actin reorganization , we used affinity chromatography to study actin-gelsolin binding . This method was previously used to determine NS1 association with tropomyosin [8] and proved successful in detecting specific protein interactions with partially insoluble cytoskeleton components in the cellular context . A9 cell lines expressing GST-coupled actin under the control of the parvoviral P38 promoter were MVM or mock infected . Twenty-four hours post-infection , cell extracts were prepared , matched for the GST-actin content due to viral induction of recombinant protein expression , and passed through Glutathione Sepharose columns to trap the fusion protein and associated polypeptides . After extensive washing steps , proteins bound to the trapped GST-actin were then eluted with high salt and cell matched volumes of input and eluates were then tested by Western blotting for the presence of gelsolin . As shown in Figure 1C ( actin lanes ) , gelsolin was invariably recovered from the GST-actin-loaded columns , whether the actin-bound proteins were from infected or non-infected cells . The specificity of the actin-gelsolin interaction was demonstrated by failure to detect gelsolin in eluates from columns loaded with extracts of A9 cells expressing only GST-free actin ( A9 lanes ) or GST-coupled tropomyosin ( TM5 lanes ) . Altogether , these results strongly suggest that gelsolin can interact with both filamentous actin and virus-processed actin structures such as the cytoplasmic patches . Since gelsolin is induced by MVM and associates with actin , we hypothesized that it might play a role in MVM-induced alteration of the actin network and in MVM propagation . To test this hypothesis , we transfected A9 cells with control serum/IgG or with antibodies ( αGln ) known to specifically inhibit gelsolin activity [19] , infected them with MVM , and placed them in culture . At different times , cells and their medium were collected separately ( Infection 1 ) . The collected cells were tested by immunofluorescence ( IF ) staining for expression of the viral protein NS1 , taken as indicator of successful infection ( Figure 2A ) . Southern blots were also produced from the cells , showing the different forms of DNA typically encountered in infected cells: the double-stranded replicative forms ( RF ) and the single-stranded DNA ( ssDNA ) of progeny virions ( Figure 2B , Inf1 ) . Neither the efficiency of infection nor the levels of the different viral DNA forms appeared to be altered by the presence of gelsolin-neutralizing antibodies . In a second phase of the experiment , the release and infectivity of progeny virions was determined by inoculating cultures of naive A9 cells with supernatant medium from infection-1 cultures ( Infection 2 ) . In this case , gelsolin activity appeared essential to either the formation or the release of infectious progeny virions , since cells exposed to infection-1 culture supernatants showed markedly lower levels of MVM DNA when the infection-1 cells had been treated with αGln ( Figure 2B , Inf2 ) . This result was confirmed by measuring production and release of infectious virions at 24 h p . i . by standard plaque assay . While cell-associated titers between αGFP-IgG and αGln-treated samples varied only marginally 1 . 5 fold ( GFP: 2 . 96×108 , αGln: 2 . 12×108 ) , inactivation of endogenous gelsolin blocked release of infectious virions leading to a 30 fold reduction of medium-associated titers in αGln-treated samples ( 2×107 vs . 6×105 ) . Similar results were obtained with human glioblastoma cells ( NCH149 ) infected with H1-PV , indicating that dependence on gelsolin is a general , late feature of PV infection ( Figure 2C ) . Is gelsolin required for the generation or for the release of infectious progeny virions ? Does its effect on progeny virion formation or release correlate with an active involvement of gelsolin in MVM-induced remodeling of the actin network ? To address these questions , we generated two cell lines , each stably transfected with a plasmid , pP38-MycGlnY438A or pP38-MycGlnD565N , driving MVM-inducible expression of a dominant-negative mutant gelsolin gene , so as to block endogenous gelsolin activity . The studied mutations were respectively: ( i ) a tyrosine-to-alanine substitution at position 438 , disrupting a phosphorylation site for Src kinases [20] regulating the PIP2 interaction and actin-severing activity [21] and ( ii ) a glutamic acid-to-glutamine substitution at position 565 , disrupting a conserved Ca2+-binding site regulating the activity of gelsolin through conformational alterations [22] . In IF microscopy and fractionation experiments , both mutant gelsolins ( GlnY438A and GlnD565N ) were found to accumulate in the perinuclear region and being associated with large , insoluble scaffold structures ( Figure S1 ) , i . e . , were able to interfere with the actin-processing activity of endogenous gelsolin . Furthermore , both mutant gelsolins were found to protect actin fibers from PV-induced remodeling , notably preventing the formation of patches ( Figure 3A , upper panel ) , and to impair the degradation of rigid actin filaments ( Figure 3A , lower panel ) . These results both confirm the dominant-negative character of the mutations introduced and demonstrate that gelsolin is instrumental in altering the actin network after PV infection . The same stable transfectants were then used to assess the role of gelsolin in infectious virion production and release . As in the case of cells treated with gelsolin-neutralizing antibodies ( Figure 2B and 2C ) , transfectants expressing GlnY438A or GlnD565N retained the ability to amplify MVM DNA . In contrast , the release of infectious viruses into the medium was drastically impaired ( Figure S2 ) . This lack of “free” viruses in the corresponding culture supernatants was not attributable to efficient readsorption onto neighboring cells , since it was also observed in cultures treated with neuraminidase to prevent MVM recapture . The question was thus: does gelsolin inactivation affect the formation or the shedding of infectious progeny viruses ? To address this question , we determined the infectious titers of both cell-associated and released virions ( Figure 3B ) . Cultures of transfectants expressing a mutant gelsolin displayed similar cell-associated infectious virion titers as A9 cultures , but a drastically ( 20- to 40-fold ) reduced titer of infectious virions shed into the medium . This strongly suggests that gelsolin is essential for efficient egress of progeny virions during MVM infection . We next investigated the subcellular location at which progeny particles get stuck in the absence of functional gelsolin . To this end , A9 cells and cells expressing GlnY438A or GlnD565N were infected with purified MVM , fixed at the indicated time p . i . , and examined by confocal microscopy for the presence of capsids . As shown for representative cells in Figure 4A and as quantified in Figure 4B , newly synthesized capsids were rapidly exported from the nucleus and transported to the periphery of A9 cells , so that capsid staining was distributed from the nucleus through the cytoplasm to the plasma membrane ( Nuc+Cytoplasmic capsids ) . In contrast , cells expressing either of the mutant gelsolins displayed trapping of a considerable proportion of the progeny virions in or around the nucleus at least until 48 h p . i . ( [Peri]nuclear capsids only ) . Capsid staining of infected A9 cells was noticeably spotty ( Figure 4A ) , suggesting that progeny viruses might be transported by vesicular structures . This possibility was tested by confocal microscopy of infected cells after double IF labeling of assembled capsids and either vesicular markers or proteins known to be involved in vesicle formation . We took several measures to make sure we were observing virus release and not virus entry: ( i ) we checked that no incoming capsids were detected under the conditions used ( Figure 4A ) ; ( ii ) we prevented re-infection by neuraminidase treatment of the cells after infection; ( iii ) we checked that the results were similar when the cells were transfected with viral DNA rather than exposed to virus particles . In the parental A9 cells , newly synthesized capsids were found , 24 and 48 h p . i . , to colocalize with Lamp2 ( Figure 5A ) , cathepsin B , and Rab6 , but not with the mitochondria ( Figure S3 ) . Cells expressing GlnY438A or GlnD565N showed no colocalization with Lamp2 , cathepsin B , or Rab6 . These data strongly support a role of ( virus-modified ) lysosomes or late endosomes in gelsolin-dependent export of progeny particles . In agreement with its involvement in ( endosomal ) vesicle formation [23] , MVM infection caused dynamin to accumulate in the perinuclear region , where it was found to colocalize with newly synthesized capsids ( Figure 5B ) . To further substantiate the association of newly synthesized infectious virions with cellular vesicles , extracts prepared from infected A9 cells or A9 derivatives expressing GlnY438A were treated to separate nuclei , large organelles ( HMF ) , a light mitochondrial fraction ( LMF ) , and a soluble cytosolic fraction , the LMF being further fractionated in a self-forming iodixanol gradient . In A9 extracts , as shown in Figure 5C , ssDNA was found not only in the nuclei along with RF DNA , but also in the HMF and LMF , co-migrating with Lamp2 , a profile suggestive of a vesicular localization . Very little virion DNA was found in the cytosolic fraction . Interestingly , only minute amounts were detected in the LMF fractions derived from cells expressing a mutant gelsolin . This suggests a possible involvement of this actin-processing protein in the formation of capsid-containing vesicles or their release from larger compartments . To further examine the role of actin in this process we determined whether infectious virions can physically bind to ( virus-modified ) actin . Protein complexes formed with GST-β-actin or GST-α-tubulin were extracted from MVM- or mock-infected cells and trapped on Glutathione Sepharose columns . The partners of β-actin or α-tubulin were then recovered and the eluates tested for the presence of virion DNA ( by Southern blotting ) and capsid proteins ( by Western blotting ) . Parental A9 cells served as negative controls . In contrast to free replicative form DNA which appeared to interact nonspecifically with the column material , MVM progeny virions were found to bind specifically to GST-actin and to elute from the column at high salt concentration , as evidenced by the presence of VP2 and ssDNA in the eluates from MVM-infected cells expressing GST-actin ( Figure 5D ) . In agreement , with our findings of virus-induced actin association with ( cellular ) membranes ( Figure 1B ) and the requirement for gelsolin to egress virions from the nucleus , this supports a hypothesis that rapid actin remodeling might be required for the formation and/or motility of virion-containing vesicles . Previous investigations with an MVM-inducible cell line expressing a dominant-negative mutant of CKII ( A9-P38:CKII-E81A ) have shown that functional CKII is essential to the release of progeny virions into the culture medium [7] . As illustrated in Figure 6A and quantified in Figure 6B , these cells are distinguishable from the parental line A9 by a striking retention of progeny viruses in the nucleus and perinuclear region . This defect , similar to that observed after functional inactivation of gelsolin , suggests that CKII might take part in regulating gelsolin in MVM-infected cells . This hypothesis was first tested by determining whether expression of the dominant-negative form of CKIIα can interfere with gelsolin-dependent remodeling of actin filaments in infected cells . A9 and A9-P38:CKII-E81A cells were infected with purified MVM and examined by confocal microscopy . As shown in Figure 6C , inhibition of CKIIα was found to correlate with prolonged persistence of rigid actin filaments and delayed formation of actin patches . Furthermore , gelsolin/actin colocalization was strongly reduced upon MVM infection . All of these observations are in agreement with the involvement of this kinase in controlling gelsolin-driven cytoskeletal changes . We then investigated whether gelsolin might be a target of phosphorylation by cellular protein kinases . Bacterially expressed purified gelsolin was incubated with recombinant PKC ( isoform α , δ , η , or λ ) or CKIIαβ in the presence of [32P]-γ-ATP . CKIIαβ was tested either alone or with the GST-NS1 polypeptide , given our recent finding that NS1 can act as an adaptor and modulate the substrate specificity of this kinase [8] . [32P]-labeled proteins were then analyzed by SDS-PAGE and autoradiography . As shown in Figure 7A , gelsolin proved to be a poor substrate for most of the tested protein kinases , including CKIIαβ alone and the NS1-modifying PKCη and PKCλ . In contrast , it was readily phosphorylated by the NS1/CKIIα complex . Two-dimensional phosphopeptide analyses confirmed that NS1 endows CKII with the capacity to phosphorylate gelsolin at multiple sites ( Figure 7B ) . This NS1 dependence is specific , since the viral product failed to modulate the CKII-driven modification of tubulin . These results raise the intriguing possibility that MVM-induced modification of gelsolin by the NS1/CKIIα complex results in actin network alterations that facilitate virus egress .
Previous studies have shown that cytoplasmic collapse at the end of parvovirus MVM infection does not reflect mere cell exhaustion , but results from a readily observable remodeling of cytoskeleton filaments [10] . In contrast to tropomyosin , which is degraded upon NS1/CKIIα phosphorylation [8] , and to microtubules , which are actively protected through PKCλ-mediated phosphorylation [10] , actin appears to undergo remodeling as a result of a virus-induced imbalance between the activator protein for polymerization , N-WASP , and the severing factor gelsolin . We show here that gelsolin activity is required not only to trim actin filaments but also to drive virus export from the nucleus ( or its immediate periphery ) to the outside of infected cells . This transport of progeny virions to the cell periphery is mediated by vesicles bearing protein markers of lysosomes and/or late endosomes , suggesting that gelsolin may play a role in the formation , loading , and/or trafficking of these vesicles . Little is known about the egress of non-enveloped lytic viruses from infected cells , commonly thought to occur as a virus burst after cell disintegration . Although cell lysis can be expected to contribute considerably to virus spread in tissue cultures , its importance in animal infection is unclear , as there is some evidence of non-lytic egress of non-enveloped viruses [24] , including PVs [18] . Here we present evidence suggesting that PV egress from infected cells is controlled by a cytoskeleton-regulating protein , gelsolin , and mediated by vesicles , arguing strongly for an active egress mechanism . This process bears some resemblance to parvovirus entry , involving clathrin-dependent endocytosis and endosomal transport via microtubules [25] . For virus egress , gelsolin is essential at an early step taking place at or near the nuclear envelope . Although the transfer of incoming PVs to the nucleus is partly understood [26] , [27] , the gelsolin-dependent event enabling progeny virions to leave the nuclear area is currently a matter of speculation . We show here that in MVM-infected cells , gelsolin accumulates not only along rigid actin fibers but also at the plasma membrane and , at a later stage of infection , within actin patches . These findings are in agreement with the actin-processing function of gelsolin , but provide no obvious clue to its involvement in the peripheral transport of PV particles . Interestingly , rapid actin remodeling , a known gelsolin activity , is reported to be associated with the formation of vesicles [23] , [28] . Our present finding that progeny virions are detectable in vesicular structures only if gelsolin is functional leads us to suggest that gelsolin may drive the assembly , loading , or mobilization of vesicles involved in transferring viral particles from the perinuclear region to the cell periphery . There are multiple reports demonstrating actin-dependent transport of intracellular pathogens ( including viruses ) and cellular vesicles [24] , [28] . While entry and egress of vaccinia virus involves movement along the microtubules , propulsion of this virus during dissemination from infected cells to adjacent tissue requires rapid actin polymerization induced by a WASP-like viral protein constitutively activating the Arp2/3 complex and leading to the appearance of “actin-tails” [29] . The movement of vesicles , on the other hand , has been shown to be driven by myosins along intact actin filaments [30] . Our data do not support such a role of the actin scaffold in guiding ( vesicle-contained ) MVM towards the plasma membrane , since ( i ) there is no homology between PV-encoded proteins and ENA/WASP-family proteins , ( ii ) recruitment of endogenous N-WASP is unlikely , as this protein is strongly down-regulated at late stages of infection [10] , and ( iii ) infection induces actin filament degradation , known to inhibit this transport system [30] . Interestingly , there is recent evidence of cross-talk between actin- and microtubule-dependent transport [23] , [31] , [32] . This raises the intriguing possibility that gelsolin-dependent actin processing might trigger the picking-up of virions or virus-loaded vesicles by the microtubule network for their transport from the nucleus to the periphery . This would be in agreement with the maintenance of microtubules until late in infection [10] and with the capsid-dynamin colocalization reported here . Besides inducing the accumulation of gelsolin , MVM infection alters its subcellular distribution and membrane-binding affinity . This suggests that gelsolin may be subject to virus-induced post-translational modifications . This possibility is in keeping with the observation of multiple gelsolin species after SDS-PAGE . Although we were unable to immunoprecipitate enough endogenous gelsolin to allow characterization of its in vivo phosphorylation pattern , we present strong in vitro evidence of its regulation through phosphorylation by the NS1/CKIIα complex: NS1 can retarget CKII to gelsolin , leading the kinase to phosphorylate this protein at multiple sites . These modifications may be relevant to the role of gelsolin in virus egress , since CKII inhibition and gelsolin inactivation similarly impair the outward transfer of progeny virions from the nucleus . Investigations in progress aim to pinpoint gelsolin functions involved in MVM infection , and particularly virus egress . As stated above , there may be a direct connection between gelsolin-dependent actin processing and virus transport systems . On the other hand , we show here that gelsolin also localizes to the cytoplasmic actin patches appearing late in infection and which are not associated with capsids . Although the role of these patches remains elusive , one might speculate that they fulfill a signaling function . Actin structures can indeed serve as scaffolds for signaling cascades , and because of its high affinity for PIP2 , gelsolin is thought to affect cell pathways , notably ones involving PLC and PLD . Altogether , these observations point to lipid-dependent signaling as a potential alternative gelsolin target to be studied for its impact on the outcome of parvovirus infection .
A9 mouse fibroblasts , derivatives thereof , and NCH149 human glioma cells [15] were maintained as monolayers in Dulbecco's Modified Eagle Medium ( DMEM ) containing 10% fetal calf serum ( FCS ) . MVMp ( MVM ) and H1-PV were propagated respectively in adherent A9 and NCH149 cells . Virus stocks were prepared by freezing and thawing in TE pH 8 . 3 . When indicated , full ( DNA-containing ) MVM particles were separated from empty capsids , on the basis of their buoyant density , by CsCl-gradient centrifugation . Stable transfectants were generated with pP38-MycGlnY438A , pP38-MycGlnD565N or pP38:GST-Tubα and the selection plasmid pSV2neo at the molar ratio of 25∶1 [3] . Colonies were pooled after growth under selection and frozen stocks prepared . Experiments were performed in absence of G418 . Transfectants were kept in culture for less than 25 passages . Previously established cell lines contained pP38-GST-βactin , pP38-GST-TM5 [8] , pP38-CKII:E81A [7] , or pP38-PKCηT512A [3] . Accumulation of MVM DNA was determined by Southern blotting [34] . When indicated , transfection with gelsolin-neutralizing antibodies was performed 4 h prior to virus infection , with 7 µg IgG and 15 µl Provectin ( Imgenex ) . To prevent secondary rounds of infection , the cells were treated with 100 ng/ml neuraminidase ( Sigma ) 4 h p . i . They were harvested in TE buffer , digested with proteinase K , and total DNA was sheared by passage through a syringe . Viral DNA was analyzed by agarose gel electrophoresis and detected , after blotting onto nitrocellulose membranes , with a 32P-labeled probe corresponding to nts 385–1885 of the NS1-encoding region of MVM DNA . To measure the formation and release of progeny virions , cultures were infected with MVM as described above . At the indicated times p . i . , medium was removed and kept separately . Adherent cells were washed , harvested in DMEM without serum by scraping from the dish , and collected by centrifugation . Medium- and cell-associated virions were quantified , after repeated freezing and thawing , in standard plaque assays [5] . Protein extracts were fractionated by discontinuous SDS-PAGE and blotted onto nitrocellulose membranes . Proteins of interest were detected by incubation for 18 h with appropriate primary antibodies in 10% dry milk/PBS and staining with HRP-conjugated secondary antibodies for 1 h followed by chemiluminescence detection ( Amersham ) . Cells were grown on spot slides ( Roth ) , mock- or MVM-infected , and further incubated for the appropriate time . Cultures were fixed with 3% paraformaldehyde and permeabilized with 0 . 1% Triton X-100 . Specimens were preadsorbed with 20% FCS , incubated with primary antibodies , and stained with specific Alexa Fluo 594- , CY2- , CY3- , or rhodamine-conjugated anti-species antibodies . After mounting with Elvanol , cells were analyzed by laser scanning microscopy with a Leica DMIRBE apparatus ( 63× lens , laser: red 543 nm , green 488 nm ) and Powerscan software or by spinning disk confocal microscopy with a Perkin Elmer ERS 6Line microscope ( 100× lens , laser: red 568 nm , green 488 ) presenting a single slice of a stack . Quantitative analyses were performed on all slides of a stack , mean colocalization being calculated with ImageJ software . Proteins and viral components interacting with cytoskeletal proteins were identified by affinity chromatography [8] . A9 cells or derivatives thereof ( A9-P38:GST-βactin , A9-P38:GST-TM5 , A9-P38:GST-Tubα ) were infected ( or not ) with MVM ( 30 pfu/cell ) . Extracts were prepared as nuclear squeezes into the cytoplasm , loaded onto Glutathione Sepharose columns under isotonic conditions , and washed extensively . Components binding to the glutathione-S-transferase-tagged ( GST-tagged ) baits were then specifically eluted with 700 mM NaCl . Viral and cellular proteins were detected by western blotting . The single-stranded DNA genomes of infectious virions were detected by Southern blotting after extensive treatment with proteinase K . Protein extracts or fractions thereof were diluted in 700 µl Co-Ip-buffer ( 20 mM Hepes-KOH pH 7 . 5 , 300 mM NaCl , 1 mM EDTA , 0 . 2 % NP-40 ) and pre-cleared by addition of FCS ( 5 µl ) and protein G-Sepharose ( 40 µl ) for 2 h at room temperature . After centrifugation , soluble proteins were incubated with specific antibodies/antiserum for 18 h at 4°C before addition of protein G-Sepharose for 2 h at room temperature . After extensive washes with Co-Ip buffer , immune complexes were collected and analyzed . In vitro kinase reactions and tryptic phosphopeptide analyses were performed as described [35] with recombinant CKIIαβ ( Roche ) and the various PKC isoforms ( Sigma ) . When indicated , purified GST-tagged wild-type or mutant ( S473A ) NS1 protein was added [8] . Gelsolin and tubulin used as substrates were produced in bacteria and purified as described for tropomyosin [8] . Assays were performed for 40 min at 37°C with 30 µCi γ-[32P]ATP in 50 µl of 20 mM HEPES-KOH [pH 7 . 5] , 7 mM MgCl2 , 150 mM NaCl , 1 mM DTT , in the presence of the appropriate cofactors . The reactions were stopped and the reaction products analyzed either directly or after immunoprecipitation , by 10% SDS-PAGE and semi-dry transfer onto polyvinyldifluoride ( PVDF ) membranes ( Millipore ) . Phospholabeled proteins were then digested with trypsin and analyzed by two-dimensional thin-layer electrophoresis/chromatography ( electrophoresis at pH 1 . 9/phosphochromatography ) . | Rodent parvoviruses are non-enveloped lytic viruses that are thought excellent tools for a virotherapy of cancer because of their strong natural oncolytic potential and low pathogenicity in humans . Egress of non-enveloped lytic viruses is commonly thought to occur as a virus burst after cell disintegration . Indeed , we showed in the past that autonomous parvoviruses induce severe cytopathic effects to the host cell , manifested in restructuring and degradation of cytoskeletal filaments , thereby supporting such mode of virus spread . Here , we focus on the impact of virus-induced actin degradation , and particularly the functioning of the actin-severing protein gelsolin . Although not required for DNA replication or progeny particle production , gelsolin appears to facilitate a regulated virus egress from the nucleus to the cell periphery via ( virus modified ) lysosomal/late endosomal vesicles . These results challenge the current view that lytic virus egress is just a passive process at the end of infection and suggests that these pathogens are endowed with the ability to efficiently spread from cell to cell potentially in solid ( tumor ) tissue . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"cell",
"biology/cell",
"signaling",
"virology/virion",
"structure,",
"assembly,",
"and",
"egress",
"cell",
"biology/membranes",
"and",
"sorting",
"biochemistry/cell",
"signaling",
"and",
"trafficking",
"structures",
"virology/viruses",
"and",
"cancer",
"cell",
"biology/cytoskeleton"
]
| 2008 | Vesicular Egress of Non-Enveloped Lytic Parvoviruses Depends on Gelsolin Functioning |
Dogs are the predominant domestic reservoir for human L . infantum infection . Zoonotic visceral leishmaniasis ( ZVL ) is an emerging problem in some U . S . dog breeds , with an annual quantitative PCR prevalence of greater than 20% within an at-risk Foxhound population . Although classically Leishmania is transmitted by infected sand flies and phlebotomine sand flies exist in the United States , means of ongoing L . infantum transmission in U . S . dogs is currently unknown . Possibilities include vertical ( transplacental/transmammary ) and horizontal/venereal transmission . Several reports have indicated that endemic ZVL may be transmitted vertically . Our aims for this present study were to establish whether vertical/transplacental transmission was occurring in this population of Leishmania-infected US dogs and determine the effect that this means of transmission has on immune recognition of Leishmania . A pregnant L . infantum-infected dam donated to Iowa State University gave birth in-house to 12 pups . Eight pups humanely euthanized at the time of birth and four pups and the dam humanely euthanized three months post-partum were studied via L . infantum-kinetoplast specific quantitative PCR ( kqPCR ) , gross and histopathological assessment and CD4+ T cell proliferation assay . This novel report describes disseminated L . infantum parasites as identified by kqPCR in 8 day old pups born to a naturally-infected , seropositive U . S . dog with no travel history . This is the first report of vertical transmission of L . infantum in naturally-infected dogs in North America , emphasizing that this novel means of transmission could possibly sustain infection within populations . Evidence that vertical transmission of ZVL may be a driving force for ongoing disease in an otherwise non-endemic region has significant implications on current control strategies for ZVL , as at present parasite elimination efforts in endemic areas are largely focused on vector-borne transmission between canines and people . Determining frequency of vertical transmission and incorporating canine sterilization with vector control may have a more significant impact on ZVL transmission to people in endemic areas than current control efforts .
Zoonotic visceral leishmaniasis ( ZVL ) is a vector-borne disease caused by obligate intracellular protozoan parasites of the genus Leishmania . In South America , dogs are the primary domestic reservoir host for ZVL [1] . Control measures for this disease are focused on vector-control and euthanasia of seropositive dogs [2] , [3] . In 2000 , Leishmania infantum , the causative agent of ZVL , was determined to be the cause of death in four Foxhounds in New York [4] . Much like human disease , signs of canine ZVL include weight loss , depression , splenomegaly , heptaomegaly , generalized lymphadenomegaly and serosanguineous nasal discharge [5] . Currently , canine ZVL in the United States is a growing problem in the Foxhound population , with an annual quantitative PCR prevalence of greater than 20% within an at-risk Foxhound population [6] , [7] . Despite this and the obvious public health concerns , primary means of transmission has yet to be determined [5] . In historically-endemic regions , the sand fly is the primary vector for this disease , and although sand flies are present within the southern United States , it has not been determined if these species are competent vectors for L . infantum [5] , [8] . There have been multiple cases of autochthonous canine ZVL in the United States . Infected dogs had not visited endemic regions nor had direct contact with other infected animals [9] , [10] . Recently , two reports have demonstrated that L . infantum infection in endemic regions was transmitted vertically across the placenta [11] to unborn fetuses . da Silva et . al . described natural transplacental transmission of Leishmania to stillborn pups from a dog in South America [12] . Another study determined that 32% of fetuses from naturally infected dogs were PCR positive for Leishmania kinetoplast DNA . Although there were no gross lesions in the fetuses or placentas , low numbers of parasites were present via histology in the liver , spleen , lymph node and bone marrow [11] . In this report we describe disseminated L . infantum infection in multiple live day-old pups born to a naturally-infected seropositive female dog . Gross and histological findings in the dam were consistent with canine visceral leishmaniasis . While the pups were yet to acquire any gross or histologic lesions consistent with disease , L . infantum kinetoplast-specific qPCR analysis of multiple tissues from 10/12 pups indicated highly disseminated infection . Moreover , the dam and all pups tested had L . infantum-specific CD4+ T cell proliferative responses , suggesting an ongoing immune response specific to the parasite and not a naïve immune response . To our knowledge , this is the first determination of transplacental transmission of Leishmania infantum in naturally-infected dogs in North America .
A pregnant , seven-year old American Foxhound female was donated to Iowa State University , Department of Veterinary Pathology in March of 2009 following demonstration of seropositivity via Centers for Disease Control Indirect Immunofluorescent assay ( IIF ) a whole parasite-based serological method ( 1∶128 ) . Three weeks after arrival , the dam gave birth to 12 pups , of which 8 were euthanized within 24 hours , and 4 were euthanized 12 weeks after birth , along with the dam . All animal use involved in this work were according to International AAALAC accreditation standards and ISU institutional IACUC approval . Animals were donated to ISU for use after signed informed consent . ISU animal facilities and programs are annually inspected and found to be above all guidelines by NIH , USDA and recently AAALAC . At the time of necropsy a complete set of tissues from all animals were collected and fixed in 10% neutral buffered formalin . Tissues were routinely processed and stained with hematoxylin and eosin ( H&E ) for histopathologic evaluation . A North American canine isolate of Leishmania infantum , ( LIVT-2 ) [13] , was grown to stationary phase in complete Grace's medium ( Incomplete Grace's supplemented with 20% fetal bovine serum , 100 U/ml penicillin , 100 µg/ml streptomycin and 2 mM L-glutamine ) . Freeze-thawed whole antigen was prepared as described previously [14] . PBMC were isolated from heparinized whole blood samples using Ficoll-Histopaque 1077 ( Sigma , St . Louis , MO ) gradient centrifugation . Red blood cells were removed using ACK lysis buffer ( 0 . 15 M NH4Cl , 1 . 0 mM KHCO3 , 0 . 1 mM Na2EDTA , pH 7 . 4 ) . PBMC were labeled with CFSE ( Molecular Probes , Eugene , OR ) as described previously [15] . PBMC were washed twice in phosphate-buffered saline ( PBS ) and resuspended in complete medium ( RPMI 1640 supplemented with 10% fetal bovine serum , 100 U/ml penicillin , 100 µg/ml streptomycin , 2 mM L-glutamine , and 25 mM HEPES buffer ) . PMBC were counted and adjusted to 4×106/ml for further analysis . CFSE-labeled PBMC ( 4×105/well ) were plated into 96-well plates and incubated with media alone , stimulated with concanavalin A ( ConA ) ( 5 µg/ml ) for 4 days or with freeze-thawed , whole L . infantum antigen ( 10 µg/ml ) for 7 days at 37°C with 5% CO2 . Cells were harvested , washed in FACS buffer ( 0 . 1% albumin , 0 . 1% sodium azide in PBS ) and labeled with PE-conjugated anti-canine CD4 antibody ( Serotec , Raleigh , NC ) . Cells were fixed in 1% paraformaldehyde and analyzed using the FACSCanto flow cytometer ( BD Pharmingen , San Diego , CA ) . Data was analyzed using FlowJo software ( Tree Star Inc . , Ashland , OR ) . Serum samples were collected from all animals , stored at −20°C and sent to the Centers for Disease Control and Prevention for IIF testing for antibodies to Leishmania spp . as previously described [16] . DNA from whole blood samples collected in heparinized tubes ( BD Pharmingen , San Diego , CA ) was isolated using the Qiagen blood DNA isolation kit according to manufacturer's instructions . Samples of placenta , bone marrow , liver , lymph node , lung , spleen , thymus , and umbilicus were collected individually and stored at −20°C until processed for DNA extraction similar to whole blood . DNA quality and quantity was measured using a NanoDrop spectrophotometer ND1000 ( Wilmington , DE ) . L . infantum kinetoplast DNA ( kDNA ) -specific primers and probe F 5′-CCGCCCGCCTCAAGAC , R 5′-TGCTGAATATTGGTGGTTTTGG , ( Integrated DNA Technologies , Coralville , IA ) , Probe 5′-6FAM-AGCCGCGAGGACC-MGBNFQ ( Applied Biosystems , Foster City , CA ) were used . ( FAM: laser-activated reporter dye; MGBNFQ: 3′-minor-groove binder non-fluorescent quencher ) . BLAST analysis indicated that these primers and probe were specific for L . infantum . DNA from L . amazonensis or L . major parasites did not amplify using this primer and probe set . Blood DNA samples were assayed via qPCR in duplicate of three dilutions ( straight , 1∶10 , 1∶20 ) using a Stratagene Mx3005P qPCR System in a 96-well format and Platinum qPCR SuperMix-UDG Master Mix ( Invitrogen , Carlsbad , CA ) as previously described [7] . Results were analyzed via MxPro QPCR software version 4 . 01 in conjunction with Microsoft Excel . Ten serial 1∶5 dilutions of a carefully calibrated sample containing 109 whole parasites/ml were made . 50 µl of each respective parasite dilution was subsequently spiked into 150 µl of fresh canine whole blood that was collected in heparinized tubes ( BD Pharmingen , San Diego , CA ) as described previously [7] . Each parasite-spiked blood sample was subsequently extracted for DNA using the Qiagen blood DNA isolation kit as above . Assuming a parasite extraction efficiency of ∼90% , 6 µl ( the amount used for each qPCR reaction ) of each of the ten resulting full-strength , straight ( 0 . 75-strength blood ) DNA extracts were calculated to contain 1 , 350 , 000 , 270 , 000 , 54 , 000 , 10 , 800 , 2 , 160 , 432 , 86 . 4 , 17 . 28 , 3 . 46 and 0 . 69 total parasites/sample , respectively , while the 1∶2 . 4-diluted series ( to give a 1∶10 dilution in well ) was calculated to contain 562 , 500 , 112 , 500 , 22 , 500 , 4 , 500 , 900 , 180 , 36 , 7 . 2 , 1 . 44 and 0 . 29 total parasites , respectively . Each of the 20 samples were analyzed by qPCR in triplicate using a Stratagene Mx3005P qPCR System in a 96-well format and Platinum qPCR SuperMix-UDG Master Mix ( Invitrogen , Carlsbad , CA ) as described previously . 1∶10-diluted DNA samples yielded the most consistent qPCR results and longest dynamic range . Each L . infantum parasite on average contains from 1 to 7 . 4 copies of the kinetoplast DNA sequence targeted by our qPCR primer and probe set [17] . The limit of detection ( LOD ) of this assay was determined to be ∼7 . 2 kinetoplast copies , or 1 parasite , per qPCR reaction . This represents an LOD range of 400–3 , 000 parasites per ml of canine blood , dependent on kinetoplast copy number and parasite stage . Parasite load was calculated as follows: initial copies per reaction = EAMP ( b – Cq ) , Copieso = 2 . 0255 ( 46 . 764 – Cq ) , where EAMP is the exponential amplification , as determined by the slope of the standard curve ( EAMP = 10−1/m ) , b = y intercept or limit of detection cycle threshold and Cq is the cycle of detection of a particular sample . This equation was directly applied to the Cq derived from each diagnostic/clinical dog blood sample extracted for DNA and subjected to kqPCR .
Twelve puppies were born to a L . infantum-naturally infected female Foxhound . At the time of birth the dam was both serologically ( 1∶128 ) and kqPCR positive for L . infantum . On histopathologic evaluation of the dam , findings in the bone marrow , liver and spleen were consistent with lymphoplasmacytic and histiocytic inflammation , likely due to disseminated visceral leishmaniasis . Rare amastigotes were noted within the spleen and liver . Of the twelve puppies , eight ( 1–8 ) were euthanized within 24 hours of birth and 4 ( A–D ) were euthanized 12 weeks after birth . All animals were submitted for necropsy . Gross findings indicated none of the first eight pups had any gastric or intestinal contents beyond amniotic fluid and scant meconium suggesting the pups had not yet suckled , ruling out the possibility of transmammary transmission . No signs of clinical leishmaniasis were noted in the pups . Samples from various tissues were collected during necropsy and analyzed for the presence of L . infantum kinetoplast DNA via qPCR . Seven of the eight puppies euthanized 24 hours after birth were positive for L . infantum in at least one tissue tested . Most pups , 1–2 and 5–8 , showed systemic disseminated infection as the parasite was detected on multiple tissues ( Figure 1A ) . All tissues tested from the dam were positive for L . infantum DNA , including the placenta , indicative of disseminated visceral infection ( Figure 1B ) . kqPCR analysis of dogs euthanized at 12 weeks demonstrated systemic parasite dissemination only in 1 out of 4 the pups , pup D ( Figure 1B ) . Pups A and C tested positive for L . infantum only in bone marrow , and pup C was not kinetoplast-qPCR positive on any of the tissues analyzed ( Figure 1B ) . In addition , pups that tested positive for L . infantum via kqPCR also had a high number of parasite genomic copies ( Figure 1A and B ) indicative of high parasite loads in these tissues . These data suggest that not only were the puppies infected transplacentally , but also that the parasite is able to disseminate systemically in utero leading to high parasite loads in multiple tissues . Whole blood samples from the dam , pups 4–7 and A–D were collected prior to euthanasia . Peripheral blood mononuclear cells ( PBMC ) were isolated , stained with CFSE , and stimulated with concanavalin A ( ConA ) , L . infantum antigen , or were left untreated . PMBC were analyzed for CD4+ T cell proliferation via flow cytometry . CD4+ T cells from all dogs proliferated in response to stimulation with ConA , indicating that the CD4+ T cell compartment was not mitogenically deficient ( data not shown ) , as previously indicated [7] . In response to L . infantum antigen stimulation all dogs excluding pups A and 7 , had strong antigen-specific CD4+ T cell proliferative responses ( Figure 2 ) . These data suggest that despite likely in utero transmission , these pups were able to mount antigen-specific adaptive immune response at birth and were neither naïve nor immune-tolerant to L . infantum antigen .
Vertical transmission of L . infantum has been previously demonstrated in experimentally-infected beagles and mice , and in naturally-infected dogs in endemic areas [11] , [12] , [18] , [19] . In the last four years L . infantum kqPCR incidence in kennels with clinical visceral leishmaniasis in the Midwest has remained at 12% ( Petersen unpublished data ) despite de-population efforts of seropositive animals . This suggests that ongoing disease transmission is occurring within the Foxhound population . Although leishmaniasis is classically described as a vector-borne disease , alternative modes of transmission including horizontal ( via direct blood to blood or sexual contact ) [20] and vertical ( transplacental or transmammary ) transmission are likely to have a role during ZVL transmission [21] . Our study demonstrates that vertical transmission of L . infantum occurs with a high penetrance demonstrated in one litter within the Foxhound population in the United States . If there is not a competent vector species , vertical transmission may be a primary mode of canine L . infantum transmission in the U . S . We demonstrate the presence of L . infantum parasite kinetoplast DNA within various tissues including the placenta of an infected Foxhound dam . Given the high blood flow through the placenta during pregnancy , L . infantum amastigotes are likely to be carried to the placenta and on to the pups . With 10 out of 12 pups demonstrating L . infantum infection ( 83% ) via kqPCR , transplacental infection is likely a primary mechanism of transmission in this population . Transvaginal infection has been suggested to play a role during vertical transmission of L . infantum in canines [22] , however , the level of parasite dissemination ( multi-organ ) , adaptive CD4+ T cell immune response to parasite antigen at birth and the high number of parasite copies detected via qPCR ( Figure 1 ) at birth , suggests transplacental rather than transvaginal transmission , supporting previous findings [11] , [18] . Histological examination of bone marrow , liver , lymph nodes and spleen tissue samples did not reveal the presence of Leishmania amastigotes in these pups . However , the sensitivity for identification of amastigotes in hematoxylin and eosin-stained sections is extremely low . Previous studies using parasitologic and histopathologic examination , as well as PCR which was less diagnostic than microscopic identification , have incorrectly declared that vertical transmission does not occur after infection with L . infantum chagasi [23] . We suggest that using well tested kqPCR in conjunction with stringent standard curves provides a more sensitive method of detection . Parasites were detected via kqPCR in all shown tissues ( Figure 1 ) , consistent with the pattern of L . infantum infection observed in adult canines . Four pups were euthanized 12 weeks after birth and their tissues were analyzed histologically and via qPCR . Three of the four animals were positive for L . infantum in at least one tissue , but only one demonstrated disseminated infection , as kinetoplast DNA was detected in multiple tissues ( Figure 1A ) . When compared to pups that were euthanized right after birth , there appears to be less disseminated infection and a lower parasite load in older pups . We postulate that this could be a result of differential parasite transmission among the pups , and/or immune-mediated control of L . infantum infection in the months following birth . While we cannot rule out the possibility of differences in parasite transmission , analysis of the L . infantum-specific CD4+ T cell response in infected animals indicates neonates and pups are all able to mount antigen-specific responses against the parasite ( Figure 2 , pups 4–7 ) . While exposure to L . infantum antigen in utero could have led to the development of immunological tolerance [24] , our data indicates these pups are responsive to L . infantum antigen and possibly able to clear the parasite from detection in many tissue sources . This may be supported by our findings that pups euthanized 12 weeks after birth show decreased parasite dissemination and parasite numbers within infected tissues ( Figure 1B ) . To our knowledge this is the first description of a neonatal immune response to L . infantum infection . In North America , four species genus Lutzomyia sand flies feed on mammals . Lutzomyia anthorphora and Lu . diabolica are found in Texas , and Lu . cruciata is found in Florida and Georgia [16] . Lu . diabolica , as isolated in Texas , has been shown to be infected with Le . mexicana , and is likely to transmit cutaneous leishmaniasis in this region [25]–[27] . Lu . shannoni has been identified in Alabama , Arkansas , Delaware , Florida , Georgia [28] , Louisiana , Mississippi , North Carolina , South Carolina , New Jersey [9] and recently into the Midwest in Kentucky [29] and Ohio [30] . Experimental infection identified that Lu . shannoni , as found in South America , fed on clinically ill Le infantum-infected dogs became infected with Le . infantum [8] . Attempts have been made to find Le . infantum-infected sand flies in the environment of U . S . dog kennels without success [9] , [16] , [21] . In endemic areas the frequency of Leishmania-infected sand flies may range from only 0 . 2–1% however this frequency may also be as high as 5–7% [31]–[33] . Many entomologists believe that sand flies are playing a yet to be determined role in transmission of Leishmania in this country . Sand fly transmission of canine visceral leishmaniasis in the United States is a frightening possibility , but until parasite infection is found in domestic sand flies , and in the absence of human cases or cases in other breeds of dogs co-housed with Foxhounds , such transmission thankfully appears to be the exception and not the rule . The data presented here poses new challenges and considerations for the control of ZVL transmission . Disease prevention methods that solely target the vector may not be sufficient to control canine disease dissemination . In support of this , studies focused on the effect of collar or topical insecticides to prevent ZVL transmission do not observe transmission reduction below 4% [34] , [35] suggesting that infection may be maintained within a population through vertical transmission despite vector control methods . Altogether our data demonstrates for the first time vertical transmission of ZVL in North American dogs . Without evidence of a competent vector , we propose that vertical transmission may be a main mechanism for autochthonous L . infantum dissemination in the United States Foxhound population . | Dogs are a favored feeding source for sand flies that transmit human L . infantum infection . Zoonotic visceral leishmaniasis ( ZVL ) is an emerging problem in some U . S . dog breeds , with over 20% of at-risk Foxhounds infected . Although classically Leishmania is transmitted by infected sand flies which exist in the United States , no role has yet been determined for vector-borne transmission . Means of ongoing L . infantum transmission in U . S . dogs is unknown . Possibilities include transplacental and horizontal/venereal transmission . Aims for this study were to establish whether transplacental transmission occurred in Leishmania-infected U . S . dogs and determine the effect of this transmission on immune recognition of Leishmania . This novel report describes wide-spread infection as identified by kqPCR in 8 day-old pups born to a naturally-infected , seropositive U . S . dog with no travel history . This is the first report of transplacental transmission of L . infantum in naturally-infected dogs in North America . Evidence that mom-to-pup transmission of ZVL may continue disease in an otherwise non-endemic region has significant implications on current control strategies for ZVL . Determining frequency of vertical transmission and incorporating canine sterilization with vector control may have a more significant impact on ZVL transmission to people in endemic areas than current control efforts . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"infectious",
"diseases",
"public",
"health",
"and",
"epidemiology",
"diagnostic",
"medicine",
"women's",
"health",
"clinical",
"immunology",
"global",
"health",
"immunology",
"public",
"health",
"veterinary",
"science"
]
| 2011 | Transplacental Transmission of Leishmania infantum as a Means for Continued Disease Incidence in North America |
Endosymbiotic bacteria from different species can live inside cells of the same eukaryotic organism . Metabolic exchanges occur between host and bacteria but also between different endocytobionts . Since a complete genome annotation is available for both , we built the metabolic network of two endosymbiotic bacteria , Sulcia muelleri and Baumannia cicadellinicola , that live inside specific cells of the sharpshooter Homalodisca coagulata and studied the metabolic exchanges involving transfers of carbon atoms between the three . We automatically determined the set of metabolites potentially exogenously acquired ( seeds ) for both metabolic networks . We show that the number of seeds needed by both bacteria in the carbon metabolism is extremely reduced . Moreover , only three seeds are common to both metabolic networks , indicating that the complementarity of the two metabolisms is not only manifested in the metabolic capabilities of each bacterium , but also by their different use of the same environment . Furthermore , our results show that the carbon metabolism of S . muelleri may be completely independent of the metabolic network of B . cicadellinicola . On the contrary , the carbon metabolism of the latter appears dependent on the metabolism of S . muelleri , at least for two essential amino acids , threonine and lysine . Next , in order to define which subsets of seeds ( precursor sets ) are sufficient to produce the metabolites involved in a symbiotic function , we used a graph-based method , PITUFO , that we recently developed . Our results highly refine our knowledge about the complementarity between the metabolisms of the two bacteria and their host . We thus indicate seeds that appear obligatory in the synthesis of metabolites are involved in the symbiotic function . Our results suggest both B . cicadellinicola and S . muelleri may be completely independent of the metabolites provided by the co-resident endocytobiont to produce the carbon backbone of the metabolites provided to the symbiotic system ( . , thr and lys are only exploited by B . cicadellinicola to produce its proteins ) .
Intracellular symbiosis involves a unicellular organism ( the endocytobiont ) which durably lives inside the cells of the other partner ( the host ) . In the last century , the crucial role of intracellular symbiosis in the ecology and evolution of many eukaryotes was many times demonstrated [1] , [2] . Intracellular mutualism ( where the presence of the endocytobiont increases the fitness of both host and endocytobiont ) was particularly well described in several associations between insects and bacteria [3] , [4] . The association is most often metabolic: each partner provides metabolites that the other one cannot produce nor find in its environment . The complete genome annotation of mutualistic endocytobionts associated with insects revealed for all of them an extreme genome reduction paired with an extreme metabolism reduction [5] . Many metabolic functions of the bacterium are thus provided by the host and , inversely , the metabolism of the endocytobiont appears specialised into functions that are absent in the metabolism of the host . In addition , it often occurs that a host provides a habitat for more than one mutualistic intracellular bacterium species . This is the case for instance of the sharpshooter ( Homalodisca coagulata ) which hosts two bacteria: the -proteobacterium Baumannia cicadellinicola and the Bacteroidetes Sulcia muelleri . The complete genome annotation of the two endocytobionts revealed that their metabolic capacities are broadly complementary [6] , [7] . The metabolism of B . cicadellinicola is globally devoted to cofactor and vitamin biosynthesis whereas the metabolism of S . muelleri is specialised in the essential amino acid biosynthesis that the sharpshooter cannot produce nor find in its diet , the xylem sap . Nevertheless , the partition of these metabolic roles is not so perfect: B . cicadellinicola produces two essential amino acids , methionine and histidine , that S . muelleri cannot produce while the latter appears to be able to synthesise menaquinone , a vitamin . Moreover , the complementarity between the two metabolisms also concerns the biosynthesis of some metabolites not needed by the insect host , such as the fatty acid biosynthesis pathway , supplied by B . cicadellinicola , except for one step which is provided by S . muelleri [7] . However , these previous analyses were essentially manually performed by comparing the lists of annotated metabolic genes using as reference the metabolic pathways available in metabolic databases such as KEGG [8] or MetaCyc [9] . Even when highly reduced , a metabolic network is however complex enough that such an approach based on lists of genes and metabolic pathways could only give a partial description of the metabolic exchanges in the symbiotic system , even for those directly involved in the symbiotic functions of the endocytobionts . The aim of this study was therefore to determine the possible metabolic exchanges in the symbiotic system by a systematic and automatic exploration of the full metabolic networks of the two endocytobionts in order to detail those leading to the biosynthesis of metabolites involved in the symbiotic function of each bacterium . Defining in an exhaustive way the metabolic exchanges in a symbiotic system implies to be able to indicate all the metabolites needed by one partner and produced by another partner . Our first task was thus to identify for each endocytobiont the metabolites potentially imported from the host cell ( that is , the so-called “seeds” ) and produced by another partner . Focusing on the biosynthesis of the specific compounds that each bacterium produces and provides to the symbiotic system ( from now on , we denote such compounds by “targets” ) then necessitates to determine which sets of exchanged metabolites lead to their production . Our second task was thus to identify for each endocytobiont the subsets of seeds ( from now on , we denote such subsets by “precursor sets” ) that are sufficient to produce the targets , and to identify them all , that is all alternative precursor sets for each target . These two tasks , and particularly the second one , are hardly feasible by just manually inspecting the metabolic pathways inferred from genomic annotations . Such broad and systematic analyses require working with the full metabolic network to consider it in a systemic way . An intuitive way to define the seeds of a metabolic network is to consider as nutrient a metabolite not produced by any reaction but consumed by one or several ones . However , in particular because of reversible reactions that produce nutrients , this definition is not sufficient . Borenstein et al . extended the seed definition in a metabolic network by decomposing the metabolic graph into strongly connected components and then detecting those without incoming edge [10] . Whatever way is adopted to define the sources of a metabolic network , the next question is to determine which precursor sets are able to produce the targets . Romero et al . ( 2001 ) proposed a method returning alternative precursor sets for a set of target compounds [11] . Their algorithm was based on a backtrack traversing of the metabolic graph from the target compounds to the seeds . Unfortunately , how cycles are dealt with during backtracking is not described in the method . Another method proposed by Handorf et al . to find precursor sets is based on a forward traversing of the metabolic graph [12] . Their algorithm is based on the concept of scope defined as the set of compounds that a set of initial seeds is able to produce [13] . The way to find precursors proposed by Handorf et al . is then to test the reachability of several sets of seeds heuristically defined [12] . However , the method does not take into account cycles that may appear between the seeds ( see Methods ) and only provides a subpart of the possible precursor sets for a target compound . Recently , we proposed the first definition of minimal precursor sets that explicitly addresses the problem of cycles and the first exact method to find them [14] . In addition , our method , called PITUFO ( for “Precursor Identification To U For Observation” ) , is able to deal with any definition of seeds . To explore the metabolic exchanges occurring between B . cicadellinicola and S . muelleri , we first defined the set of seeds for each bacterium thanks to the method developed by Borenstein et al . [10] . By comparing with the compounds produced by each metabolic network , we were able to discriminate between the seeds produced by the co-endocytobionts and those potentially produced by the insect host or found in its diet . We then applied PITUFO to determine which subsets of seeds are involved in the biosynthesis of compounds already known to participate in the symbiotic function of each bacterium . The two steps are summarised in Figure 1 . Our study offers the first detailed and systematic description of the metabolic exchanges occurring in a symbiotic system . Our results also demonstrate the usefulness of graph-based dedicated methods in the metabolic analysis of multi-species systems .
Draft metabolic reconstructions for the bacterial genomes of Baumannia cicadellinicola and Sulcia muelleri were downloaded from the MaGe annotation system [15] . This platform makes available metabolic networks built from re-annotated genomes . Each metabolic network reconstruction is available in the pathway-tools format [16] . We restricted the networks to the small-molecule metabolism , meaning that reactions involving macromolecules such as nucleic acids or proteins were removed from the final metabolic networks . A first manual curation consisted in removing what may be considered as “fake” reactions . Indeed , an enzyme is potentially able to catalyse several reactions but a limited number of them actually takes place in a given organism . The reactions that clearly do not happen correspond to those that either are disconnected from the network ( they use as inputs compounds that are not produced in the organism and produce compounds not used as substrates by other reactions ) , or that are connected to the network only by cofactors . The topology of the network provides thus a clue to remove 14 reactions from the metabolic network of S . muelleri and 37 reactions from the metabolic network of B . cicadellinicola ( see Tables S1 and S2 ) . Such reactions can be automatically detected by a topological analysis but their elimination requires a manual inspection and some biological a priori since some reactions appear disconnected because of a hole in the network , that is , of a single reaction that is missing due to an error or incompleteness in the annotation process . The main clue to fill such holes is to inspect the completeness of the metabolic pathways predicted in the organism . By comparing these predictions with data from the literature , we are able to complete some metabolic pathways , and thus the metabolic network . In addition , several reactions involve generic compounds ( for instance , an aldehyde ) in the draft metabolic networks . When the same reaction existed with specific compounds , the generic reaction was simply removed . This is the case of nine reactions in the metabolic network of B . cicadellinicola . If specific reactions do not exist , the reaction has to be duplicated into several reactions so that they involve compounds already existing in the metabolic network . This is the case of the reaction RXN-8972 for which a substrate is “lysine or meso-diaminopimelate” . This reaction was thus splitted into two reactions ( RXN-8972BIS and RXN-8972TER in Table S4 ) that involve lysine and meso-diaminopimelate respectively . In automatic metabolic reconstructions , several reactions can be assigned to annotated enzymes . These reactions often use the same main substrates but different cofactors or even completely different substrates . When several reactions were assigned to a same enzyme , we removed the reactions for which the substrates required were absent in the metabolic network . This is the case of eight reactions in the metabolic network of S . muelleri and of 19 reactions in the metabolic network of B . cicadellinicola . We removed also six reactions in the metabolic network of S . muelleri and seven reactions in the metabolic network of B . cicadellinicola that were classified into the small molecule metabolism whereas they clearly involve macromolecules ( see Tables S1 and S2 ) . In the end , a total of 16 reactions in the metabolic network of S . muelleri and 58 reactions in the metabolic network of B . cicadellinicola were thus removed , using the clues mentioned above ( see Tables S1 and S2 ) . The gene ( epd ) that catalyses the production of erythronate-phosphate from erythrose-phosphate appears as absent in the genome of B . cicadellinicola . Wu et al . made the assumption that this role could be carried by glyceraldehyde 3-phosphate dehydrogenase [6] . We thus added this reaction to the metabolic network of B . cicadellinicola ( ERYTH4PDEHYDROG-RXN in Table S4 ) . McCutcheon et al . mentioned that no gene in S . muelleri could be assigned as argE or dapE , potentially coding for an enzyme catalysing one step in either the lysine biosynthesis and the arginine biosynthesis [7] . Interestingly , a gene ( SMGWSS-116 ) was assigned as argE in the genome of S . muelleri by the MaGe annotation system . We then considered the corresponding reactions in the two metabolic pathways as present ( SUCCDIAMINOPIMDESUCC-RXN and ACETYLORNDEACET-RXN in Table S3 ) . The direction of the reactions was first assigned based on the pathways where they are involved in the MetaCyc database [9] . A reaction is thus assigned as irreversible if it occurs in the same direction in all the MetaCyc pathways . If it was not possible to infer a unique direction , then the reaction remained reversible . The direction of most of the reactions in both metabolic networks were assigned in this way ( Tables S3 and S4 ) . Various manual corrections were also performed , essentially based on the constraints brought by the topology of the network and the biology of the organism , exemplified in Figure 2 . For instance , the molecular weight of compound in Figure 2 could be too large to allow a transport of the molecule . Furthermore , the classification of the reactions that are in the same metabolic pathway where and classically appear as intermediate metabolites can be an additional clue to assign the direction of . Eight reactions in the metabolic network of S . muelleri and 28 reactions in the metabolic network of B . cicadellinicola were hence assigned as irreversible ( Tables S3 and S4 ) . The whole set of reactions of the metabolic networks of S . muelleri and of B . cicadellinicola are displayed in Tables S3 and S4 . The metabolic networks of S . muelleri and of B . cicadellinicola are available in SBML format [17] in Datasets S1 and S2 . We restricted our study to the metabolism involving transfers of carbon atoms between molecules . In each reaction , we thus removed sets of molecules that do not participate into carbon exchanges . We called these metabolites “side compounds” . First , we established a list of 24 classical transformations between side compounds ( e . g . ) present in the metabolic networks of the two bacteria ( see Table S5 ) . When one of these transformations is identified in a reaction , the corresponding side compounds were removed from the reaction . Since whether metabolites are side-compounds in a given reaction is not always clear , some reactions were then manually corrected . The following inorganic compounds were also removed: water , proton , phosphate , diphosphate , ammonia , hydrogen peroxyde , sulfite , sulfate and oxygen . Reactions that do not imply a transfer of carbon atoms are also eliminated . This is for instance the case of the reactions involved in the sulfate reduction . The filtered metabolites are written in non-bold in Tables S3 and S4 . The filtered metabolic networks of S . muelleri and of B . cicadellinicola are available in SBML format [17] in Datasets S3 and S4 . In order to describe the metabolic exchanges between the endocytobionts , the first step consisted in identifying which metabolites each bacterium potentially acquires from its environment . For this , we based ourselves on the definition of Borenstein et al . of the seed set of a network: “the minimal subset of the occurring compounds that cannot be synthesized from other compounds in the network ( and hence are exogenously acquired ) ” [10] . To apply the Borenstein method to identify the seed sets , the metabolic network of each bacterium was modelled as a directed compound graph . In such a graph , nodes represent compounds and there is an arc between two compound nodes if at least one reaction produces one of the compounds ( possibly more ) from the other ( possibly more ) . A reversible reaction between two metabolites is modelled by two arcs with opposite directions linking them . Since the side compounds were previously filtered ( see previous Section ) , we avoid paths between metabolites that are biologically meaningless for our study . The seeds identification is based on the detection of the strongly connected components ( SCC ) in the compound graph . An SCC is a subgraph that contains a maximal set of nodes such that for any pair of nodes and in , there exists a path between and and a path between and . An SCC with no incoming arc is called a source component . Any compound inside a source component is a potential seed or , in our case , just seeds . This definition of seeds allows to take into account the uncertainty about the direction of some peripheral reactions and about the presence of reactions producing metabolites that are actually exogenously acquired . We invite the reader to refer to the paper of Borenstein et al . for more precisions [10] . For each symbiont , we further inspect the collection of seed sets identified in order to classify these seeds as potentially provided by the insect or by the other co-symbiont . This is done by analysing the feeding source of the host as described in the literature [6] , [7] or the metabolic network of the other co-symbiont to check whether these metabolites may be produced , and may therefore be supplied . We implemented a version of the Borenstein's method using the Igraph package [18] and applied it to the compound graph of each bacterium . Once the set of seeds was defined for each metabolic network , the next step was to identify which subsets of the seeds , henceforward called “precursor sets” , are sufficient to produce the metabolites known to be involved in the symbiotic metabolic association , that from now on we call the “targets” . Those are metabolites output by one bacterium that may then be used by the other co-symbiont or the host . We first put as targets the metabolites reported as involved in the symbiotic association by McCutcheon et al . [7] . We then added erythrose-4-phosphate , phosphoenolpyruvate , oxaloacetate and ribose-5-phosphate to the list of target compounds for B . cicadellinicola because of their presence both in its metabolic network and in the precursor sets identified for S . muelleri . These additional targets are particularly interesting since they could directly correspond to metabolic pathways shared between the two metabolic networks . For the same reasons , we added homoserine and 2-ketovaline to the list of target compounds for S . muelleri . To identify the precursor sets , we used the PITUFO method that we recently developed [14] . Given as input a metabolic network , a list of seeds and a set of target metabolites , PITUFO returns the list of all minimal precursor sets for the target metabolites . For the purposes of this paper , we consider single sets of target metabolites , that is sets with only one element . Notice that once we get as result all minimal sets of precursors that are able to produce a target , we are covering all alternative paths that may lead to its production . understand the reasoning . The strength of PITUFO comes from the fact that it takes into account cycles in the definition of precursor sets in a fully formalised manner . This allows to find paths from the precursor sets to the targets that pass through cycles in the network but are still feasible . Previous methods , such as those that compute the scope of a subset of the seeds as defined by Handorf et al . [13] and were later used to test the reachability of a target compound from a set of seeds [12] , fail to link some sets of seeds to a target compound if there is such a cycle in the paths between them . The scope of an initial seed set is itself and then any metabolite that can be produced using only substrates already in and added to it until no new compound can be produced [13] . This iterative process is called forward propagation [11] or network expansion [13] . The strategy of PITUFO to deal with cycles is to allow the use of metabolites involved in cycles if they are also produced ( regenerated ) in the forward propagation from the seeds to the targets . Indeed , in Figure 3 , the scope of the set does not contain but if we allow the use of or in the forward propagation process , then the scope of contains . However , this could lead to clearly unrealistic paths without the constraint of regeneration of the compounds involved in cycles . For instance , in the same figure , if we allow the use of ( and possibly also ) , the scope of contains also but uses up all of and unless both were in infinite supply , in which case they should be considered as seeds . The metabolites or , inside a cycle in Figure 3 , that may be used and that are regenerated when the network is fired from a subset of the seeds ( the set in the figure ) , are called “self-generating metabolites” [14] . Observe that these are defined in relation to a subset of the seeds . They do not need to be given as input but will be identified by the algorithm together with the sought precursor sets . The following definitions allowed us then to formally establish what is a precursor set of a given target compound: a subset of the set of seeds is considered as the precursor set of a target compound if there exists a set of metabolites such that the scope of , allowing the use of in the forward propagation process , contains and all the metabolites in , which ensures the regeneration of . The set is considered as a minimal precursor set if there is no set strictly contained in that verifies this property . In the above definition ( and in [14] ) , a metabolic network is considered as an hypergraph . Nodes are metabolites and there is an hyperarc between two sets of metabolites if there is a reaction that produces one of the sets from the other . Contrary to simple graphs where the nodes represent either compounds or reactions and arcs link individual nodes , the topology of an hypergraph takes into account the need in general for more than one substrate to activate a reaction [19] . The use of hypergraph modelling allows the formalisation of hyperpaths between precursor sets and targets since when a reaction is modelled as an hyperarc , it already explicitly establishes that all of its substrates are needed . On the other hand , an hypergraph could lead to some confusion for the method previously applied to identify the seed sets and proposed by Borenstein et al . [10] since there is no clear definition of strongly connected components of an hypergraph . For instance , isolated vertices may not be considered as SCCs and there is no unique definition of cycles in hypergraphs . For these reasons , the method was applied as in the original work on a compound graph representation of the metabolic network . Since PITUFO is an exact method , it is enough to describe its input and output without recalling how the second is produced from the first . For those interested in the method itself , the algorithm is described in detail in [14] . The current version of the algorithm was implemented in Java and takes as parameter an SBML file describing a metabolic network [17] and a file containing a list of seeds and one or a list of target compounds . The reconstructed metabolic networks complete or filtered are available in the Supplementary material . The method used is available at this address: http://sites . google . com/site/pitufosoftware/ . Figures S1 to S20 display the sub-networks linking each target metabolite to their precursor sets . These reconstructions were performed from the PITUFO results using the visualisation software Cytoscape [20] .
Table 1 shows the number of reactions and compounds in each metabolic network as indicated in the Reconstruction Section . As mentioned in previous studies , both metabolic networks are extremely reduced . The metabolic network of B . cicadellinicola is less than half the size of the metabolic network of the free bacterium Escherichia coli . The reduction is even more important in S . muelleri since , with only 64 reactions , its metabolic network is less than ten percent the size of the network of E . coli . In both cases , the extensive manual curation allowed to highly reduce the number of reversible reactions as we succeeded to assign a direction to most of the reactions in the two metabolic networks . Figure 4 displays the set of seeds identified in the metabolic network for each bacterium . Coloured arrows mark those produced in the metabolic network of the co-endocytobiont and those potentially provided by the insect host according to the literature . Seeds that correspond to annotated transport reactions are also tagged . We recall that the reactions involving big molecules are not taken into account in this analysis . For instance , the reactions charging amino acids onto their corresponding tRNAs do not appear in the metabolic network we built . This means two things . First , an amino acid involved only in the production of proteins does not appear in the seeds that we identified , which explains the absence of some essential amino acids in the set of seeds identified in B . cicadellinicola . Second , this also means that an amino acid identified as a seed is involved in the small molecule metabolism and not only in the production of proteins . Moreover , we focused on the transfers of carbon atoms . There are then no inorganic metabolites in the sets of seeds . Furthermore , the organic compounds not involved in carbon atom transfers do not appear in this list . This is the case for instance of glutamine that appears as a source of nitrogen but not of carbon in the metabolism of B . cicadellinicola . This metabolite is thus a seed in the original metabolic network of the bacterium but not in the filtered one . We identified 19 seeds in the metabolic graph of B . cicadellinicola and 10 seeds in the metabolic graph of S . muelleri . Only three seeds are common to the two sets: serine , aspartate and bicarbonate ion . Whereas none of the seeds in S . muelleri could be linked to a transport reaction , four seeds ( lysine , glutamate , aspartate and glucose ) correspond to transport reactions annotated in B . cicadellinicola . Furthermore , a general amino acid ABC transporter annotated in its genome enables B . cicadellinicola to import also other amino acids identified as seeds: threonine , glycine , tyrosine and alanine . Among the 10 seeds identified in S . muelleri , three are amino acids ( cysteine , aspartate and serine ) and three are sugars: erythrose-4-phosphate and ribose-5-phosphate are classically produced by the pentose phosphate pathway and ribose-5-phosphate by the glycolysis pathway . Among the 19 seeds identified in B . cicadellinicola , we found 13 amino acids or related metabolites ( such as homoserine or 2-ketovaline ) and only one sugar , glucose . In the compound graph of B . cicadellinicola , serine belongs to the same source component ( see Methods ) as threonine and glycine . In fact , there are two reversible reactions that , respectively , link serine and threonine to glycine ( Figure 5 ) . It is thus impossible to distinguish which one ( s ) actually produces the other ( s ) . All were considered as potential seeds and were taken into account by the PITUFO method for detection of the precursor sets . There is only one other example of such alternative seeds detected by the method of Borenstein [10]: these are oxaloacetate and aspartate linked by the same reversible reaction in the metabolic network of S . muelleri ( Figure 6 ) . All the other seeds found are metabolites that are not produced by any reaction . Among the 16 seeds specific to B . cicadellinicola , five are produced by S . muelleri and two are certainly not provided by the insect host: threonine and lysine . Five seeds were already mentioned as potentially provided by the sharpshooter: glucose 6-phosphate , tyrosine , glycine , glutamate and alanine . Protoheme and porphobilinogen were mentioned by Wu et al . as needed to be imported by B . cicadellinicola to complete the siroheme biosynthesis pathway [6] . They seem not to be produced by S . muelleri and should be provided instead by the insect . Five seeds identified in the metabolic graph of S . muelleri are produced by B . cicadellinicola: erythrose-4-phosphate , phosphoenolpyruvate , ribose-5-phosphate , octaprenyl-diphosphate and oxaloacetate , but all could be also available in the insect cell . Other seeds that do not seem to be produced by the co-endocytobiont were not reported before as potentially provided by the host . We assume that these seeds are produced by the insect or present in its diet . Knowledge of the metabolic network of the sharshooter should confirm or disprove the production of these metabolites by the insect host . Figures 7 and 8 indicate the precursor sets for the target metabolites selected in the metabolic networks of the two endocytobionts ( see Methods ) . The parsimony of the metabolic network of both bacteria is reflected in the small number of precursor sets found for the target metabolites to which we applied PITUFO: the maximum number of solutions for a target is only three and the maximum total number of involved precursors is seven . For S . muelleri , apart from menaquinone , all targets are amino acids , which explains the uniformity of the results . Two seeds are present in all minimal precursor sets computed for these amino acids ( except homoserine ) : erythrose-4-phosphate and phosphoenolpyruvate . Both are potentially provided by B . cicadellinicola . We found oxaloacetate and aspartate as alternative precursors for the synthesis of isoleucine and lysine . Indeed , each one can produce the other by the same reversible reaction in the metabolic network of S . muelleri ( Figure 6 ) . This leads to two possible scenarii , depending on which one of them is actually provided . Aspartate is one of the primary components of the xylem sap , it is thus reasonable to think that this compound should be provided by the host . In the metabolic network of S . muelleri , aspartate is involved in other reactions that in particular participate in the synthesis of other amino acids . Oxaloacetate is only involved in the reaction that produces aspartate . Since S . muelleri is able to produce aspartate from oxaloacetate , and since the former is not used in other reactions , the import of oxaloacetate by S . muelleri seems to be more realistic than the import of aspartate . Moreover , B . cicadellinicola could provide oxaloacetate while the bacterium is able to synthesise it from aspartate ( see Figure 8 ) . Some seeds appear as obligatory in the synthesis of several targets in B . cicadellinicola . Glucose and aspartate , reported as provided by the insect cell , thus appear as obligatory for the synthesis of , respectively , twelve and five target compounds . As mentioned previously , serine , glycine and threonine have been detected as alternative seeds by the Borenstein method [10] . For B . cicadellinicola , they appear in the minimal precursor sets for methionine , coenzyme A , glutathione and thiamine . McCutcheon et al . suggested that homoserine and 2-ketovaline , potentially provided by S . muelleri , could be precursors of metabolites supplied by B . cicadellinicola . Homoserine was reported as a precursor of methionine and 2-ketovaline as a precursor of coenzyme A [7] . Our results confirm these hypotheses . For methionine , our method adds precision by indicating also the alternative precursors serine-glycine-threonine . For coenzyme A , our method further suggests this triplet and also -alanine as obligatory precursors . Interestingly , we observed that only methionine and coenzyme-A require metabolites provided by S . muelleri . Moreover , the metabolites needed by the other targets could be all potentially acquired from the host cell by B . cicadellinicola .
Graph-based modelling of the metabolic networks of B . cicadellinicola and S . muelleri enabled us to complete and precise the description of the metabolic exchanges between these two endocytobionts and with their host , the sharpshooter Homalodisca coagulata . By automatically computing the set of seeds for each metabolic network , we thus offer the first exhaustive list of metabolites potentially imported by S . muelleri and B . cicadellinicola . By using our method to find precursor sets for given target compounds , we provide a general and detailed view of the metabolic exchanges that potentially lead to the synthesis of metabolites involved in the mutualistic association . The definition of seeds by Borenstein et al . [10] allowed to indicate alternative ones that could not be found only by defining the seeds as metabolites not produced by any reaction . We found only two instances of such alternative seeds: oxaloacetate-aspartate in the metabolic network of S . muelleri and glycine-serine-threonine in the metabolic network of B . cicadellinicola . The method of Borenstein et al . remains highly suitable to detect seeds in metabolic networks where many reactions cannot be assigned a direction . From the list of seeds previously defined , the method we developed , PITUFO , was able to find the precursor sets of metabolites reported as involved in the symbiotic association . Contrary to previous methods , PITUFO is an exact algorithm and returns all the precursor sets for a given target compound . In addition , by explicitly taking into account cycles in the definition of precursors and in the algorithm , PITUFO is able to find solutions not reachable by the previous methods . Unfortunately , because the implementation of previous methods is not available or is dataset-dependent , we were not able to compare their performance with the one of PITUFO . Most of our results could only hardly be found by manual analysis of the metabolic pathways . However , the pertinence of our or of previous analyses is highly linked to the quality of the metabolic network reconstructions . The most time-consuming part in this study was then to refine the metabolic reconstructions available for the two bacteria ( see Methods ) . Interestingly , the methods we used to find seeds and precursor sets also helped us to refine the metabolic reconstructions when some inconsistencies were found . There are several ways to complete this study and to improve the tools that we used . First , PITUFO only returns sets of precursors and not the possible hyperpaths between them and the target compounds . The identification of key metabolites , such as those involved in hyperpath intersections , and compression of the information contained in the metabolic hyperpaths could be a way to provide results easier to interpret for the analyst . When alternative precursor sets are indicated , it would be interesting to point to those that are the most likely to be actually used . Taking into account the stoichiometric coefficients would allow to prune precursor sets not consistent with the stoichiometric constraints . Measuring the production rate of the target compounds would be a way to sort the precursor sets . Finally , PITUFO is restricted to the identification of all the minimal precursor sets leading to the production of the set of targets specified by the user , putting aside the other metabolic functions , even vital for the organism . The identification of all minimal precursor sets leading to the production of both essential metabolites ( e . g . those participating to the biomass ) and metabolites involved in the mutualistic association was beyond the scope of this study but is certainly of interest and will be developed in the future . For both bacteria , the number of seeds that we identified is very reduced , even considering that our study is limited to the carbon metabolism of small molecules . This means that the global reduction of the metabolism in the symbionts comes with a reduction in the number of metabolites imported from the host cell . The identification by PITUFO of a unique precursor set for most of the selected target compounds shows that there are almost no alternative sources to produce essential compounds . Indeed , the mutualistic association of the symbionts with their host is very ancient ( 70 to 100 millions years for B . cicadellinicola and approximatively 280 millions years for S . muelleri ) [21] . The stability of their environment , particularly because of their vertical mode of transmission , made their metabolism specialised in the exploitation of a restricted set of substrates . However , our results showed that the two bacteria use very differently their environment . Indeed , only three seeds common to the two metabolic networks have been identified . Even the common seeds have a completely different fate in the two bacteria as they are not involved in the same metabolic pathways . The complementarity of the two metabolisms is then not only manifested in the metabolic capabilities of each organism but also by their different use of the nutrients available in the host cell . The set of seeds identified in the metabolic network of B . cicadellinicola is mainly composed of amino acids or related metabolites such as 2-ketovaline or homoserine . Three seeds identified in the metabolic network of S . muelleri are also amino acids . The presence of amino acids in the seeds identified in B . cicadellinicola ( glutamate , lysine , alanine , serine , aspartate and glycine ) or in S . muelleri ( serine , aspartate and cysteine ) is interesting in the sense that they are not only provided as essential building blocks of proteins but also as starting points of the biosynthesis of other metabolites . This clarifies the role of the exchanged amino acids as reported in earlier studies [6] , [7] . Conversely , the absence of other amino acids in the seeds indicates that they do not participate to the formation of the carbon backbone of other compounds . Some seeds automatically defined in our study were already mentioned in earlier studies [7] . Aspartate , identified as a common seed in both metabolic networks , is an amino acid indicated to exist in great concentration in the xylem sap that the sharpshooter feeds upon [7] . Its large availability in the direct environment of the two bacteria makes of it an efficient source for the production of other metabolites . Furthermore , a specific aspartate transporter has been annotated in the B . cicadellinicola genome . Other metabolites such as non-essential amino acids and glucose , the only sugar identified as seed in the metabolic network of B . cicadellinicola , were also mentioned as components of the xylem sap and then available for the two symbionts [7] . However , some metabolites mentioned as highly present in the xylem sap were not identified in our set of seeds . For instance , arginine , an amino acid which is abundant in proteins , is absent from the metabolic network of B . cicadellinicola and appears only as output in the metabolic network of S . muelleri . Glutamine in both metabolic networks is only used as a nitrogen source and thus does not appear in the filtered metabolic networks . Malate is completely absent from the metabolic network of S . muelleri . It is produced from fumarate in the metabolic network of B . cicadellinicola and then does not need to be imported . Three seeds identified in the metabolic network of S . muelleri are glycolytic products . Erythrose-4-phosphate and ribose-5-phosphate are commonly produced in the pentose phosphate pathway . Phosphoenolpyruvate is commonly produced during glycolysis . PITUFO returned erythrose-4-phosphate and phosphoenolpyruvate as obligatory precursors in the synthesis of the metabolites that S . muelleri provides to the symbiotic system , except homoserine and menaquinone . Ribose-5-phosphate was identified as obligatory precursor for tryptophan . These three metabolites , as well as oxaloacetate and octaprenyl diphosphate , are produced by the metabolic network of B . cicadellinicola . However , it is likely that these metabolites could be made available by the insect host . This means that the carbon metabolism of S . muelleri may be completely independent of the metabolic network of B . cicadellinicola . On the contrary , the two essential amino acids ( threonine and lysine ) identified as seeds for B . cicadellinicola are certainly not produced by the insect host nor present in its diet and must be provided by S . muelleri . The carbon metabolism of B . cicadellinicola therefore appears as dependent on the metabolism of S . muelleri , at least for these two amino acids . This dependence is added to the obligatory supply of other essential amino acids by S . muelleri that are required for protein biosynthesis in B . cicadellinicola [6] , [7] . However , among the precursors identified for the synthesis of metabolites that B . cicadellinicola passes on to the symbiotic system , only methionine and coenzyme A need metabolites produced by S . muelleri: homoserine and 2-ketovaline . The first one may be produced by the plant and then be present in the diet of the insect , and the second one may be produced by the insect via the degradation of valine . This suggests that B . cicadellinicola , as well as S . muelleri , may be only dependent on the metabolites obtained from the insect to produce the metabolites provided to the symbiotic system . Indeed , threonine and lysine which are supplied by S . muelleri to B . cicadellinicola , are only exploited by the latter to produce its proteins . The reconstruction of the metabolic network of the host and a better knowledge about the metabolome of the plants that the insect feeds upon will inform us whether some seeds , such as intermediates in the biosynthesis of essential amino acids , are actually produced by the insect or present in its diet . One challenging issue concerns how the metabolites are exchanged between the three partners . The annotation of transporters for amino acids and sugars in B . cicadellinicola [6] gives only a partial answer to this question . Indeed , very few transporters were found during the annotation of the genome of S . muelleri , and none corresponds to the seeds that we identified [7] . This means that other scenarii have to be proposed to explain the exchanges of metabolites in the symbiotic system . In particular , the cells of B . cicadellinicola often appear to adhere to the surface of the much larger cells of S . muelleri [6] . This proximity should facilitate the exchanges between the two bacteria . However , the two bacteria seem to be not always in the same cells [6] . This poses the problem of how the essential amino acids are provided to B . cicadellinicola . One other remaining interesting question is evolutionary: how did the reductions of the metabolism of the two symbionts get organised during evolution to reach their current complementarity ? A recent study compared the metabolic gene sets of two pairs of co-resident endocytobionts . One pair was formed by the two endocytobionts of the sharpshooter studied in this paper and the other pair was formed by another strain of S . muelleri ( SMDSEM ) and by Hodgkinia cicadicola , found in the cells of cicadas [22] . The authors showed that the two strains of S . muelleri exhibit almost identical metabolic capabilities . They suggested also that , although phylogenetically distant , H . cicadicola and B . cicadellinicola have converged on similar metabolic functions , especially those that are complementary to the metabolism preserved in S . muelleri . The application of such methods as we used in this study to this other pair of co-resident endosymbionts should allow to identify some common patterns in the sharing of a set of nutrients and their use in the metabolic networks of the different partners . The extension to other endosymbiotic systems could provide us with crucial information to understand the establishment of such nutritional associations . | Some bacteria , called endocytobionts , permanently live inside the cells of a pluricellular organism and often bring an adaptative advantage to their host by providing compounds that the latter cannot produce or find in its diet . The association may involve several species of bacteria within the same host . The sap-feeding insect called glassy-winged sharpshooter thus maintains a permanent metabolic association with two different species of bacteria that it hosts within specialised cells . Complete genome annotations of the two endocytobionts allowed to reconstruction of their metabolism . By manually inspecting those annotations and comparing them to reference metabolic functions , earlier studies revealed a great complementarity between the metabolisms of the two endocytobionts and indicated potential metabolic exchanges between them . However , the metabolism of an organism is complex enough that such an approach could only give a partial description of the metabolic exchanges in the symbiotic system . We therefore determined all the metabolic exchanges in the symbiotic system by a systematic and automatic exploration of the full metabolism of the two endocytobionts in order to detail those leading to the biosynthesis of compounds involved in the symbiotic function of each bacterium . Our results highly refine our knowledge about the complementarity and the connections between the metabolisms of the two bacteria and their host . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"computational",
"biology/metabolic",
"networks"
]
| 2010 | Graph-Based Analysis of the Metabolic Exchanges between Two Co-Resident Intracellular Symbionts, Baumannia cicadellinicola and Sulcia muelleri, with Their Insect Host, Homalodisca coagulata |
Accumulating evidence suggests that many tumors have a hierarchical organization , with the bulk of the tumor composed of relatively differentiated short-lived progenitor cells that are maintained by a small population of undifferentiated long-lived cancer stem cells . It is unclear , however , whether cancer stem cells originate from normal stem cells or from dedifferentiated progenitor cells . To address this , we mathematically modeled the effect of dedifferentiation on carcinogenesis . We considered a hybrid stochastic-deterministic model of mutation accumulation in both stem cells and progenitors , including dedifferentiation of progenitor cells to a stem cell-like state . We performed exact computer simulations of the emergence of tumor subpopulations with two mutations , and we derived semi-analytical estimates for the waiting time distribution to fixation . Our results suggest that dedifferentiation may play an important role in carcinogenesis , depending on how stem cell homeostasis is maintained . If the stem cell population size is held strictly constant ( due to all divisions being asymmetric ) , we found that dedifferentiation acts like a positive selective force in the stem cell population and thus speeds carcinogenesis . If the stem cell population size is allowed to vary stochastically with density-dependent reproduction rates ( allowing both symmetric and asymmetric divisions ) , we found that dedifferentiation beyond a critical threshold leads to exponential growth of the stem cell population . Thus , dedifferentiation may play a crucial role , the common modeling assumption of constant stem cell population size may not be adequate , and further progress in understanding carcinogenesis demands a more detailed mechanistic understanding of stem cell homeostasis .
Certain aspects of the cancer stem cell hypothesis have previously been addressed by mathematical models . It has been shown that having a hierarchical tissue design , where a small population of stem cells maintains a transient population of differentiating cells , may slow the accumulation of mutations and protect against cancer [26]–[28] . The question of whether genetic instability ( resulting in hyperactive mutation rate ) is an early or later event in mutation acquisition leading to cancer has been addressed by several groups ( see [4] for review ) . Most mathematical models find that the onset of genetic instability should be an early event , if at least some of the mutations are neutral . However , sequencing suggests that the mutator phenotype is expressed relatively late in cancer progression [9] . Stem cell populations are typically small . Hence , the dynamics of mutant cells in the stem cell population are highly sensitive to stochastic fluctuations . A tumor begins with a single mutated cell , so there is a substantial chance of mutant extinction due to random events . Genetic drift and stochastic clonal extinction in stem cell lineages have been experimentally demonstrated for both normal tissue stem cells [29]–[31] and cancer stem cells [2] in several tissue types . Consequently , a deterministic model of mutation acquisition in stem cells will significantly underestimate the time to cancer establishment [32] . Many models of mutation acquisition use a stochastic approach and are concerned with calculating time to emergence or fixation ( or when the number of mutant cells reaches some threshold value used in diagnosis ) of a mutant cell with fitness in a population of size . The waiting time for cancer is often defined as the time until a particular number of mutation events have occurred in at least one cell . Iwasa et al . [33] considered a two-stage Moran model and described conditions under which “stochastic tunneling” can occur . ( In this phenomenon , cells with two mutations reach fixation before cells with one mutation reach fixation . ) Durrett et al . [34] obtained asymptotic estimates of waiting times until a cell with mutations first appears under the assumption of neutrality ( ) . These models typically consider a fixed population size [5] , [23] , [25] , [35]–[39] . The fixed population assumption is supposed to reflect homeostasis in the stem cell population , though how homeostasis is achieved is typically not addressed . Although the Moran model captures the stochastic nature of mutation acquisition , this type of model is not capable of describing mutations that change the stem cell division pattern and result in possible expansion of the stem cell pool , which in turn leads to tumor growth . Some recent models also consider mutation accumulation in exponentially growing cell populations [40]–[43] . Beerenwinkel et al . [6] used the Wright-Fisher model with exponentially growing population size to look at the effect of selection on the waiting time to cancer , and they predicted that the observed genetic diversity of colorectal cancer genomes can arise under a normal mutation rate ( taken to be per cell division ) if the average selective advantage per mutation is on the order of 1% . Similar calculations using a discrete branching process found % given [40] . Note that increased mutation rates due to genetic instability would allow even smaller selective advantages during tumorigenesis , but neutral mutants ( ) result in waiting times that are too long compared with disease incidence . Other groups have also concluded that for normal mutation rates and neutral mutants , mutations in multiple genes in acquired hematopoietic disorders are most likely very rare events , as acquisition of multiple mutations typically requires development times that are too long compared to disease incidence [36] . Spencer et al . [44] and Ashkenazi et al . [45] have focused on the sequential order of mutations associated with increased rate of proliferation , decreased rate of death , increased mutation rate , and other hallmarks of cancer that must accumulate before emergence of cancer . The sequence of mutations with the shortest waiting time to getting all the necessary mutations is considered the most likely mutational pathway [25] , [44] . However , these models do not consider the possibility that dedifferentiation of progenitor cells can affect the time to multiple mutation acquisition . The dividing progenitor cell population has previously been described by multi-compartment ODE models , with cells moving between compartments as they age [45]–[47] . Note that in these models the exact number of different stages of differentiation is ambiguous and does not exactly correspond to mitotic events , as cells may undergo more than one division in each compartment stage [46] . Most of these models of age-structured cell populations assume a stem cell proliferation rate that is dependent on the total number of cells and thus incorporate negative feedback as a means of achieving homeostasis [48] , [49] . These deterministic models have focused on mechanisms that could regulate cell numbers that are necessary for homeostasis and efficient repopulation . We use a similar mathematical approach to model the progenitor population as [49] , but we couple it to stochastic dynamics in the stem cell compartment . Upon division a stem cell can produce zero , one , or two stem cells with probabilities , , and , respectively ( Fig . 1A ) . The mean number of stem cell offspring is given by . If symmetric divisions are permitted , the stem cell population can be described by a branching process with the expected number of cells at time given by . However , a branching process either goes extinct or undergoes exponential growth , and thus it cannot capture stem cell dynamics at equilibrium . One solution is to use a conditional branching process [50] , where the probabilities for a branching process are conditioned to the total population size remaining constant by an unspecified sampling mechanism ( i . e . , assuming that the stem cell population remains in homeostasis ) . Some theoretical studies have previously considered the impact of the asymmetry of cell division on stem cell dynamics . However , these stochastic models all assumed a fixed stem cell population size , either through a variant of the Moran process [35] , [51] or conditional branching process [39] . We utilize a different approach to get a time-varying but bounded stem cell population size in our models . We use mathematical modeling to study how the possibility of “dedifferentiation” of mutant progenitor cells into a stem cell-like state affects the waiting time to carcinogenesis . Dividing progenitor cells have large growing populations , so we use a deterministic model to describe their evolutionary dynamics . For stem cell populations , stochastic effects are important , because the proliferating stem cell population is typically small . We use a stochastic model for stem cell dynamics as a boundary condition to the PDE governing differentiated cell expansion ( Fig . 1B and C . ) There is also feedback from the deterministic progenitor population to the stochastic stem cell population as a rate of “dedifferentiation” . To assess the effect of dedifferentiation on time to carcinogenesis , we consider models for stem cell dynamics with both fixed and variable stem cell numbers ( Fig . 1D ) . The main questions we address are: Our general compartment model can be applied to different tissues , such as colonic crypts , mammary cells , and hematopoiesis .
Extending Eq . ( 1 ) to account for mutations between multiple subpopulations of progenitor cells ( Fig . 1C ) we obtain ( 2a ) ( 2b ) ( 2c ) Here is the mutation rate per cell per unit time and is the number of progenitor cells of “age” from the subpopulation with mutations . We assume , and no back mutation is allowed . Let be the number of stem cells with mutations at time . Let be the probability of a symmetric division that gives rise to two differentiated cells , be the probability of an asymmetric division that gives rise to one stem cell and one differentiated cell , and be the probability of a symmetric division that gives rise to two stem cells . Then ( 3 ) If we neglect mutation , the steady wave-form solutions of Eq . 2 have the form ( 4 ) where is the average number of stem cells of type produced per division and is the age-dependent growth rate of the differentiated cell population ( Text S1 ) . Hence , the long-term age distribution is largely determined by the functional forms of the differentiated cell birth and death rates ( Fig . S1 and S2 . ) Altered birth and death rates due to mutations can result in mutant subpopulations growing to higher plateaus in size , but the final population size will be bounded . Our PDE system can be easily modified to have a maximal carrying capacity for each sub-population . This does not qualitatively change the age distribution of progenitor cells ( Fig . S1 ) and does not significantly affect the fraction of -mutation cells in the total progenitor population ( Fig . S3 ) , so we do not consider it further . To mimic a maturity switch for cellular proliferation and death , we took the proliferation and death rates of differentiated cells per unit time to be ( 5a ) ( 5b ) Here and are the maximal proliferation and death/removal rates of progenitor cells . The age at which the proliferation switch occurs ( i . e . , half the progenitor cells stop dividing ) is given by , and the steepness of the proliferation switch is determined by . Similarly , the age at which half the cells begin to undergo apoptosis is given by , and the steepness of the death switch is controlled by . If , then differentiated cells between the ages of and are not replicating ( senescent ) . Note that setting either of these values to zero results in a uniform rate of birth/death . Effects of varying proliferation/death parameters are shown in Fig . S2 . The parameters governing proliferation , in particular and , have much larger influence on the final differentiated cell population size than parameters governing death/removal . The steepness of the switch does not substantially change the age distribution . Parameters used are summarized in Table 1 . We used parameter estimates from the human hematopoietic system because parameters for other cancers are less well known . We used as the number of necessary mutations to develop a cancerous phenotype . Although it has been estimated that for the human hematopoietic system there are 11 , 000–22 , 000 stem cells [59] , which give rise to all blood and immune system cells , most of these cells are quiescent and only divide when body sustains an injury and needs to repopulate the hematopoietic system . Our model only considers actively dividing stem cells , which have been estimated by various methods to number around 100 [32] , [60] . The entire actively dividing stem cell population has previously been modeled as turning over once per year [32] , but most recent estimates have an individual stem cell dividing every 25–50 weeks [61] . However , this is likely an over-estimate , as it is difficult to distinguish between actively dividing and quiescent stem cell populations . We assume that an active stem cell divides every 20 weeks , which when multiplied by results in an active stem cell population turnover time of weeks . ( The entire stem cell population including quiescent cells turns over on a much longer timescale . ) Whereas the size of the active hematopoietic stem cell pool is small , the number of progenitor cells such as granulocyte , erythroid , monocyte , and megakaryocyte colony-forming units ( CFU–GEMM ) and granulocyte and monocyte colony-forming units ( CFU–GM ) is much larger . There are approximately CFU–GEMM cells and CFU–GM cells [62] . There are estimates that each CFU–GEMM may contribute to hematopoiesis for an average of 60 days ( range of 40–340 days ) and that it replicates at an average rate of once every 50 days ( range of 35–285 days ) [62] . We track the progenitor populations for weeks , and assume that their proliferative potential rapidly drops off after 10 weeks . The maximal proliferation and death rates , and were chosen so that 100 stem cells results in progenitor cells of all ages . Not much is known about the selective advantage provided by driver mutations for different cancer types , except that it is small ( ) . Unless stated otherwise , we assume neutral fitness in the stem cell pool ( ) in our stochastic models throughout the paper , to focus on the effect of dedifferentiation . We use a range of for the progenitor cells in the deterministic model . Mutation estimates per cell division per gene range from about in normal cells to in the case of chromosomal instability [63] . ( Note that the rate of epigenetic change has been estimated to be orders of magnitude higher than that of genetic change and could also play a role in cancer initiation [10] . ) A common value used in many mathematical models is a driver mutation rate of per division , obtained by assuming a somatic mutation rate of per gene , and about 100 genes that could be mutated to give same phenotype [40] , [45] . In normal hematopoietic cells the mutation rate has been measured as per division [64] . Note that in the stochastic model , which considers every cell division , the mutation rate can be used as is , but using chronological time ( i . e . , weeks or months ) means that this value should be multiplied by the average number of divisions per unit time to obtain . ( Mutations that speed up the cell cycle will then speed up the apparent mutation rate per unit of chronological time in our progenitor model . ) The expected number of doublings from stem to progenitor cells is [46] , and the total number of progenitors cells of type is . Using values from Table 1 , this results in cell divisions that take place over 10 weeks , so in equations ( 2 ) .
We first considered whether mutation and reproduction in the progenitor population could by itself generate a sustained population of two-mutation cancerous cells . We thus modeled a scenario in which no stem cell mutations occur , so the boundary condition to the progenitor population system in equations ( 2 ) is simply . Because selection in the progenitor population might favor mutants , we also assumed that progenitor cells with mutations have a proliferation rate ( Eq . ( 5 ) ) . This yields a steady-state age distribution of normal and mutant progenitor cells ( Fig . S2 ) . Fig . 2 summarizes results for typical parameter values , showing that for mutant cells to be an appreciable fraction of the population , the mutation rate and proliferative advantage must both be unreasonably high . This is true both if the total progenitor population can grow without bound ( Fig . 2a ) and if its growth is restricted ( Fig . 2b ) . Similar findings are obtained if competition between progenitor subpopulations is included in the model ( Fig . S3 ) . Consistent with previous work [26] , [27] , [36] , these results show stem cell dynamics cannot be ignored in considering time to carcinogenesis , so we next considered stochastic models of the stem cell population . Our first models for stem cells did not incorporate dedifferentiation , so the dynamics were entirely governed by the stem cells . In modeling cancer , the time to carcinogenesis can be defined as the time for a single -mutation cell to emerge , the time for -mutation cells to pass some threshold number or fraction , or the time for -mutation cells to fix in the population . If the mutation rate is low ( such that ) , then all three definitions are similar , because the time to emergence of a successful -mutation cell is long compared to the time from emergence to fixation . However , there is large uncertainty regarding effective mutation rates in carcinogenesis ( Table 1 ) , so the assumption of low mutation rate may not always be valid , and we thus calculated times to fixation . We began our stem cell modeling by considering fixed population size , corresponding to strict homeostasis . In this constant case , we could leverage several analytic results , with which our simulations agreed well . Fig . 3A shows a typical simulation . The full probability density distribution of time to fixation is given by Eq . ( 12 ) and agrees well with our simulations for high mutation rates ( Fig . 3B ) . The time to emergence of a successful mutant is of order 1/ ( ) stem-cell generations ( Eq . ( 9 ) ) . For normal mutation rates of per cell division , the mean time until emergence of a two-mutation cell is stem cell generations , which is very long even with a short stem cell generation time . Because homeostasis is likely imperfect , we also considered a stochastically fluctuating stem cell population size . We found that , without dedifferentiation , the distributions of times until fixation are very similar for models with and without fluctuations in the stem cell population size , as long as we condition on non-extinction of the stem cell population ( Fig . 3B ) . This is true for a wide range of probabilities of asymmetric division and strengths of mean reversion ( Eq . ( 15 ) ) . This agrees with previous findings that demographic stochasticity does not alter fixation times of neutral mutants in a large population [65] , provided that the carrying capacities of the mutants are the same . Our results suggest that dynamics within either the progenitor or stem cell compartments considered separately do not result in carcinogenesis in the hematopoietic system on a realistic time-scale , provided that cancer-causing mutations occur at normal mutation rates , selection advantages relative to wild-type stem cells do not appear until mutations , and the stem cell population size is constant or varies stochastically around a carrying capacity . We thus turned our attention to coupled model systems in which progenitor cells can dedifferentiate into stem cells . For the coupled system , we first considered stem cell homeostasis caused by strict asymmetric division in the stem cell population , so the stem cell population size remains fixed . To model dedifferentiation in this case , we built off the Moran model and assumed that when a stem cell dies and another enters the population , the new entrant comes from the two-mutation progenitor population with probability equal to times the proportion of two-mutation cells in the progenitor population . Otherwise the new stem cell comes from replication of another stem cell . Roughly speaking , in this model the death of a stem cell leaves a opening in the niche , which can potentially be filled by a dedifferentiated progenitor cell . The number of progenitor cells which can successfully dedifferentiate is controlled by the number of niche openings ( stem cell deaths ) , not by the absolute number of progenitor cells . Typical simulation results are shown in Fig . 4A . We found that dedifferentiation dramatically shortens the time to fixation of two-mutation cells ( Fig . 4B ) . For small dedifferentiation rates , we also saw good agreement between our simulations and a semi-analytical approximation for the time to fixation of two-mutation cells with selective advantage ( Eq . ( 12 ) ) . This agreement suggests that under strict stem cell homeostasis , dedifferentiation is effectively equivalent to a growth advantage for mutant stem cells . Distributions of times to fixation of two-mutation stem cells are plotted as a function of both dedifferentiation rate and mutation rate in Fig . 4C . Dedifferentiation had two major effects in this model: increasing the probability that an emergent two-mutation stem cell would fix and reducing the time between emergence and fixation . Both of these effects act only after a two-mutation cell has been generated in the stem cell population . ( Recall that , as shown in Fig . 2 , the mutation rate and selective advantage must be unrealistically high for a nontrivial fraction of two-mutation progenitor cells to exist in the absence of underlying two-mutation stem cells . ) For all mutation rates , the distribution of times to fixation was roughly constant for dedifferentiation rates , consistent with population genetics theory that selection is only effective when the selection coefficient is greater than the reciprocal of the effective population size . For small mutation rates , increasing beyond this threshold only marginally shortened the total time to fixation . This is because in this case the total time to fixation is dominated by the time for a successful two-mutation cell to emerge , and dedifferentiation only reduces this time by a factor of ( Eq . ( 9 ) ) , where is the probability of a emergent two-mutation stem cell fixing . Under neutrality , so for our model with , dedifferentiation can shorten the time to emergence by at most a factor of 10 . The dedifferentiation rate needed to significantly change this waiting time scales linearly with ( Fig . S4C ) . Hence , for larger stem cell population sizes , a small dedifferentiation rate would have a larger effect . For high mutation rates , the effect of dedifferentiation is more dramatic , because the time from emergence to fixation of two-mutation cells , which dedifferentiation also shortens , is comparable to the time to emergence ( Fig . 4D ) . The model considered in Fig . 4 assumes that only two-mutation progenitor cells can dedifferentiate . We also considered an alternate model in which any progenitor cell can dedifferentiate ( Text S1 ) . In this alternate model , dedifferentiation again had little effect for . Past that threshold the effect was substantial , because in this model dedifferentiation speeds up the time to emergence of two-mutation cells , because one-mutation cells fix much more quickly when they too can dedifferentiate ( Fig . S4D ) . In addition , we considered the case in which the dedifferentiation rate is additionally weighted by the progenitor proliferation rate , and our results did not change qualitatively ( Text S1 , Fig . S4B ) . Our analytical and numerical results suggest that , with intact homeostasis in the stem cell population and normal mutation rates , dedifferentiation plays a fairly minor role in speeding up the time to cancer initiation . We thus turned to consider the case in which homeostasis is not strict . In the previous section , we assumed that the stem cell population size was constant because homeostasis was maintained by all divisions being strictly asymmetric . Consequently , dedifferentiated progenitor cells could only occupy newly created openings in the stem-cell niche created by a death event in the stem cell population . Because homeostasis is likely maintained at the population level [66] , with each stem cell division producing not strictly one stem cell but rather on average one stem cell , we next considered a model in which the stem cell population could stochastically fluctuate around a carrying capacity . In this model , stem cell homeostasis was maintained by dynamically altering the probabilities of the three possible outcomes of a stem cell division: two stem cells , one stem and one progenitor cell , or two progenitor cells ( Eq . ( 15 ) ) . Two-mutation progenitor cells each had a probability per unit time of dedifferentiating , and dedifferentiated cells were simply added to the stem cell pool . Thus in this model the total influx of dedifferentiated cells depended on the total number of two-mutation progenitor cells , not on the creation of openings in the stem cell niche . ( Note that , in our previous model with constant stem cell population size , the rate of dedifferentiation per reproduction event was denoted . To distinguish the present model , we denoted the progenitor dedifferentiation rate per cell per unit time as . ) Again , we asked whether dedifferentiation substantially speeds the time to carcinogenesis . Fig . 5A and 5B show typical results from this model for a moderate dedifferentiation rate . After a waiting time , the population of stems cells began to grow exponentially , because the influx of dedifferentiated two-mutation progenitor cells exceeded the capacity of stem-cell division homeostasis . For larger dedifferentiation rates , the exponential growth rate is larger ( Fig . 5C and 5D ) , and the distribution of progenitor ages can be distorted , with many young cells , as seen in Fig . 5E and 5F . Exponential growth eventually occurs whenever the dedifferentiation rate exceeds a threshold . Solving self-consistently for the influx of dedifferentiated cells and the growth rates of the stem and progenitor cell populations , we obtained an integral equation for the growth ( 18 ) which provides an excellent fit to the numerical simulations ( Fig . 5 and 6A , B ) . ( For derivation details , see Text S1 . ) Setting this growth rate to zero , we found ( 19 ) Here is probability of asymmetric stem cell division ( producing one stem and one progenitor cell ) , and is the mean time between stem cell divisions . ( Note that if , this model reduces to the Moran model with the population size monotonically increasing due to dedifferentiation . ) Lastly , in Eq . ( 19 ) is the average number of progenitor offspring produced by a two-mutation stem cell . Because changes as the system attempts to maintain stem-cell homeostasis , is actually a stochastic variable that depends on the stem cell population size . During exponential growth , because the probability of symmetric divisions that give rise to two stem cells goes to zero , and all new stem cell growth comes from dedifferentiated progenitor cells . In Eq . ( 19 ) , is the growth rate of two-mutation progenitor cells as a function of age , so is the number of progenitors produced by one two-mutation stem cell . Increasing the amplification of mutant stem cells into progenitors increases the net dedifferentiation rate , lowering the threshold . Because the threshold depends on the age distribution of the two-mutation cells , for a given ( small ) rate of dedifferentiation , evolving a mutant that proliferates faster ( increasing ) can destabilize a system in which the number of cancerous cells is stable and take it into exponential growth regime . The dependence of the critical dedifferentiation rate on the growth-rate advantage of two-mutation progenitor cells and probability of asymmetric cell division is shown in Fig . 6B . The critical decreases rapidly as the selective advantage of two-mutation cells increases . Increasing or also lowers the critical dedifferentiation rate , because homeostasis is less effective when asymmetric stem cell divisions are less frequent . Note that the exponential growth rate does not depend on the mutation rate ( Fig . S5A ) , and although the critical given by ( 19 ) needed for exponential growth is a function of the probability of asymmetric division , the actual growth rate and the time to exponential growth are not significantly affected by changing ( see Fig . S5B ) . For dedifferentiation rates below , two-mutation stem cells eventually fix in the population , but for , the stem cell population is likely to begin exponential growth before fixation of two-mutation stem cells . Thus in Fig . 6C and 6D we report the time to carcinogenesis as the time for the two-mutation stem cell population to exceed , the nominal carrying capacity of the stem cell compartment . In this case of stochastic stem cell homeostasis , dedifferentiation can dramatically shorten the time to carcinogenesis , even for low mutation rates . This is because the first two-mutation stem cell often arises not from direct mutation of a stem cell , but rather from dedifferentiation of a progenitor cell generated by mutations within the progenitor compartment ( Fig . 6E ) . Although mutations in the progenitor compartment do not affect a large fraction of progenitors , because the number of progenitor cells is so large , the absolute number of two-mutation progenitor cells is non-negligible . Thus even small rates of dedifferentiation can have dramatic effects . This is in contrast to the case of strict stem cell homeostasis , in which the absolute number of two-mutation progenitor cells was unimportant , because they needed an opening in the stem cell niche to successfully dedifferentiate . Our results show that the case of stochastically controlled stem cell homeostasis is qualitatively different from the case of strict homeostasis . If homeostasis is controlled at the population level ( where stem cell decisions between symmetric and asymmetric division are stochastic ) , dedifferentiation can overwhelm it , leading to exponential growth of the stem cell population . Moreover , if dedifferentiated cells do not depend on openings to colonize the stem cell niche , dedifferentiation can dramatically hasten the time to carcinogenesis , even for low mutation rates .
Progression to cancer is associated with expansion of the cancer stem cell ( CSC ) population , but the origin of these CSCs remains unclear . Although CSCs may arise directly from adult stem cells , they may also arise from somewhat differentiated cells that have dedifferentiated and acquired stem cell-like characteristics [13] , [14] , [18] , [19] , [67] . Stems cells replicate indefinitely , giving them a long time to accumulate the mutations that drive carcinogenesis , but the population of actively dividing stems cells ( ) is small . Progenitor cells replicate only a small number of times , but the population of progenitor cells is typically several orders of magnitude larger than the stem cell population . Thus , as a population , progenitors undergo many more divisions , potentially letting some of these cells acquire mutations that enable them to dedifferentiate and drive carcinogenesis . Here , using mathematical modeling , we have shown that even a small rate of dedifferentiation may drastically shorten the time to cancer emergence , even for low mutation rates . Recent studies suggest stem cell dynamics during homeostasis are governed by neutral competition and genetic drift [10] , [29] , [30] . Traditionally , stem cells were thought to always undergo asymmetric division , always yielding a stem cell and a progenitor cell , resulting in a fixed stem cell population size . This scenario is represented by our first model for stem cell dynamics , based on the popular Moran model . It has been recently shown , however , that symmetric divisions also occur in adult stem cells and may be the predominant form of division [68] , [69] . Moreover , cancer stem cells have been shown to undergo more symmetric divisions than normal stem cells [70] . Little is known , however , about how the stem cell population size is regulated [29] . Hence , in our second model for stem cell dynamics , we made the simplifying assumption of an a priori carrying capacity . We considered a density-dependent stochastic process , in which the degree of mean reversion is controlled through the probabilities of producing zero , one , or two stem cell offspring . In this model , the non-constant stem cell population size tends to return to the carrying capacity , because the mean number of stem cells produced per division is greater than one when and less than one when . ( Although the stem cell population size could , in principle , be maintained by regulating apoptosis rather than biasing division , previous modeling suggests that regulating division probabilities rather than cell cycle time or removal is more important for maintaining homeostasis [46] , [71] . ) If the stem cell population size varies and is regulated by biasing division , we found two distinct regimes . If the dedifferentiation rate is much less than a critical value , then the initial two-mutation stem cell often arises from a normal stem cell , so the time to fixation of such a cell is similar to the case with constant population size . If the dedifferentiation rate exceeds the critical value , however , then the initial two-mutation stem cell often arises from a dedifferentiated progenitor cell , so the time to fixation is dramatically shorter than the case with constant population size . Moreover , in this regime the stem cell population eventually grows exponentially , as dedifferentiating progenitor cells overwhelm stem cell homeostasis . Note that the threshold between these two regimes is independent of the overall mutation rate , if stem and progenitor cell mutation rates are proportional . When the stem cell population size is constant , dedifferentiation simply acts like a selective advantage for mutant stem cells . When the stem cell population size is allowed to vary , however , dedifferentiation can additionally drive exponential growth of the stem cell population . If the stem cell population size is constant , our results imply that stem cell dynamics in the coupled stem cell-progenitor system can be approximated by a population genetics model of the stem cells alone , as long as that model includes positive selection . In this case , we found that the dedifferentiation rate must exceed to substantially shorten the time to cancer acquisition , similar to classical population genetics results that the selection coefficient must exceed the inverse population size to be effective . Hence , our model predicts that in tissues where the niche contains fewer cells , smaller rates of dedifferentiation are sufficient to influence the time to cancer . It is interesting to note that cancers where dedifferentiation has been shown to occur have a small niche size ( i . e . intestinal crypts ) [15] , [16] . For the hematopoietic system , based on the available literature , we assumed that the number of actively dividing stem cells is , so the dedifferentiation rate must be or higher to significantly shorten the time to cancer . Here we focus on the hematopoietic system , in which the stem cell compartment consists of active cells , and two mutations are necessary for carcinogenesis . For some other cancers , such as colon cancer , the number of stem cells per compartment is much smaller , there are many compartments , and the number of necessary mutations is larger . For high mutation rate , the mean time to fixation scales linearly with ( see Fig . S4C ) . So in cancers with small two-mutation stem cells will fix much faster . However , the need to accumulate more mutations will slow carcinogenesis . We expect , however , that the qualitative effects of dedifferentiation will be similar to the hematopoietic system we analyzed . The fact that mutants take a long time to reach an appreciable fraction of the stem cell population is not typically considered in the cancer modeling literature , which often makes an implicit assumption that a newly emerged mutant cell will not go extinct and will fix quickly . Our results show that , for high mutation rate , the time for a mutation to fix in the population is comparable to time for a successful mutant to first emerge , in accordance with classical results of Kimura and Ohta [56] . This is especially important if division events are rare and the population size is large . Considering the time to some predetermined diagnosis threshold is similar to considering the time to fixation , because the time between a selected mutation becoming common and fixing is typically short [72] . Hence , elevated mutation rate ( genetic instability ) may not speed up time to carcinogenesis as much as is typically assumed , suggesting that some form of selection ( potentially through dedifferentiation ) is necessary . Most tumors accumulate hundreds of mutations , but the number of necessary “driver” mutations depends on the type of cancer . We considered mutations , because sequencing of acute myeloid leukemia genomes suggest that there are two driver mutations present [54] . Moreover , recent findings on induced pluripotent stem cells also suggest , as loss of both copies of the tumor suppressor protein p53 [73] or the activation of two oncogenes [67] may be necessary for dedifferentiation . Disabling both copies of p53 improves the efficiency of reprogramming to a stem-like state and greatly enhances the production of induced pluripotent stem cells [74] , [75] . The loss of p53 also leads to the emergence of tumor cells bearing functional and molecular similarities to stem cells [22] , [73] . Finally , inactivation of p53 changes the ratio of symmetric to asymmetric division in mammary stem cells , allowing the total stem cell population to escape homeostasis [70] . Our model only considers actively dividing stem cells , which in the human hematopoietic system have been estimated to be roughly 100 [32] out of 11 , 000–22 , 000 total stem cells [59] . A more complete model would consider both the active and quiescent stem cell populations . Transitions between these states may be influenced by the progenitor population size , potentially acting as a negative feedback and regulating the proliferation of cancer stem cells . In our models , cancerous cells take over the stem cell population , but the ratio of cancer progenitor cells to cancer stem cells is fixed by the progenitor growth process . Even when dedifferentiation drives exponential growth of the stem cells , it is their absolute number that increases , not their proportion in the population . This is in concordance with some in vitro studies , which suggest a fixed proportion of CSCs in a tumor [18] . Many theoretical models find that in order to accumulate multiple mutations on a reasonable time scale , the onset of elevated mutation rate ( i . e . , genetic instability ) should be an early event in tumorigenesis ( reviewed in [4] , [7] ) . The importance of genetic instability , however , depends on assumptions about symmetric self-renewal and differentiation of stem and progenitor cells . In particular , mutations that alter stem cell division or make committed progenitors somewhat immortal may also lead to an early onset of cancer , diminishing the impact of genetic instability [45] . Similarly , our results show that different assumptions about how dedifferentiation occurs ( frequency-dependent reproduction versus absolute numbers of dedifferentiating cells ) dramatically alter time to carcinogenesis . A large body of modeling work in this area ( reviewed in section Prior Mathematical Models ) has focused on calculating the time to carcinogenesis under the assumption of constant population size ( not specifying the mechanism of homeostatic regulation ) . We compared the times to multiple mutation acquisition in our constant and variable stem cell population size models and found that without dedifferentiation both models yield similar results . With dedifferentiation , however , we found that the two models differ substantially . We explicitly considered different ways that homeostasis can be maintained in the stem cell population , and showed that these assumptions can lead to very different results . Our results suggest that if homeostasis is controlled through division asymmetry and if de-differentiated cells do not depend on openings to colonize the stem cell niche , then for de-differentiation rate larger than a critical threshold , the cancer stem cell will most likely originate in a progenitor cell that has undergone de-differentiation . This is a prediction of our model that can be experimentally tested using inducible genetic labeling , the same technique that permitted lineage-tracing experiments allowing quantification of symmetric versus asymmetric divisions [29] , [30] . A similar method was previously used to identify oligodendrocyte precursor cells as the tumor cell of origin in glioma [76] . Our model contributes to the existing literature on the trade-offs between symmetric and asymmetric divisions of the stem cell population in stem cell-driven [35] , [51] , [77] . To our knowledge , our model is the first to quantify the effects of dedifferentiation on the time to carcinogenesis . There are a number of important aspects of homeostasis our model does not consider , such as spatial aspects of stem cell position in the niche ( we assume the cells are well-mixed ) or negative feedback to stem cell divisions from the progenitor population ( we assume progenitors only influence the stem cells by dedifferentiating ) . Lander et al . have previously shown that negative feedback control is needed for homeostasis , and that feedback regulating replication probabilities is more effective than feedback regulating cell cycle lengths [71] . In ongoing work , we are investigating what effect including spatial structure and feedback from progeny will have on dedifferentiation times . The effect of spatial structure on mutation acquisition is still not fully resolved . Some groups argue that time to acquire mutations is actually decreased in a spatial model compared to the space-free model [78] . Other groups find that time to multiple mutations is increased when space is considered [79] , [80] . Like other mathematical models , our model suggests that eradication of cancer is dependent on eradication of cancer stem cells [81]–[83] . The potential for progenitor cells to dedifferentiate and repopulate the stem cell compartment , however , may complicate successful treatment . Our work suggests that further progress in understanding initiation and treatment of cancer requires a more detailed understanding dedifferentiation and of stem cell homeostasis . | Recent evidence suggests that , like many normal tissues , many cancers are maintained by a small population of immortal stem cells that divide indefinitely to produce many differentiated cells . Cancer stem cells may come directly from mutation of normal stem cells , but this route demands high mutation rates , because there are few normal stem cells . There are , however , many differentiated cells , and mutations can cause such cells to “dedifferentiate” into a stem-like state . We used mathematical modeling to study the effects of dedifferentiation on the time to cancer onset . We found that the effect of dedifferentiation depends critically on how stem cell numbers are controlled by the body . If homeostasis is very tight ( due to all divisions being asymmetric ) , then dedifferentiation has little effect , but if homeostatic control is looser ( allowing both symmetric and asymmetric divisions ) , then dedifferentiation can dramatically hasten cancer onset and lead to exponential growth of the cancer stem cell population . Our results suggest that dedifferentiation may be a very important factor in cancer and that more study of dedifferentiation and stem cell control is necessary to understand and prevent cancer onset . | [
"Abstract",
"Introduction",
"Models",
"Results",
"Discussion"
]
| [
"mathematics",
"theoretical",
"biology",
"applied",
"mathematics",
"biology",
"evolutionary",
"biology"
]
| 2014 | Effect of Dedifferentiation on Time to Mutation Acquisition in Stem Cell-Driven Cancers |
Precise patterns of spatial and temporal gene expression are central to metazoan complexity and act as a driving force for embryonic development . While there has been substantial progress in dissecting and predicting cis-regulatory activity , our understanding of how information from multiple enhancer elements converge to regulate a gene's expression remains elusive . This is in large part due to the number of different biological processes involved in mediating regulation as well as limited availability of experimental measurements for many of them . Here , we used a Bayesian approach to model diverse experimental regulatory data , leading to accurate predictions of both spatial and temporal aspects of gene expression . We integrated whole-embryo information on transcription factor recruitment to multiple cis-regulatory modules , insulator binding and histone modification status in the vicinity of individual gene loci , at a genome-wide scale during Drosophila development . The model uses Bayesian networks to represent the relation between transcription factor occupancy and enhancer activity in specific tissues and stages . All parameters are optimized in an Expectation Maximization procedure providing a model capable of predicting tissue- and stage-specific activity of new , previously unassayed genes . Performing the optimization with subsets of input data demonstrated that neither enhancer occupancy nor chromatin state alone can explain all gene expression patterns , but taken together allow for accurate predictions of spatio-temporal activity . Model predictions were validated using the expression patterns of more than 600 genes recently made available by the BDGP consortium , demonstrating an average 15-fold enrichment of genes expressed in the predicted tissue over a naïve model . We further validated the model by experimentally testing the expression of 20 predicted target genes of unknown expression , resulting in an accuracy of 95% for temporal predictions and 50% for spatial . While this is , to our knowledge , the first genome-wide approach to predict tissue-specific gene expression in metazoan development , our results suggest that integrative models of this type will become more prevalent in the future .
Gene expression is regulated through the interplay of transcription factors binding to cis-regulatory modules ( CRMs ) , chromatin modifications and the basal transcriptional machinery recruited to promoter elements . CRMs function as discrete regulatory elements [1] , [2] , that can act at varying genomic distances from their target genes [3] . Despite recent advances in our understanding of the regulatory steps of transcription , the ability to predict both spatial and temporal aspects of gene expression remains surprisingly limited . Efforts in this direction can be broadly divided into two groups: ( 1 ) Predicting cis-regulatory or enhancer activity , where recent studies in yeast [4] Drosophila [5]–[7] and C . elegans [4] have made substantial progress . In one such study the tissue specificity of the neighboring gene's expression was used to guide the search for specific TF combinations [7] , while in another the combination of sequence motif matches was used to predict gene expression [4] . Although , these are important steps , integrating the activity of multiple cis-regulatory elements that regulate overlapping or distinct aspects of a gene's spatio-temporal expression remains a key challenge ( Fig . 1a , Fig . S1 ) . ( 2 ) Using chromatin state dynamics to predict gene expression [8]–[12] ) , with or without information on transcription factor ( TF ) and insulator data . For example , in Drosophila a logistic regression was used to predict temporal ( not tissue-specific ) gene expression in embryogenesis [11] , showing a performance better than random for 23 . 6% genes , with a 2 . 5 fold enrichment over control experiments where the connectivity between TFs and their targets was reshuffled . In c . elegans an SVM classifier was used for a similar task of discerning highly and lowly expressed transcripts based on measured chromatin marks [13] , although tissue specificity was not examined . This approach , based on transcripts and chromatin marks in their immediate vicinity ( +/−4 kb ) achieves high accuracy ( average AUC for all stages = 0 . 82 ) , reflecting the strong correlation between transcription and chromatin marks on the gene body , such as the H3K79 methylation and Pol II occupancy consistent with the results by Karlic et al [14] . However , while virtually all regulatory elements appear to reside within 5 kb of the transcriptional start site ( TSS ) in C . elegans , this is not the case in other species . In Drosophila , mouse and humans there are many examples of remote CRMs acting at large distances from the TSS [15]–[18] spanning many intervening genes [19] , [20] , where large chromatin loops are thought to bring the enhancers and the target gene's promoter in close physical proximity [21] . In addition , genes , especially developmental regulators , are controlled by multiple CRMs , giving rise to partially overlapping patterns of activity [22] , [23] . In order to capture , and thereby predict , the full spectrum of a gene's spatial expression , two key issues need to be addressed directly: ( 1 ) accurately linking CRMs to their target genes and ( 2 ) integrating the activity from multiple CRM , as is done naturally for most developmental genes in multicellular species . There is currently very little biological information or understanding of how the activity of multiple elements is integrated at the promoter level . While some studies have suggested that each CRM acts in an additive manner so that the gene's expression pattern is the simply sum of all elements , other studies have shown that the gene can be expressed in a broader [8] , [24] ) or more restricted [9] , [10] spatial domain than the sum of its individual regulatory elements . It is therefore currently not clear how best to integrate separate computational models of cis-regulatory elements to accurately reflect this convergence of regulatory information controlling a gene's expression in vivo . These difficulties have limited spatial predictions of gene expression to a small number of very well characterized genes [6] , [25] , or more globally to focusing on predicting on-off states in single cell systems [26] , [27] , thereby circumventing the inherent complexity of spatial expression within a multicellular organism . High-resolution ChIP-chip or ChIP-seq approaches facilitate the mapping of distant regulatory elements based on transcription factor occupancy [26]–[32] , co-factor binding [33] or chromatin marks [34] , providing new possibilities to develop better predictive models of global gene expression patterns . However , there are still several levels of information missing , including a complete catalog of all enhancers active during specific stages of development , information on the identity and timing of the TFs recruited to each enhancer , cell-type specific information on chromatin status , the activity state of the associated target gene and a general lack of information on the physical association of CRMs with promoter elements . Despite this incomplete knowledge , we asked if the current level of information is sufficient to accurately predict spatio-temporal gene expression within the context of a multicellular embryo , reasoning that the predictive power of the model should only improve as more information becomes available .
We developed a probabilistic model , integrating diverse types of data generated from whole embryos and thereby containing mixed signals from many tissues , to predict both spatial and temporal aspects of gene expression , with particular emphasis on the mesoderm and derived muscle types . More specifically , using Drosophila embryogenesis as a model system , we integrated six types of data relevant to transcriptional regulation: ( i ) 19 , 000 TF binding peaks derived from ChIP-chip experiments for mesoderm specific TFs , clustered into 8008 non-overlapping cis-regulatory modules ( ChIP-CRMs ) , ( ii ) spatio-temporal activity data for 343 CRMs from in vivo transgenic reporter assays , ( iii ) the genomic distance of CRMs in relation to transcriptional start sites , ( iv ) 37 , 923 occupancy peaks for 6 insulator binding proteins , ( v ) H3K4me3 enrichment measured for promoter regions of 14689 genes , and ( vi ) spatio-temporal expression of 5 , 995 genes derived from in-situ hybridization ( see Table S1 for a detailed data description ) . Note , as chromatin modifying enzymes for canonical histones and insulator binding proteins are ubiquitously expressed , the whole embryo data from ( iv ) and ( v ) does not contain any inherent cell-type specific ( spatial ) information , and ( v ) represents merged temporal signal over the entire period of embryogenesis , which is 24 hr in Drosophila . TF occupancy ( i ) and gene expression ( vi ) data provide information on potential regulatory input and the final spatio-temporal output , respectively , but little means to connect the two , highlighting the need to integrate diverse layers of information . Previous studies suggest that cis regulatory elements function , to a large extent , independently of each other [35] . Assuming that this is correct , there are two natural levels to model gene expression based on: ( i ) the relationship between TF occupancy and CRM activity and ( ii ) the relationship between models of multiple CRMs' activity and a gene's expression ( Fig . 1a ) . This first step was recently addressed using support vector machine ( SVM ) models , which demonstrated that TF occupancy alone is sufficient to predict spatio-temporal CRM activity during mesoderm development [5] . It was postulated [36] that the same method could in principle be adapted to model gene expression prediction , although this would require linking CRMs to their appropriate target genes and integrating inputs from multiple CRM models to reflect a target gene's expression . Taking advantage of the wealth of data on TF occupancy at mesodermal CRMs [5] , we tested this assumption by building a simple additive model that assigns each CRM to the nearest gene and then sums the SVM prediction scores for all assigned CRMs to obtain a spatio-temporal expression prediction at the gene level . Overall , the predictions were of poor quality ( Fig . S2 ) , indicating that a model based on these simple assumptions does not reflect the biological complexity of the system . Using well-characterized gene loci to examine why the model failed revealed that enhancers do not always regulate the nearest gene , but often a more distant gene ( Fig . 1a twist locus ) or can even act across an intervening inactive gene to reach its appropriate target ( Fig . S1a bagpipe locus ) . Such inactive ‘bystander’ genes [37] can be located within the intron of a target gene ( Fig . S1c Fas3 locus ) or vice versa ( CG6981 ) , further confounding the problem of appropriate target gene assignment . This demonstrates the need to move to a more integrative model that includes information on promoter activity ( H3K4me3 enrichment ) and insulator occupancy within a gene locus . As insulator binding proteins mediate long-range regulatory interactions between enhancers and their target genes [3] , [38] , we reasoned that insulator occupancy could improve the ability to recognize ‘bystander genes’ , while the presence of H3K4me3 at promoters will identify active genes within the vicinity of active CRMs . To deal with this complexity , we applied a Bayesian model to probabilistically integrate diverse types of data in an iterative manner , which has the advantage of being able to cope with uncertainty and incompleteness within each dataset using conditional probabilities . The model consists of three components ( Fig . 1b ) : ( i ) a Bayesian Network ( BN ) that describes the probability of a CRM being active in a tissue or time-point as a function of its occupancy by different TFs , ( ii ) a custom probabilistic model that describes the probability of a gene being expressed at a given stage and tissue depending on the activity of surrounding CRMs , the location of CRMs and insulators relative to the promoter , and the activity state of the promoter ( Fig . 1 ) , and ( iii ) an expectation maximization ( EM ) procedure [39] that functions to find an optimal set of parameters within the overall Bayesian model , iterating between the BN and custom model until convergence . To accurately predict gene expression , the model must be able to cope with dynamic changes in the regulatory context of genes , which determines their activity state at different stages of development and in different tissues . To account for this , we trained the model using spatio-temporal expression information of 5 , 082 non-ubiquitous genes generated from large-scale in-situ hybridization experiments [40] , describing when and where genes are expressed during embryogenesis . As a proof-of-principle we focused on five temporal windows of development and five tissue classes ( 10 prediction classes; Supplementary text S1 . ) . In more detail , the first component , modeling CRM activity as a function of TF binding events , was achieved using a BN , allowing for accurate representation of conditional probability ( Fig . 2a , described in detail in Supplementary text S1 – in “Layer 1-TF binding” and “Layer 2-CRM activity” ) . The model uses measured TF binding events on CRMs as input ( from ChIP-chip data ) and spatio-temporal CRM activity data as output ( from in vivo transgenic-reporter assays ) ( depicted in Fig . 1a ) . The nodes within the BN are of two types: specific TF binding events ( factor-F at time-point-T , representing 15 variables ) and activity classes ( tissue or time-point , representing 10 variables ) . Each edge between nodes represents the probability of a CRM being active in a given activity class as a function of a particular binding event ( e . g . CRM activity in tissue-A depends on the binding of factor-F at time-point-T ) . The correct topology of connections was reconstructed using the Bayesian Dirichlet equivalence score as implemented in the BNfinder software [41] . Once the most likely topology was known , the conditional probabilities of CRM activity in different classes ( temporal and spatial ) were calculated from the training data using the maximum likelihood principle . The trained BN and the conditional probability distributions were then used to provide probability estimates for the spatio-temporal activity of all 8008 CRMs , not only the 147 used in the training dataset . Based on these probability estimates , we compared the BN model with the previously published SVM approach [5] . Overall , our model gives slightly better predictions of previously unseen CRM activity ( Fig . S3b ) , even though it was not explicitly optimizing the accuracy at the CRM level . In addition , unlike ‘black box’ type models such as SVMs , the learned BN network topology provides interpretable insights into the most important TF binding events for each spatio-temporal activity . For example , the BN revealed that Biniou ( a FoxF TF ) enhancer occupancy is the key predictive signal for visceral muscle activity ( Fig . 2a ) , which matches the known essential role of this TFs for visceral muscle development from genetic studies [2] . The second component of the Bayesian model addresses how genes integrate probabilistic signals from one or many CRMs by relating this information to known gene expression patterns within the training set ( described in detail in Supplementary text S1 – “Layer 3-gene activity” ) . For each gene , we consider the location of its transcriptional start site ( TSS ) and the CRMs present in its broad environment ( +/−100 kbp , where there is one gene per ∼8 kb in the Drosophila genome ) . As the majority of known Drosophila enhancers are located within +/−20 kb of their target gene's promoter , the probability of activation decreases linearly with respect to the distance from the TSS . The only parameter that the model fits is the maximal distance between a CRM and the TSS within a +/−100 kbp window . To facilitate linking CRMs to their appropriate target gene , the model integrates information on the occupancy of six insulator binding proteins [42] relative to the location of CRMs and surrounding genes ( Fig . 1a ) . As insulator proteins can block enhancers from inappropriately activating nearby promoters [43] , CRM-promoter interactions are considered blocked if they operate across an insulator boundary ( see Methods ) . To obtain a probability for a promoter being in an active or inactive state , we used the presence of H3K4-trimethylation ChIP-seq signal at promoters as an indicator of promoter activity [44] ( Fig . 1a , Fig . S1 , methods ) . The model requires both an active promoter and at least one active CRM to activate a gene in a given spatio-temporal context . The classifier accuracy was determined using the area under a receiver-operator curve ( AUC ) for varying posterior probabilities of gene activation . To train the model for tissue specific or developmental stage specific chromatin context , we used in-situ hybridization data of 5082 genes [40] to identify genes in specific spatio-temporal classes . For simplicity , expression patterns were divided into a number of binary classes: focusing on 5 tissue classes ( mesoderm , somatic muscle , visceral muscle , mesoderm+somatic muscle and visceral+somatic muscle ) and 5 time-windows ( stages 4–6 , 7–8 , 9–10 , 11–12 , 13–16 , spanning ∼85% of embryonic development ) . Separate variables were incorporated for a gene or CRM activity in each class , allowing each class to be evaluated individually using the probability of a gene to be expressed in a particular spatial or temporal domain . The coupling between the two mentioned components of the model is through the intermediate layer representing the activity of the CRMs ( Supplementary text S1 “Integrating the different layers of the model using iterative optimization” ) . Since the activity of the vast majority of the ChIP-defined CRMs is unknown , the variables in the intermediate layer are latent . Under this setting , an iterative Expectation maximization ( EM ) [39] procedure was used to facilitate using data of varying degrees of completeness at different levels of the model . The TF binding data is very extensive for all 8008 CRMs ( at least within the scope of the five TFs ) , as is the insulator occupancy and promoter activity data , although the later two represent merged signals from mixed tissue types and have very low temporal resolution . Spatio-temporal expression data is available for a substantial number of genes ( ∼33% of predicted Drosophila genes ) , which contrasts with the scarcity of knowledge on CRM activity , which is available for only ∼4% of CRMs . This level of CRM activity data is sufficient to train a predictive model of CRM activity , using a BN ( Fig . 2a ) or SVM [5] approach . However , there is not a single gene in the Drosophila genome where the activity of all ChIP-defined CRMs in its vicinity are known . As such , there are no complete examples that could be used to fit a model representing convergence of multiple CRM activities to a single gene's expression . To address this , the activity of CRMs was consistently treated as a hidden variable in the model , and the CRM activity information was only used for model initialization . EM was used to iteratively improve both the CRM activity predictions and gene expression predictions ( see methods and Fig . 1b ) , resulting in an effective model with local maximal likelihood . By performing the EM procedure in a 10-fold cross-validation framework , we assessed the ability of the model to predict gene expression for genes not used for training . The average AUC value for all 10 prediction classes exceeds 0 . 8 ( Fig . S4 ) , a significant improvement over the simple additive SVM method ( p-value<10−7; Fig . S3a ) . Importantly , the cross-validation estimated performance is comparable to that of the model trained on the full dataset ( Fig . S5 ) , indicating that the model is not over-fitted . The difference in AUC slightly underestimates the improvement of the model as it is based on predictions made for all genes , while only a minority of Drosophila genes are expected to be specifically expressed in each activity class and the majority of genes are correctly predicted not to be regulated by mesodermal CRMs . For example , from all 5082 Drosophila genes with characterized non-ubiquitous expression , only 137 have annotated expression in the activity class somatic muscle , 135 in mesoderm and 60 in VM [40] . Extrapolating these numbers to the entire genome estimates that the percentage of genes expressed in each activity class is in the range of 1–2% , excluding ubiquitously expressed genes . With this in mind , we examined the top 2% of predictions from the trained Bayesian model , which identified on average a 15-fold enrichment in gene expression in the predicted tissue compared to a random classifier , for all activity classes , with the best class having a 45-fold enrichments ( Fig . 2b ) . To investigate the most important aspects of the model's predictions , we compared the results to simpler approaches that do not use either chromatin state ( insulator binding data or H3K4me3 ) or an EM procedure , all of which obtained inferior results ( Fig . 2b , Fig . S6 ) . Adding H3K4me3 promoter activity signal to TF binding , for example , reduces the number of false-positive predictions by 1 . 5 fold , thereby increasing the enrichment of correct predictions ( Fig . 2b ) . The method also demonstrates improved performance over a simpler two-layer model predicting gene expression directly from ChIP peaks , skipping the intermediate CRM layer [45] ( , Methods ) . Although this 2-layer model is not accurate enough to make reliable predictions , the approach can be very valuable for initiation of the EM algorithm in cases where there is no CRM activity database available . In many organisms obtaining information on CRM activity for a large number of regulator elements is difficult . We therefore tested whether our approach could provide comparable results without providing the measured activity of selected CRMs . To avoid random fluctuations we have used the gene expression data for genes with very closely ( <500 bp ) associated CRMs as a proxy for enhancer activity . While this is certainly introducing some erroneous information by both erroneous target assignment and by assigning total gene activity to only one selected enhancer , it seems to give only slightly worse results for classes with multiple genes associated to it ( VM , SM , MESO , see Fig . S10 ) . To validate the true performance of the model we took advantage of spatio-temporal expression data for more than 600 genes not included in our training set that was part of the third release of the Berkeley Drosophila Genome Project ( BDGP ) in-situ database [40] . We used models trained on the whole training dataset and assessed their performance on the genes present only in the new dataset by calculating the AUC for each activity class ( Fig . S8 ) . The performance was comparable to the cross-validated-based estimates , with the average AUC of 0 . 78 ( compared to 0 . 82; Fig . 2c ) . To further validate the quality of the trained model , we chose a tissue with a relatively restricted spatial expression , the visceral muscle ( Fig . 3a , AUC 0 . 87 ) , and manually curated the top 100 genes predicted to be expressed in this tissue ( VM ) . Examining the literature and BDGP , we identified spatio-temporal expression for 46 of the 100 genes , 67% of which are expressed in visceral muscle , while the expression of the remaining 33% did not fit with our prediction ( Dataset S8 ) . We randomly selected 22 genes for which there was either no expression data available , or were apparent prediction errors from the model ( within the 33% ) . Double fluorescent in-situ hybridization using a visceral muscle specific marker revealed that the timing of expression of 21 out of 22 genes match their temporal prediction ( 95% ) , while the expression of 50% match their spatial prediction ( Fig . 3c , Fig . S9 ) , representing a 42-fold enrichment in gene expression in visceral muscle compared to the 1 . 2% of genes annotated in the BDGP database ( Fig . 3b ) . The high success rate of the model , despite the presence of inaccurate expression annotations within the training dataset , demonstrates the general robustness of this iterative approach .
This study represents a first attempt to build an integrative probabilistic genome-wide model that predicts both the spatial and temporal aspects of gene expression , within the context of a multicellular embryo . The Bayesian model integrates diverse types of genomics data , including transcription factor occupancy , chromatin modification and insulator binding information , using in vivo CRM activity information and gene expression data to train the model . In addition to predicting gene expression , introspection of the model's parameters reveals a number of additional insights . First , the iterative trained Bayesian network improved the accuracy of the previously published SVM approach for CRM and gene activity prediction [5] ( Fig . S3 ) , and recovered , without any prior information , the known dependencies between specific TFs and respective tissues . Second , through expectation-maximization , the model learns the optimal distance of a CRM to its target gene . This revealed extensive long-range enhancer activity , which may be much more widespread in Drosophila than previously anticipated . Although there are a handful of known enhancers acting >30 kb from their target gene [19] , [20] , the majority of CRMs identified to date are <+/−20 Kb of their target gene . This apparent close proximity , however , most likely reflects the current biases in how CRMs are identified in single gene studies , starting from the gene moving out , or in global studies where CRMs are typically linked to the closest proximal gene . The iterative Bayesian model revealed that CRMs as far as 50 kbp from the transcriptional start site have important contributions to accurately predict a gene's activity . Third , the model suggests that enhancer sharing between genes may be an inherent property of developmental enhancers where a CRM contributes to the predictions of 3 genes on-average . This observation , which came directly from the trained model , has recently been observed using an experimental technique to link CRMs to genes ( chromatin conformation capture ) [12] , and has exciting implications for how transcriptional networks are regulated during development . Taken together , this approach provides a method to move from descriptive ‘omics’ type data to predictive models of gene expression . Given the exponential increase in measurements of chromatin state and TF occupancy in the coming years , we expect this type of iterative analyses to become increasingly useful as a complement to ongoing attempts to map global gene expression patterns by experimental approaches and as a tool to uncover novel properties of transcriptional regulation .
CRM occupancy data and CAD database were used as published by Zinzen et al . [5] . Gene expression patterns were obtained from the BDGP in-situ hybridization database [46] - release 2 served as the training data , while release 3 ( beta release downloaded on May 27th 2010 ) was used as the testing dataset . Only genes with tissue specific expression ( excluding ubiquitous and maternal expression ) were analyzed . Anatomical terms from BDGP were grouped into more general classes ( mesoderm , somatic muscle , visceral mesoderm ) , similarly to the procedure used by Zinzen et al . [5] . Temporal classes were based on the staged groups used by the BDGP in their annotations ( st . 4–6 , 7–8 , 9–10 , 11–12 , 13–16 ) . Whole embryo ChIP-seq data of histone H3K4 tri-methylation was from ModEncode [47] for three time-points: 4–8 h , 8–12 h and 12–16 h ( ModEncode sample IDs:790 , 791 , 792 ) . Averaged processed signal was calculated for a region surrounding all transcriptional start sites ( −100 , +400 bp from TSS ) and then discretized into low and high values ( threshold 0 . 3 ) for training the Bayesian network . Whole embryo ChIP-chip data for the six insulator proteins was obtained from Negre et al . [42] ( using a 1% FDR ) . The Bayesian model is composed of three main layers of different nature . The first layer represents variables corresponding to Transcription Factor binding to CRMs; second represents the CRMs activity under different conditions and the third is concerned with gene activity under the same set of conditions . We made the assumption that the only True causal connections are either coming from the first to second layer ( TF binding causing CRM activity ) or from the second to third layer ( CRM to gene activity ) . No direct dependencies from first to third layer are allowed . A Bayesian Network was used to model the dependency between binding and CRM activity , while gene expression was assumed to be independently initiated by any active CRM within an acceptable range . Bipartite Bayesian network was used to describe dependencies between TF binding and CRM activity . For each CRM , a quantitative binding score was computed for each of 15 TF/stage combinations ( as previously described [48] ) representing quantitative measurements for actual binary events of TF binding ( Dataset S1 ) . Each expression class ( temporal or spatial ) was represented by a separate binary variable . There were 5 temporal classes , representing stages 4–6 , 7–8 , 9–10 , 11–12 and 13–16 , following the BDGP nomenclature and 5 tissue-specific classes mesoderm ( MESO ) , somatic muscle ( SM ) , visceral muscle ( VM ) , mesoderm and SM ( MESO+SM ) , somatic and visceral muscle ( SM+VM ) . Edges in the network represent dependencies of the conditional probability function of the variable corresponding to the CRM being active in a given condition on any variables representing TF binding events . Measured binding and activity for each CRM were considered to be a single observation from the same underlying joint distribution and they were used to find an optimal network . The network structure was constrained to only contain edges of this kind and probability distributions were optimized using BNfinder [24] software using Bayesian Dirichlet equivalence ( BDe ) score . No constraints on the resulting cpd function were set , however the binding signal was converted by the BNfinder software to probabilistic readouts of binary variable using a Gaussian mixture model . For detailed parameters used see Supplementary Text S1 . All distances between a CRM and a transcriptional start site of a gene that were lower than 100 kb were tabulated based on FlyBase genome annotations , version 5 . 17 [49] ( Dataset S2 ) . For each CRM-promoter pair , the total number of insulator peaks was calculated in between them . Each gene is assumed to be able to respond to the activation signal from any of the paired CRMs , depending on the distance and the number of insulator peaks between them . It is assumed that the probability of activation by a CRM over a given distance d is linearly decreasing with d until it reaches 0 at the distance dmax or when the predefined limit of insulator peaks have been exceeded . Each promoter is assigned a probability of being activated in development based on the histone modification ( H3K4me3 ) level measured within the 500 bp around the TSS , using non tissue-specific data ( Dataset S4 ) . For details see Supplementary text S1 . The majority of CRMs ( >95% ) have unknown activity , so we treat all variables corresponding to CRM activity as latent and use a maximum likelihood principle to estimate them . We define the likelihood function Lwhere G represents gene activity , i indexes CRMs , j indexes genes , Ai represents activity of the i-th CRM , Wij represents the weight of CRM-promoter interaction ( depending on distance and insulators , as described earlier ) and Rj representing the probability of a given promoter responding under specific conditions . Given this likelihood function we aim to find the most likely parameters of the model , i . e . the Bayesian Network and the optimal dmax . We use the Expectation Maximization ( EM ) strategy , by iteratively improving our current estimate of the parameters . Since the EM is a local optimization strategy , the result is highly dependent on the starting model . Normally this can be solved by starting from multiple randomized models , however in our case properly sampling a space of all Bayesian Networks would be difficult and likely to produce non-biological models . Instead we begin by initializing the BN parameters based on known CRM activity data ( CAD [5] ) by making the first inference not on the full training dataset but on the subset of the training set with experimentally measured tissue-specific activity . The dmax parameter could also have a strong impact in the initial stage of EM if it is set too low and therefore excludes some CRM data from the whole learning process . For this reason we initially set dmax to the maximum possible value and allow it to change freely from then on . The EM procedure is composed of alternating iteration of the expectation ( E-step ) and maximization ( M-step ) steps until convergence ( improvement in the likelihood below 2% ) . In the E-step , we calculate estimated probability of each CRM being active in each condition based on our current model parameters ( BN and dmax ) . Since the model has three layers and we are interested in the estimation of the hidden variables from the middle one , we use an approach based on the forward-backward algorithm frequently used to infer the probabilities of the hidden variables in Hidden Markov Models [7] . In our case , the forward probability is the probability of the CRM being active given the TF binding data , and can be easily computed using the BN for all CRMs . The backward probability is the probability of the CRM being active given the gene expression data . We can ignore all genes j such that wij = 0 as the change of the i-th CRM activity will not affect the total likelihood . For each CRM we need to consider all genes with wij>0 and the CRM can only be inactive if each of the genes in its range is turned on by another CRM , which , by assumption of independent action of CRMs , can be computed using Bayes theorem and total probability . The overall activity of each CRM is determined by a smoothing step as the product of the respective forward and backward probabilities . In the M-step , current estimates of CRM activity ( the latent variables ) are used for finding the model parameters ( BN and dmax ) maximizing the likelihood function L . For the BN , we are using the Bayesian Dirichlet equivalence optimization implemented in the BNfinder [24] library . Due to the constrained structure of the BN , it is possible to find a globally optimal network representing observed combinations of binding and activity very efficiently . As the likelihood function is not monotonous with respect to dmax we employed an exhaustive strategy to find the optimal dmax giving the maximum likelihood under assumed CRM activity and gene expression . This can be done quite efficiently with a step size of 200 bp , equal to the minimal size of the CRM . The process was repeated until convergence; in the tested cases ∼10 iterations were required to reach improvement in one step below 2% . For a more detailed description see Supplementary text S1 . For each expression class ( temporal or spatial ) the posterior probability calculated from the model was used as the ranking criteria to calculate the area under the curve ( AUC ) for the receiver operator characteristic ( ROC ) curve . The AUC value can be interpreted as the probability of a random positive example to have a higher posterior probability of expression than a random negative example . To assess the significance of the achieved AUC measures in comparison to random classifier or in comparison between different models we used the procedure proposed by Hanley and McNeil [50] . To avoid over-fitting , all models were trained in a 10-fold cross-validation scheme based on BDGP gene expression database release 2 . Then the entire BDGP release 2 dataset was used for training the final models ( Dataset S5 ) , which were then tested on the gene expression patterns from BDGP release 3 ( Dataset S6 ) , excluding those from the training set . The same models were used to select genes from the visceral muscle activity class for validation by in-situ hybridization experiments . All training sets are available in Dataset S7 . The EM algorithm was implemented in Python using the BNfinder [24] library for estimating Bayesian networks , ROC curves were plotted with ROCR [51] package for R . All the scripts are available at https://code . launchpad . net/bnfinder/GEpredict In-situ hybridizations in Drosophila embryos were carried out using standard protocols as described previously [52] . The following ESTs from the Drosophila Gene Collection ( DGC ) were used to generate Digoxigenin-labeled probes: GM02640 ( Eip75B ) , LD09907 ( Hex-A ) , RE05370 ( CG9194 ) , GM10074 ( bt ) , AT24194 ( Rya-r44f ) , LP05734 ( Hsp22 ) , GH06348 ( CG1516 ) , RH17388 ( CG10654 ) , GH24653 ( A3-3 ) , SD01953 ( by ) , LP03829 ( CG6981 ) , GH27027 ( Ncc69 ) , SD11716 ( CG14709 ) , HL01392 ( fau ) , LP06027 ( Cpr78E ) , GH06222 ( CG13124 ) , LD02379 ( nrv1 ) , RE70568 ( Lim3 ) , LD44720 ( CG7530 ) , GH23506 ( CG14655 ) , LP04481 ( CG6770 ) , GH19382 ( CG4945 ) . biniou cDNAs ( from M . Frasch ) was used to generate Biotin-labeled probe . Double in-situs hybridizations were performed by using anti-Digoxigenin-POD and anti-Biotin-POD antibodies ( Roche ) and detected sequentially with FITC and Cy3 ( Perkin-Elmer TSA kit ) . A Zeiss LSM 510 confocal microscope was used for imaging . | Development is a complex process in which a single cell gives rise to a multi-cellular organism comprised of diverse cell types and well-organized tissues . This transformation requires tightly coordinated expression , both spatially and temporally , of hundreds to thousands of genes specific to any given tissue . To orchestrate these patterns , gene expression is regulated at multiple steps , from TF binding to cis-regulatory modules , general transcription factor and RNA polymerase II recruitment to promoters , chromatin remodeling , and three-dimensional looping interactions . Despite this level of complexity , the regulation of gene expression is typically modeled in the context of transcription factor binding and a single enhancer's activity as this is where the majority of experimental data is available . Recent advances in the measurement of chromatin modifications and insulator binding during embryogenesis provide new datasets that can be used for modeling gene expression . Here we use a Bayesian approach to integrate all three levels of information to combine the activity of multiple regulatory elements into a single model of a gene's expression , implementing an expectation maximization strategy to overcome the problem of missing data . Importantly , while the data for histone modifications and insulator binding represents merged signals from all cells in the embryo , the model can extract cell type specific and stage-specific predictions on gene expression for hundreds of genes of unknown expression . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"systems",
"biology",
"genome",
"expression",
"analysis",
"developmental",
"biology",
"genomics",
"functional",
"genomics",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics",
"cell",
"fate",
"determination"
]
| 2012 | Predicting Spatial and Temporal Gene Expression Using an Integrative Model of Transcription Factor Occupancy and Chromatin State |
The use of high quality disease surveillance data has become increasingly important for public health action against new threats . In response , countries have developed a wide range of disease surveillance systems enabled by technological advancements . The heterogeneity and complexity of country data systems have caused a growing need for international organizations such as the World Health Organization ( WHO ) to coordinate the standardization , integration , and dissemination of country disease data at the global level for research and policy . The availability and consistency of currently available disease surveillance data at the global level are unclear . We investigated this for dengue surveillance data provided online by the WHO . We extracted all dengue surveillance data provided online by WHO Headquarters and Regional Offices ( RO’s ) . We assessed the availability and consistency of these data by comparing indicators within and between sources . We also assessed the consistency of dengue data provided online by two example countries ( Brazil and Indonesia ) . Data were available from WHO for 100 countries since 1955 representing a total of 23 million dengue cases and 82 thousand deaths ever reported to WHO . The availability of data on DengueNet and some RO’s declined dramatically after 2005 . Consistency was lacking between sources ( 84% across all indicators representing a discrepancy of almost half a million cases ) . Within sources , data at high spatial resolution were often incomplete . The decline of publicly available , integrated dengue surveillance data at the global level will limit opportunities for research , policy , and advocacy . A new financial and operational framework will be necessary for innovation and for the continued availability of integrated country disease data at the global level .
Threats to public health around the world have become increasingly complex and the importance of high quality disease surveillance for preparedness and disease control will continue to grow [1] . Scientific progress and global cooperation against emerging threats will depend on the availability and sharing of disease surveillance data between countries . Global health and funding agencies emphasized this in an appeal for greater availability and use of data for global health [2 , 3] . Formally , the 2005 International Health Regulations require the use and sharing of data in response to new threats [4 , 5] . The central role of the World Health Organization ( WHO ) in global disease surveillance and data dissemination has been stated in World Health Assembly resolutions for specific diseases [6] . The WHO has developed various data systems to integrate and disseminate country surveillance data such as the Global Health Observatory [7] , the Global TB Database [8] , DengueNet [9] , RabNet [10] and FluNet [11] . In addition to these global databases , WHO Regional Offices ( RO’s ) also provide disease surveillance data through their websites to inform member countries on disease patterns and trends in their region . Increasingly , country Ministries of Health post their own disease surveillance data online for their constituency , mostly in the form of epidemiological bulletins but sometimes using sophisticated online data repositories such as those developed by Brazil and Indonesia [12 , 13] . The public availability of disease surveillance data from various heterogeneous sources provides new opportunities for research , training , and policy making but can also lead to confusion on data discrepancies between sources . Limited information on methodology used at various steps along the data trail from within countries to the global level has further complicated this data landscape . Although it is generally known that surveillance data reported by different agencies may not be identical due to reporting methods and definitions , few studies have quantified the availability and consistency of publicly available disease surveillance data across sources . This information can guide policy makers , scientists , students , and others to use available data more effectively . We used the example of dengue to assess the availability and consistency of surveillance data provided online by WHO . We also provided examples of online data provided by the Ministry of Health of Brazil and Indonesia .
We extracted all online dengue surveillance data from WHO ( WHO DengueNet [9] and from the websites of the Pan American Health Organization ( PAHO ) [14] , the WHO Southeast Asia Regional Office ( SEARO ) [15] and the WHO Western Pacific Regional Office ( WPRO ) [16 , 17] ) , and by the Ministries of Health ( MOH ) of Brazil and Indonesia [12 , 13] . Brazil and Indonesia were selected as examples because they provided open access to detailed dengue surveillance data online in computer readable format . All available data up to April 12th 2013 were extracted at the highest possible spatiotemporal resolution . To obtain standardized data across these sources , we extracted data for all ages and for both genders combined . We did not extract serotype specific data because these were minimally available . We standardized indicators reported by different sources across spatial and temporal scales and also harmonized country names using the United Nations ISO country name standard ( ISO 3166 ) [18] . We assessed the availability of dengue data from each source and also measured data consistency between and within sources . We defined consistency between sources as the percent agreement of data reported for overlapping countries and time periods . We defined consistency within a source by the percent agreement of indicators that were recomputed by us from data within the source and the corresponding indicators provided by the same source . All data used in this study are made publicly available through the University of Pittsburgh Project Tycho online data system ( www . tycho . pitt . edu ) .
We extracted a total of 71 , 460 counts for 100 countries from DengueNet and WHO RO websites ( Fig 1 ) . These data represented a total of ~23 million dengue cases and ~82 , 000 deaths that have been reported to WHO between 1955 and 2012 . Of these , ~13 million cases ( 56% ) and ~20 , 000 deaths ( 24% ) were reported between 2000 and 2012 . A total of ~4 . 6 million cases were reported by WPRO ( 20% ) , ~3 . 2 million ( 14% ) by SEARO , and ~15 million ( 66% ) by PAHO countries ( Table 1 ) . The majority of dengue deaths were reported by SEARO ( 49% ) and WPRO ( 44% ) . Each source provided counts for a range of different indicators ( Fig 2 ) . Data for “all” dengue cases ( dengue fever and dengue hemorrhagic fever combined ) and “all” dengue deaths were available from DengueNet and all RO’s . Data for DHF cases were predominantly from PAHO , few counts were from WPRO and none from SEARO . Across time , DengueNet provided counts for the longest time period ( 1955–2011 ) compared to SEARO ( 1985–2006 ) , WPRO ( 2000–2011 ) , and PAHO ( 1995–2012 ) ( Fig 3 and S1 Fig ) . Across sources , data for “all” cases were provided for the longest time periods , followed by mortality data . Data for DHF counts were available for the shortest time periods ( Fig 3 and S1B Fig ) . In general , many counts were missing across years and countries . We compared data from DengueNet and RO’s to assess consistency across sources ( Table 2 and S2 Fig ) . The overall percent agreement was 83 . 8% across all indicators . Data from SEARO were the most consistent with a percent agreement of 92 . 2% and data from WPRO were the least consistent at 72 . 3% . Data for DHF cases were more consistent compared to the other indicators at 92 . 4% compared to 76 . 1% for “all” cases and 89 . 4% for “all” deaths . DengueNet values for all indicators were generally lower compared to values from RO’s . In total , DengueNet reported 426 , 808 fewer “all cases” , 17 , 854 fewer DHF cases , and 245 fewer deaths compared to RO’s ( Table 2 ) . We recomputed the number of “all” cases for DengueNet from separately reported dengue fever ( DF ) and DHF cases . We also recomputed the case fatality rate ( CFR ) for DengueNet from reported cases and deaths . Our recomputed data for “all” cases corresponded with 98 . 9% of original values and for CFR with 99 . 5% ( Table 3 ) . We also recomputed the annual number of dengue cases at country level from monthly cases at the provincial level ( in DengueNet , data were either reported at the country level by year or at the provincial level by month ) . Our recomputed annual country level data for “all” cases was > 3 million cases lower compared to reported data at that level . The recomputed values for “all” deaths were about 2000 deaths lower compared to reported mortality at country level by year . This discrepancy was likely due to missing data at the lower administrative levels . We found that provincial level data were not available for all calendar months in years before 1997 and after 2004 ( S3 Fig ) . In addition , provincial level data for countries were only available for a median of 3 . 5% of provinces before 1996 and for 85 . 7% of provinces after 1996 ( using The Second Administrative Level Boundaries data set project ( SALB ) [19] for the expected number of provinces per country ) . We also assessed dengue surveillance data provided online by the Ministries of Health of Brazil and Indonesia ( Fig 4 ) . Both these countries are dengue endemic and have developed online databases that provide publicly available dengue surveillance data . The annual number of “all” cases reported by the Brazil and Indonesia MOH corresponded to WHO data for most years except 2008 ( Brazil ) and 2000/2004 ( Indonesia ) . No WHO data were available for Indonesia after 2005 . We found discrepancies within the data provided by the MOH of Indonesia for years after 2007 . Our recomputed number of cases per year at the country level from reported provincial data ( 1st administrative level ) was higher than country level values recomputed from district data ( 2nd administrative level ) . This suggested that data from lower administrative levels were incomplete .
We integrated publicly available online dengue surveillance data from various WHO and country sources to describe the availability and consistency of globally available dengue surveillance data . We found that consistency of overlapping data between DengueNet and WHO Regional Offices was lacking and that data at subnational levels were often incomplete . This incompleteness was difficult to recognize since the absence of data for provinces or districts was not indicated explicitly . DengueNet systematically reported lower values compared to the RO’s . This may be due to a difference in timing of data reports made by countries and a lack of updating DengueNet as countries updated their figures . DengueNet was created by the WHO Headquarters in 2002 as part of the Global Health Atlas [9 , 20] . Focal points were appointed and trained in every country to upload standardized reports into the DengueNet repository [21] . This has successfully led to public sharing of dengue data across countries through a central global repository . In addition to DengueNet , RO’s also routinely release dengue surveillance data from member countries through their websites . PAHO and SEARO provide links to surveillance data sheets in PDF format and WPRO has developed an online Health Information and Intelligence Platform ( HIIP ) . The WHO is the only source of integrated disease surveillance data across countries . Numerous studies have used WHO dengue surveillance data to describe trends and patterns of this disease at the global [22 , 23 , 24] and regional level [25 , 26 , 27] . Despite their role as a core resource for international dengue surveillance data , DengueNet and some RO data have not been regularly updated over the past decade , most likely due to capacity and funding constraints . With the decline of WHO as a central global resource for dengue surveillance data , the data landscape will become increasingly scattered and difficult to navigate . Other agencies or institutes can contribute additional capacity or alternative frameworks for global disease surveillance data may be needed , such as a distributed network instead of centralized databases . Increasingly , individual countries disseminate their own disease surveillance data online in various formats ranging from epidemiological bulletins to sophisticated databases . This has greatly advanced the availability of disease data at the global level . In 2010 the 63rd World Health Assembly stated that “the WHO urges member states to improve the collection of reliable health information and data and to maximize , where appropriate , their free and unrestricted availability in the public domain” [28] . Country data systems however use a large diversity of surveillance methodology and definitions that often lack detailed documentation . The potential biases and lack of comparability of data across countries are limiting the efficient use of these data . The reporting process of dengue surveillance data from countries to WHO also lacks detailed documentation and may vary across countries . Future research should formally compare country data systems and country vs . WHO data to gain more insight in potential biases of the various sources . A standardized and curated global data system can maximize opportunities for the efficient use of country disease data for science and policy . Data standardization and curation are essential for a global data system . For example we found that ~16% of country names in DengueNet were different from country names used by the RO’s ( S1 Table ) . Across all WHO sources , ~19% of country names were different from the UN ISO standard for country names [19] . In the absence of up-to-date global platforms for disease surveillance data , alternative data systems have emerged such as Google Dengue and Flu Trends and the HealthMap project that automatically integrate data from search queries or online news items respectively [29 , 30 , 31] . Innovative technological solutions and capacity used by these projects should be applied to integrate country disease surveillance data as well to establish a state-of-the-art 21st century global data system . This system can be coordinated by WHO but can be implemented by external institutes that have already created large scale public health data systems such as the Institute of Health Metrics and Evaluation , the Malaria Atlas Project , or Project Tycho . A new and sustainable framework will be required to ensure that integrated and curated disease surveillance data from countries around the world will continue to be available to stakeholders at all levels . Innovative technology should be used for data integration that minimizes the burden on countries but maximizes data availability and use . Academic and private sector partners should step up to support international agencies with this increasingly complex mandate . | The use of high quality data and information has become essential for public health agencies to monitor and protect population health . Technological advancement has enabled the development of sophisticated disease surveillance systems by many countries . Increasingly , countries are making surveillance data publicly available to their constituencies . A key role of international agencies such as the World Health Organization is the integration and curation of country data at the global level . Because it can be confusing to navigate the current online disease data landscape , we assessed the availability and consistency of online available surveillance data for dengue provided by the World Health Organization and two example countries ( Brazil and Indonesia ) . We found that data availability declined substantially after 2005 and that consistency between sources was limited to 84% , representing a discrepancy of half a million cases . These limitations reduce opportunities for the efficient use of country data to counter public health threats . A new financial and operational model is needed to advance the use of disease data at the global level . Industry and academic partners need to step up to support this mandate . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | The Availability and Consistency of Dengue Surveillance Data Provided Online by the World Health Organization |
HIV-1 is internalized into mature dendritic cells ( mDCs ) via an as yet undefined mechanism with subsequent transfer of stored , infectious virus to CD4+ T lymphocytes . Thus , HIV-1 subverts a DC antigen capture mechanism to promote viral spread . Here , we show that gangliosides in the HIV-1 membrane are the key molecules for mDC uptake . HIV-1 virus-like particles and liposomes mimicking the HIV-1 lipid composition were shown to use a common internalization pathway and the same trafficking route within mDCs . Hence , these results demonstrate that gangliosides can act as viral attachment factors , in addition to their well known function as cellular receptors for certain viruses . Furthermore , the sialyllactose molecule present in specific gangliosides was identified as the determinant moiety for mDC HIV-1 uptake . Thus , sialyllactose represents a novel molecular recognition pattern for mDC capture , and may be crucial both for antigen presentation leading to immunity against pathogens and for succumbing to subversion by HIV-1 .
Dendritic cells ( DCs ) are the most potent antigen-presenting cells found in the organism and play a paramount role in initiating immune responses to assaulting pathogens [1] , [2] . DCs that patrol the mucosal tissue display an immature phenotype and are able to capture incoming pathogens , which leads to DC activation , maturation , and migration to the secondary lymphoid tissue , where DCs acquire the mature phenotype required to efficiently induce adaptive immune responses [1] . Given the unique role of DCs in initiating primary immune responses [1] , it is generally believed that these antigen-presenting cells are critical to induce resistance to infection [2]–[4] . In the specific case of viral infections , murine DC depletion models have provided in vivo evidence of the DC requirement to induce both humoral and cellular antiviral immune responses [5] , [6] . However , some viruses , including HIV-1 , have evolved strategies to subvert DC antiviral activity [7] , [8] . DCs can even promote HIV-l dissemination , both through the direct release of new virus particles after productive infection , and through transmission of captured viruses to susceptible T cells without DC infection , a process known as trans-infection ( reviewed in [9] ) . Direct infection and trans-infection occur to a different extent in immature DCs ( iDCs ) and mature DCs ( mDCs ) ( reviewed in [9] , [10] ) . The initial Trojan horse hypothesis suggested that HIV-1 capture by iDCs in the mucosa may protect the virus from degradation and allow its transport to secondary lymphoid organs , facilitating trans-infection of CD4+ T cells and fueling viral spread [11] , [12] . However , iDCs show rapid degradation of captured viral particles [13]– , and several lines of evidence suggest now that the long-term ability of iDCs to transfer HIV-1 relies on de novo production of viral particles after productive infection [14] , [16] , [17] . HIV-1 replication in DCs is generally less productive than in CD4+ T cells ( reviewed in [9] , [18] ) , however , probably due to the presence of cellular restriction factors such as SAMHD1 that limit reverse transcription following viral entry [19] , [20] . Maturation of DCs further reduces their ability to support HIV-1 replication [21]–[24] , but potently enhances their capacity to trans-infect bystander CD4+ T cells [15] , [21] , [25]–[27] . Trans-infection occurs via the infectious synapse , a cell-to-cell contact zone that facilitates transmission of HIV-1 by locally concentrating virus and viral receptors [26] , [28] . Strikingly , poorly macropinocytic mDCs [29] sequester significantly more complete , structurally intact virions into large vesicles than actively endocytic iDCs [30] , and retain greater amounts of virus 48 h post-pulse than iDCs immediately after viral wash [15] . Thus , enhanced mDC trans-infection correlates with increased HIV-1 capture and a longer life span of trapped viruses [15] , [27] . Furthermore , mDCs efficiently interact with CD4+ T cells in lymphoid tissues; key sites of viral replication , where naïve CD4+ T cells are activated and turn highly susceptible to HIV-1 infection [31] . Accordingly , carriage of HIV-1 by mDCs could facilitate the loss of antigen-specific CD4+ T cells [32] , [33] , favoring HIV-1 pathogenesis . However , the molecular mechanism underlying HIV-1 uptake by mDCs remains largely uncharacterized . We have previously identified an HIV-1 gp120-independent mechanism of viral binding and uptake that is upregulated upon DC maturation [15] . Furthermore , HIV-1 Gag enhanced green fluorescent protein ( eGFP ) –expressing fluorescent virus-like particles ( VLPHIV-Gag-eGFP ) follow the same trafficking route as wild-type HIV-1 in mDCs [34] , and hence share a common molecular pattern that governs entry into mDCs . In addition , we also reported that treatment of HIV-1 or VLP producer cells with inhibitors of glycosphingolipid biosynthesis yielded particles with less glycosphingolipids , which exhibited reduced entry into mDCs [34] , [35] , without affecting net release from producer cells [36] . Thus , we hypothesize that gangliosides in the outer monolayer of HIV-1 and VLP membranes could act as viral attachment factors and allow viral recognition and capture by mDCs . Here we sought to investigate the molecular determinants involved in viral binding and internalization mediated by mDCs . Using liposomes to mimic the lipid composition and size of HIV-1 , we demonstrate that gangliosides are the key molecules that mediate liposome uptake . We extended these observations to VLPs and HIV-1 , characterizing a new role for these glycosphingolipids as viral attachment factors . Furthermore , we identify sialyllactose on HIV-1 membrane gangliosides as a novel molecular recognition pattern that mediates virus uptake into mDCs .
Considering that glycosphingolipids are enriched in raft-like plasma membrane domains [37]–[39] from where HIV-1 is thought to bud ( reviewed in [40] ) , we investigated the potential role of glycosphingolipids for HIV-1 capture by mDCs . The ganglioside GM3 was previously identified in the membrane of different retroviruses including HIV-1 [41] , [42] . We were able to confirm the presence of GM3 in HIVNL4-3 derived from the T-cell line MT-4 by mass spectrometry ( Figure 1A ) . In addition , we detected several other gangliosides including GM1 , GM2 , and GD1 in the HIV-1 membrane ( Figure 1A ) . To test whether gangliosides in the outer leaflet of HIV-1 or vesicular membranes can act as attachment factors yielding mDC uptake , we prepared Texas Red ( tRed ) -labeled large unilamellar vesicles ( LUV ) mimicking the size and lipid composition of HIV-1 ( referred to as LUVHIV-tRed and prepared as in [43] ) and containing different gangliosides ( Figure S1 ) . All LUVs displayed equal fluorescence intensities ( Figure S2 ) . mDCs were pulsed with either LUVHIV-tRed or VLPs for 4 h at 37°C and the percentage of fluorescent cells was determined by fluorescence activated cell sorting ( FACS ) . Similar to our previous results with infectious HIV-1 [15] , a high percentage of mDCs captured the fluorescent VLPHIV-Gag-eGFP ( Figure 1B ) . Furthermore , VLPs produced in the CHO cell line , which is only able to synthesize gangliosides up to GM3 [44] , were also efficiently captured by mDCs ( Figure 1C ) . Uptake into mDCs was further observed for the murine retrovirus MuLV ( Figure 1D ) , which was previously shown to also contain gangliosides [41] . On the other hand , no significant uptake into mDCs was observed for LUVHIV-tRed , which contained the main lipid constituents of HIV-1 , but were devoid of gangliosides ( p<0 . 0001 , paired t test ) ( Figures 1B and S3 ) . Uptake into mDCs remained negative for LUVHIV-tRed containing ceramide ( Cer ) ( p<0 . 0001 , paired t test ) ( Figures 1B and S3 ) . This was completely different when monosialogangliosides such as GM3 , GM2 , or GM1a were incorporated into the LUVs; mDCs were able to capture these liposomes with equal efficiency as VLPHIV-Gag-eGFP ( Figures 1B and S3 ) . To ensure that this capture was not merely due to electrostatic interactions between negatively charged gangliosides and surface charges on mDCs , LUVHIV-tRed containing negatively charged phosphatidylserine ( PS ) were analyzed in parallel and were found to be negative for mDC capture ( p = 0 . 0081 , paired t test ) ( Figure 1B ) . These results reveal that monosialogangliosides mediate LUV capture by mDCs , and that the carbohydrate head group is essential for this process . To determine whether ganglioside-containing LUVHIV-tRed and VLPHIV-Gag-eGFP exploit a common entry mechanism into mDCs , we performed competition experiments . mDCs were pulsed with decreasing amounts of GM2-containing LUVHIV-tRed and a constant amount of VLPHIV-Gag-eGFP for 4 h at 37°C . After extensive washing , the percentage of eGFP- and tRed-positive cells was determined by FACS . GM2-containing LUVHIV-tRed efficiently competed for the uptake of VLPHIV-Gag-eGFP into mDCs in a dose-dependent manner ( p<0 . 0001 , paired t test ) ( Figure 1E ) . However , no competition for VLP uptake was observed for LUVHIV-tRed containing Cer or lacking glycosphingolipids ( Figure 1E ) . Hence , GM-containing LUVHIV-tRed and VLPHIV-Gag-eGFPuse a common entry mechanism to gain access into mDCs , which is dependent on the carbohydrate head group . We next investigated whether GM-containing LUVHIV-tRed and VLPHIV-Gag-eGFP reach the same compartment in mDCs using spinning-disc confocal microscopy . We had previously described three types of patterns for HIV-1 captured into mDC: random , polarized , or sac-like compartments [15] , [45] . The same patterns were also observed for GM-containing LUVHIV-tRed and the percentage of mDCs displaying the different patterns was similar regardless of the particle used ( Figure 2A ) . Thus , VLPHIV-Gag-eGFP and GM-containing LUVHIV-tRed not only compete for internalization , but also traffic to an analogous compartment within mDCs . To determine whether VLPHIV-Gag-eGFP and GM-containing LUVHIV-tRed are captured into the same compartment , mDCs were pre-incubated 3 h at 37°C with GM-containing LUVHIV-tRed and subsequently incubated with VLPHIV-Gag-eGFP for three additional hours . Confocal microscopy of fixed cells revealed that GM-containing LUVHIV-tRed and VLPs polarized towards the same cell area in mDCs ( Figure S4 ) . Furthermore , VLPs extensively co-localized with GM-containing LUVHIV-tRed ( including either GM1a , GM2 , or GM3 ) in the same intracellular compartment ( Figure 2B; Video S1 ) . The HIV-1 envelope is a liquid-ordered membrane and gangliosides are presumably enriched in this type of membranes [41] , [43] , [46] , [47] . Moreover , ganglioside interaction with cholesterol in lipid rafts [37] , [38] , [47] is known to influence ganglioside conformation and alter its activity as a cellular receptor [48] . We therefore assessed whether membrane structure or the specific lipid composition ( other than gangliosides ) of the particle membrane influenced mDC capture . mDCs were incubated with LUVPOPC-tRedcomposed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine ( POPC ) with or without different gangliosides ( Figure 3A ) . In contrast to LUVHIV-tRed , LUVPOPC-tRed have a liquid-disordered membrane structure [38] . Results for LUVPOPC-tRed were very similar to LUVHIV-tRed with efficient capture if either GM1a , GM2 , or GM3 was present , while no uptake was observed for Cer containing LUVPOPC-tRed or LUVPOPC-tRed lacking gangliosides ( Figure 3A ) . Furthermore , the percentage of mDCs displaying particles in random , polarized or sac-like compartment capture patterns was again very similar for the different particles ( Figure 3B ) . These results show that ganglioside-containing LUVs use the same trafficking pathway as VLPHIV-Gag-eGFP regardless of their membrane structure , indicating that gangliosides themselves are the key molecules responsible for mDC capture . To gain further insight into the molecular determinant structure required for efficient recognition by mDCs , LUVHIV-tRed carrying more complex gangliosides were produced , including two , three , and four sialic acid groups at diverse positions in the carbohydrate polar head group ( di- , tri- , and tetra-sialogangliosides ) ( Figure 4A ) . mDCs pulsed with an equal amount of LUVHIV-tRed-containing gangliosides with two or three sialic acids ( GD1b and GT1b , respectively ) captured these particles with the same efficiency as GM1a-LUVHIV-tRed ( Figure 4A ) . However , capture was almost completely lost for LUVHIV-tRed containing a ganglioside with four sialic acids ( GQ1b ) ( Figure 4A ) . Accordingly , LUVHIV-tRed carrying GD1b or GT1b efficiently competed for mDC uptake with VLPHIV-Gag-eGFP , while no competition was observed for LUVHIV-tRed carrying GQ1b , PS , or Cer ( Figure 4B ) . These results indicate that complex gangliosides with up to three sialic acids located in distinct positions of the carbohydrate head group share a common structure determinant for mDC uptake , which is lost in GQ1b . The lack of internalization of Cer-containing LUVs indicated that the carbohydrate head group is specifically required for mDC capture . Sialic acid has previously been identified as a cellular receptor for certain viruses [49] . We therefore tested its importance for mDC capture . Incubation of mDCs with equal concentrations of LUVHIV-tRed containing Cer , GM1a , or GM1 without the sialic acid group ( Asialo GM1 ) revealed sialic acid-dependent capture ( Figure 5A ) . In addition , in situ neuraminidase treatment of GM3-containing LUVHIV-tRed and VLPHIV-Gag-eGFP , significantly reduced particle capture ( Figure 5B ) and LUV binding to mDCs ( Figure S5 ) . Thus , the sialic acid moiety in gangliosides is necessary for specific recognition by mDCs . To assess the contribution of other components of the carbohydrate head group , we prepared LUVHIV-tRed containing either GM4 ( lacking the glucose moiety of GM3 ) ( Figure 5C ) or GalCer ( lacking both the glucose and sialic acid moieties of GM3 ) ( Figure 5C ) . mDCs incubated with GM4- or GalCer-containing LUVHIV-tRed showed only background levels of liposome capture ( Figure 5C ) , indicating that the glucose moiety of sphingolipids is also necessary for mDC capture . Given that the carbohydrate moiety within gangliosides constitutes the molecular recognition determinant for mDC capture , these head groups should compete for VLP and LUV uptake . Capture of GM3-containing LUVHIV-tRed or VLPHIV-Gag-eGFP by mDCs was completely blocked in the presence of the GM3 polar head group ( sialyllactose ) , while equal concentrations of lactose ( lacking the sialic acid group ) had no effect ( Figure 5D ) . Taken together , these data clearly show that the sialyllactose moiety of gangliosides is the molecular determinant required for efficient VLP and LUV recognition and capture by mDCs . Noteworthy , the high concentrations of the GM3 head group required for competition in Figure 5D compared to the low concentrations of gangliosides in LUVs needed to outcompete VLPs ( ∼1 , 000-fold less ) ( Figure 1C ) suggests that the attachment of sialyllactose to Cer within membranes confers a higher binding avidity . This is not surprising since viruses and toxins are multivalent: for instance , the VP1 capsid protein of SV40 is pentameric and the cholera toxin is pentavalent; both bind five GM1 molecules [50]–[52] . In addition , the hydrophilic moiety of Cer in the membrane interface could be part of the recognition domain , directly increasing the binding affinity to mDCs . Modeling of the 3-D structure of gangliosides ( Figure S6 ) suggested that mDC recognition required an exposed sialyllactose domain and that the lack of recognition of GQ1b could be caused by steric hindrance ( Figure S6 ) . Thus , although the lipid and protein context of surrounding membranes might influence the biologically active conformation of gangliosides [48] , [53]; 3-D reconstructions provide a structural basis to explain the distinct recognition patterns observed for different gangliosides . The sialyllactose recognition domain defined in this study differs from the NeuAc α 2 , 3Gal β 1 , 3GalNAc on gangliosides identified as host cell receptors for SV40 and Sendai virus , and for cholera and tetanus toxin [50] , [52] , [54] , but is identical with the α 2 , 3-sialyllactose identified as host cell receptor for other paramyxoviruses [55] , [56] . To determine whether the results obtained with LUVs and VLPs also hold true for the authentic virus , we performed experiments with wild-type HIVNL4-3 produced in primary T cells . Similar to our previous data [15] , a high percentage of mDCs captured HIV-1 , while uptake into iDCs was much less efficient ( p = 0 . 0047 , one sample t test ) ( Figure 6A ) . To confirm the importance of viral gangliosides for mDC capture , we purified HIVNL4-3 from primary CD4+ T cells pre-treated or not with the glycosphingolipid biosynthesis inhibitor NB-DNJ . HIV-1 capture was strongly reduced for virus obtained from inhibitor-treated cells compared to control virus ( p<0 . 0001 , one sample t test ) ( Figure 6B ) . To directly determine the importance of the sialyllactose head group for mDC capture of authentic HIV-1 , we performed competition experiments showing a strong reduction of virus capture in the presence of the GM3 polar head group , but not in the presence of lactose ( p<0 . 0001 , one sample t test ) ( Figure 6C ) . These data confirm the observations obtained with liposomes and VLPs for authentic HIV-1 from primary CD4+ T cells . HIV-1 capture by mDCs has been shown to promote trans-infection of CD4+ T cells and other target cells [15] , [21] , [25]–[27] , and we therefore analyzed whether sialyllactose recognition by mDCs is also important for viral transmission . Co-culturing mDCs that had been exposed to an equivalent amount of infectious HIVNL4-3 derived from NB-DNJ-treated or untreated CD4+ T cells with the TZM-bl reporter cell line revealed a strong reduction of trans-infection for the virus from inhibitor-treated cells compared to control virus ( p = 0 . 0404 , paired t test ) ( Figure 6D ) . We also observed a strong reduction of trans-infection for mDCs pulsed with HIVNL4-3 in the presence of the GM3 polar head group and subsequently incubated with TZM-bl cells ( Figure 6E ) ( p = 0 . 0197 , paired t test ) . These results were further confirmed when we co-cultured HIV-1-pulsed mDCs with activated primary CD4+ T cells . Co-cultures were performed in the presence or absence of the protease inhibitor saquinavir to distinguish net trans-infection from re-infection events ( Figure 6F and 6G; left and right panels , respectively ) . Infection of primary CD4+ T cells was strongly enhanced when they were co-cultured with HIV-1 pulsed mDCs ( Figure 6F and 6G; filled bars ) . This effect was abrogated when mDCs were pulsed with virus produced from NB-DNJ-treated cells ( Figure 6F ) or cultured with the GM3 polar head group ( Figure 6G ) . These data indicate that the sialyllactose moiety of gangliosides is the molecular determinant required for efficient HIV-1 capture by mDCs and for subsequent viral trans-infection .
Sialic acid on gangliosides has previously been shown to function as host cell receptor for several viruses [54] , [57] , [58] and for human toxins [52] . The current study clearly identifies a novel role for sialylated gangliosides in the membrane of viruses or LUVs as determinants for specific capture by mDCs . This recognition is reminiscent of the engulfment of apoptotic cells by phagocytes such as DCs , which is triggered by PS , a phospholipid normally found in the inner leaflet of the plasma membrane of living cells , but exposed on the surface of dying cells [59] . However , the viral capture described here is dependent on an exposed sialyllactose moiety on gangliosides , which we identified as a novel molecular recognition pattern . The ganglioside GM3 was previously detected in the membrane of HIV-1 and several other viruses ( SFV , VSV , MuLV ) [41] , [42] , and this was extended to GM1 , GM2 , and GD1 for HIV-1 in the present study . Gangliosides are significant components of the plasma membrane lipidome [42] , suggesting that all enveloped viruses , which bud from the plasma membrane of infected cells , may be captured into mDCs by the reported mechanism unless they exclude sialyllactose-containing gangliosides . Most studies of viral lipidomes have not included gangliosides so far , however , and it will be important to determine whether certain viruses have developed mechanisms to prevent sialyllactose presentation on their membrane lipids . This would be conceivable for influenza virus , which carries a viral neuraminidase needed for removal of the sialic acid receptor from the producer cell surface , thus allowing virus release . This neuraminidase may also remove sialic acid groups from viral gangliosides , thus preventing mDC uptake and potential antigen presentation . Consequently , neuraminidase inhibitor treatment should lead to increased DC capture and potentially enhance immunogenicity of influenza virus or VLPs . Moreover , viral ganglioside content may vary depending on the membrane composition of the producer cell . Viral replication in the nervous system , where gangliosides are particularly enriched [60] , may thus lead to the insertion of an increased amount of distinct gangliosides into virions , affecting mDC recognition and local immunosurveillance . The potential immunological role of mDC uptake implies efficient antigen capture and processing throughout the antigen presentation pathway . Interestingly , antigen-bearing cellular secreted vesicles known as exosomes also follow the same trafficking route as HIV-1 [34] and contain gangliosides such as GM3 or GM1 [61] . Hence , sialyllactose-carrying gangliosides in the membrane of viruses and cellular vesicles are targeting molecules for mDC uptake , a pathway that may normally lead to antigen processing and presentation and has been subverted by HIV-1 for infectious virus storage and transmission . Furthermore , although downregulation of endocytosis is considered a hallmark of DC maturation ( reviewed in [3] ) , there is increasing evidence that under inflammatory conditions mDCs capture , process , and present antigens without exclusively relying on prior pathogen exposure [62] , [63] . This scenario might be particularly relevant in chronic infections , such as the one caused by HIV-1 , where translocation of bacteria from the intestinal lumen [64] could stimulate DCs systemically and contribute to sustained antiviral immune responses . Remarkably , HIV-1 infected patients show enhanced GM3 content in the plasma membrane of T lymphocytes and high titers of anti-lymphocytic GM3 antibodies [65] , [66] . Paradoxically , HIV-1 capture into mDCs appears to also critically enhance viral dissemination in lymphoid tissue by efficient release of infectious virus to CD4+ T cells in the DC-T-cell synapse , thus promoting pathogenesis and disease progression through trans-infection . Hence , although myeloid cells are largely refractory to productive HIV-1 infection due to the presence of cellular restriction factors such as the recently identified SAMHD1 [19] , [20] , the sialyllactose driven trans-infection process characterized here in mDCs seems to exploit a pre-existing cellular-trafficking machinery that avoids the activation of these intrinsic immune pathways . The efficient capture of ganglioside-carrying cellular vesicles or virions suggests a model where a specific receptor present on the cell surface of mDCs ( and possibly other cells ) recognizes the sialyllactose moiety on virions or vesicle membranes . Gangliosides have been reported to function as cell adhesion molecules [67] , and this may also involve such a receptor . Specific recognition of vesicular gangliosides would then trigger uptake of the respective particles into an intracellular compartment from where they are either recycled to the surface ( as in HIV-1 transmission to CD4+ T cells ) or fed into the antigen presentation pathway . The results of the current study identify sialyllactose on membrane gangliosides as the relevant molecular recognition pattern , explaining the specificity of this process and providing the basis for its future exploitation for interventional or vaccine purposes .
The institutional review board on biomedical research from Hospital Germans Trias i Pujol approved this study . MT-4 cells were infected with HIVNL4-3 and co-cultured with uninfected cells . Virus was harvested before cytopathic effects were observed , and purified essentially as described in [43] . Briefly , medium was cleared by filtration , and particles were concentrated by ultracentrifugation through a cushion of 20% ( w/w ) sucrose . Concentrated HIV-1 was further purified by velocity gradient centrifugation on an OptiPrep gradient ( Axis-Shield ) . The visible virus fraction was collected and concentrated by centrifugation . The final pellet was resuspended in 10 mM Hepes , 150 mM NaCl , pH 7 . 4 ( hepes-sodium buffer ) , rapidly frozen in liquid nitrogen and stored at −80°C . For lipid composition analysis , samples were resuspended in methanol upon thawing and then assessed in a UPLC coupled to an orthogonal acceleration time-of-flight mass spectrometer with an electrospray ionization interface ( LCT Premier; Waters ) using the procedure previously described in [68] . Data were acquired using positive ionization mode over a mass range of m/z 50–1 , 500 in W-mode . A scan time of 0 . 15 s and interscan delay of 0 . 01 s were used at a nominal instrument resolution of 11 , 500 ( FWHM ) . Leucine enkephalin was used as the lock spray calibrant . Peripheral blood mononuclear cells ( PBMCs ) were obtained from HIV-1-seronegative donors and monocyte populations ( >97% CD14+ ) were isolated with CD14+-positive selection magnetic beads ( Miltenyi Biotec ) . DCs were obtained culturing these cells in the presence of 1 , 000 IU/ml of granulocyte-macrophage colony-stimulating factor ( GM-CSF ) and interleukin-4 ( IL-4; R&D ) . mDCs were differentiated by culturing iDCs at day five for two more days in the presence of 100 ng/ml of lipopolysaccharide ( LPS; Sigma ) . DCs were immunophenotyped at day 7 as previously described [15] . Adequate differentiation from monocytes to iDCs was based on the loss of CD14 and the acquisition of DC-SIGN , while DC maturation upregulated the expression of CD83 , CD86 , and HLA-DR . CD4+ T lymphocytes required for trans-infection experiments were isolated from PBMCs with CD4-negative selection magnetic beads ( Miltenyi Biotec ) and stimulated for 72 h in the presence of 10 IU/ml of IL-2 ( Roche ) and 3 µg/ml of phytohaemagglutinin ( PHA; SigmaAldrich ) . Primary cells were maintained in RPMI with 10% fetal bovine serum ( FBS ) , 100 IU/ml of penicillin and 100 µg/ml of streptomycin ( all from Invitrogen ) . HEK-293T and TZM-bl ( obtained through the US National Institutes of Health [NIH] AIDS Research and Reference Reagent Program , from JC Kappes , X Wu , and Tranzyme Inc . ) were maintained in D-MEM ( Invitrogen ) . CHO cell line was maintained in α-MEM . MT4 cell line was maintained in RPMI . All media contained 10% FBS , 100 IU/ml of penicillin , and 100 µg/ml of streptomycin . VLPHIV-Gag-eGFP were obtained transfecting the molecular clone pGag-eGFP ( obtained through the NIH AIDS Research and Reference Reagent Program , from MD Resh ) . VLPMLV-Gag-YFP were obtained transfecting the molecular clone pGag-MLV wt [69] . HEK-293T cells were transfected with calcium phosphate ( CalPhos , Clontech ) in T75 flasks using 30 µg of plasmid DNA . CHO cells were electroporated ( 0 . 24 Kv and 950 µF ) using 7×106 cells and 40 µg of plasmid DNA . Supernatants containing VLPs were filtered ( Millex HV , 0 . 45 µm; Millipore ) and frozen at −80°C until use . For studies with concentrated VLPs , medium was harvested , cleared by filtration , and particles were concentrated by ultracentrifugation ( 28 , 000 rpm 2 h at 4°C in SW32 rotor ) through 20% ( w/w ) sucrose . The final pellet was resuspended in hepes-sodium buffer , rapidly frozen in liquid nitrogen , and stored at −80°C . The p24Gag content of the MT4-derived viral stock and VLPHIV-Gag-eGFPwas determined by an ELISA ( Perkin-Elmer ) and by a quantitative Western blot . Detection was carried out with a LiCoR Odyssey system employing our own Rabbit anti-capsid polyclonal antibody and a purified Gag protein ( kindly provided by J Mak ) as a standard . To produce HIVNL4-3 in PBMCs , cells were stimulated with 3 µg/ml PHA and 10 IU/ml of IL-2 for 72 h prior to infection with HIVNL4-3 . To generate HIVNL4-3 in CD4+ T cells , whole blood from three HIV-1–seronegative donors were CD8+ T cell depleted using Rossettesep anti-CD8+ cocktail ( Stem cell ) . Enriched CD4+ T cells were pooled and stimulated under three different conditions: low-dose PHA ( 0 . 5 µg/ml ) , high-dose PHA ( 5 µg/ml ) , or plate-bound anti-CD3 monoclonal antibody OKT3 ( e-Bioscience ) [70] . After 72 h , cells were mixed together , infected with HIVNL4-3 and resuspended to a final concentration of 106 cells/ml in RPM1 with 10% FBS supplemented with 100 IU/ml of IL-2 . To produce glycosphingolipid deficient HIVNL4-3 , enriched CD4+ T cells were kept in the presence or absence of 500 µM of NB-DNJ ( Calbiochem ) and 10 IU/ml of IL-2 for 6 d before stimulation with 3 µg/ml of PHA and subsequent infection with HIVNL4-3 . Virus growth was monitored by p24Gag ELISA ( Perkin Elmer ) . Supernatants were harvested when the concentration of p24Gag was at least 102 ng/ml , filtered and stored at −80°C until use . Titers of all viruses were determined using the TZM-bl indicator cell line as described elsewhere [71] . LUVs were prepared following the extrusion method described in [72] . Lipids were purchased from Avanti Polar Lipids and gangliosides were obtained from Santa Cruz Biotechnology . Ganglioside source was bovine brain with the exception of GM4 , which was from human brain . The LUVHIV-tRed lipid composition was: POPC 25 mol%: 1 , 2-dipalmitoyl-sn-glycero-3-phosphocholine ( DPPC ) 16 mol%: brain sphingomyelin ( SM ) 14 mol%: cholesterol ( Chol ) 45 mol% , and when Cer , PS , or gangliosides were present ( 4 mol% ) the SM amount was reduced to 10 mol% . The LUVPOPC-tRed lipid composition was 96 mol% POPC containing or not 4 mol% of Cer , GM3 , GM2 , or GM1a . All the LUVs contained 2 mol% of 1 , 2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine ( DHPE ) -tRed ( Molecular Probes ) . Lipids were mixed in chloroform∶methanol ( 2∶1 ) and dried under nitrogen . Traces of organic solvent were removed by vacuum pumping for 1–2 h . Subsequently , the dried lipid film was dispersed in hepes-sodium buffer and subjected to ten freeze-thaw cycles prior to extruding ten times through two stacked polycarbonate membranes with a 100-nm pore size ( Nucleopore , Inc . ) using the Thermo-barrel extruder ( Lipex extruder , Northern Lipids , Inc . ) . In order to perform mDC pulse with equal concentrations of LUV displaying similar fluorescence intensities , tRed containing LUVs concentration was quantified following the phosphate determination method of [73] and the fluorescence emission spectra was recorded setting the excitation at 580 nm in a SLM Aminco series 2 spectrofluorimeter ( Spectronic Instruments ) . All capture experiments were performed pulsing mDCs in parallel at a constant rate of 100 µM of distinct LUVtRed formulations and 75 ng of VLPHIV-Gag-eGFP Gag quantified by western blot ( 2 , 500 pg of VLPHIV-Gag-eGFP p24Gag estimated by ELISA ) per 2×105 cells for 4 h at 37°C . After extensive washing , positive DCs were acquired by FACS with a FACSCalibur ( BD ) using CellQuest software to analyze collected data . Forward-angle and side-scatter light gating were used to exclude dead cells and debris from all the analysis . Competition experiments were done incubating 2×105 mDCs with 75 ng of VLPHIV-Gag-eGFP Gag at a final concentration of 1×106 cells/ml for 4 h at 37°C in the presence of decreasing amounts of GM2-containing LUVHIV-tRed or 100 µM of Cer- and PS-containing LUVHIV-tRed . Alternatively , cells were incubated with 75 ng of VLPHIV-Gag-eGFP Gag and 100 µM of LUVHIV-tRed including or not GM1a , GD1b , GT1b , GQ1b , Cer , and PS . Cells were then analyzed by FACS as previously described . Co-localization experiments were done pulsing mDCs sequentially with LUVHIV-tRed and VLPHIV-Gag-eGFP for 3 h as described in the capture assays section . Capture patterns were analyzed similarly , incubating cells with distinct LUVHIV-tRed and VLPHIV-Gag-eGFP separately . After extensive washing , cells were fixed and cytospun into glass slides , mounted in DAPI-containing fluorescent media and sealed with nail polish to analyze them in a spinning disk confocal microscope . Z-sections were acquired at 0 . 1-µm steps using a 60× Nikon objective . Spinning disk confocal microscopy was performed on a Nikon TI Eclipse inverted optical microscope equipped with an Ultraview spinning disk setup ( PerkinElmer ) fitted with a two CCD camera ( Hamamatsu ) . The co-localization signals in percentages of the compartment area and the thresholded Manders and Pearson coefficients were calculated for each image using Volocity 5 . 1 software ( Improvision ) . To obtain 3-D reconstructions , confocal Z stacks were processed with Volocity 5 . 1 software , employing the isosurface module for the nucleus and the maximum fluorescent intensity projection for LUVs and VLPs . A total of 2×105 DCs were pulsed for 2 h at 37°C with 25 µM of GM3-containing LUVHIV-tRed and 75 ng of sucrose-pelleted VLPHIV-Gag-eGFP Gag treated or not during 12 h at 37°C with 100 or 50 mU of neuraminidase from Clostridium perfringens Factor X ( Sigma Aldrich ) . The 12-h incubation was done in a glass-coated plate ( LabHut ) in hepes-sodium buffer , and the reaction was stopped adding RPMI media containing FBS . Cells were washed and assessed by FACS to obtain the percentage of tRed- and eGFP-positive cells . mDCs were preincubated with or without 5 or 10 mM of lactose ( Sigma-Aldrich ) and soluble GM3 carbohydrate head group ( Carbosynth ) for 30 min at RT . Cells were then pulsed with 50 µM of GM3-containing LUVHIV-tRed and 75 ng of sucrose-pelleted VLPHIV-Gag-eGFP Gag for 2 h at 37°C , at a final concentration of 5 or 10 mM for the compounds tested . Cells were analyzed by FACS as described previously . For experiments with HIVNL4-3 generated in primary CD4+ T cells , mDCs were equally pre-incubated with lactose and the GM3 head group . Minimal energy structures in vacuum were computed using Chem3D Ultra software employing the MM2-force field and the steepest-descent algorithm . Minimum root mean square gradient was set to 0 . 1; minimum and maximum move to 0 . 00001 and 1 . 0 , respectively . mDCs and iDCs ( 5×105 cells ) were exposed to 30 ng p24Gag of HIVNL4-3 obtained from stimulated PBMCs for 4 h at 37°C . Cells were washed thoroughly to remove unbound particles , lysed , and assayed for cell-associated p24Gag content by an ELISA . mDCs ( 2 . 5×105 cells ) were exposed to 50 ng p24Gag of HIVNL4-3 obtained from CD4+ T cells treated or not with NB-DNJ , incubated and assayed as previously described to detect the cell-associated p24Gag . Alternatively , mDCs ( 2 . 5×105 cells ) pre-incubated or not with GM3 or lactose as previously indicated were exposed to 90 ng p24Gag of HIVNL4-3 obtained from CD4+ T cells and assayed equally . For trans-infection assays , mDCs were pulsed equally but with a constant multiplicity of infection ( MOI ) of 0 . 1 , extensively washed and co-cultured in quadruplicate with the TZM-bl reporter cell line at a ratio of 104∶104 cells in the presence of 0 . 5 µM of saquinavir to assay luciferase activity 48 h later ( BrightGLo luciferase system; Promega ) in a Fluoroskan Ascent FL luminometer . Background values consisting of non-HIV-1 pulsed co-cultures were subtracted for each sample ( mean background of 8 . 668 RLUs×100 ) . Trans-infection to primary cells was performed similarly , co-culturing pulsed mDCs with activated primary CD4+ T cells for 48 h on a 96-well U-bottom plate without removal of unbound viral particles . Co-cultures were performed in the presence or in the absence of 0 . 5 µM of saquinavir . Infection of activated primary CD4+ T cells was detected with FACS , measuring the intracellular p24Gag content within the CD2-positive CD11c negative population of CD4+ T cells employing the monoclonal antibodies p24Gag-FITC ( KC57-FITC , clone FH190-1-1 , Beckman Coulter ) , CD2-PerCP Cy5 . 5 ( clone RPA-2 . 10 , BD Pharmingen ) , and CD11c-APC Cy7 ( clone Bu15 , BioLegend ) . Co-cultures containing non-pulsed cells were used as a background control for p24Gag labeling and used to set up the marker at 0 . 5% . To detect the possible cell free virus infection of activated CD4+ T cells , an equal MOI was added directly to control wells lacking mDCs . Statistics were performed using GraphPad Prism v . 5 software . | Antigen-presenting cells such as dendritic cells ( DCs ) are required to combat infections , but viruses including HIV have evolved strategies to evade their anti-viral activity . HIV can enter DCs via a non-infectious endocytic mechanism and trick them into passing infectious virus on to bystander CD4+ T cells . Immature DC ( iDCs ) are characterized by high endocytic activity and low T-cell activation potential . Interestingly , several groups have shown that DCs that have undergone “‘maturation’” ( mDCs ) , a process that occurs on contact with a presentable antigen , capture higher numbers of HIV-1 particles than iDCs when they are matured in the presence of lipopolysaccharide . mDCs move to the lymph nodes where they have more opportunity to interact with T cells than iDCs , and thus to pass on infectious virus . But the molecular mechanism underlying HIV-1 uptake by mDCs has until now been elusive . Here we show that gangliosides , basic components of the host cell's plasma membrane , have an important role in this process . Gangliosides are known to be incorporated into the viral envelope membrane during the process of viral particle budding and here we show that they serve as viral attachment factors: they are recognized and enable HIV-1 uptake by mDCs . Thus , in addition to the well-known function of gangliosides as host cell receptors that mediate virus ( e . g . , polyoma and SV40 ) attachment and transport from the plasma membrane to the ER , we now demonstrate that they can also act as determinants for capture by mDCs . Furthermore , we identify a moiety composed of sialyllactose on HIV-1 membrane gangliosides as the specific domain recognized by mDCs . We propose that this novel recognition moiety might be crucial for inducing immune responses , but also critical to disseminate HIV-1 and other ganglioside-containing viruses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"flow",
"cytometry",
"immune",
"cells",
"viral",
"transmission",
"and",
"infection",
"antigen-presenting",
"cells",
"immunity",
"to",
"infections",
"immunology",
"microbiology",
"host-pathogen",
"interaction",
"immunodeficiency",
"viruses",
"viruslike",
"particles",
"immune",
"defense",
"cytometry",
"viral",
"immune",
"evasion",
"membranes",
"and",
"sorting",
"biology",
"pathogenesis",
"host",
"cells",
"cell",
"biology",
"immunity",
"virology",
"molecular",
"cell",
"biology"
]
| 2012 | Sialyllactose in Viral Membrane Gangliosides Is a Novel Molecular Recognition Pattern for Mature Dendritic Cell Capture of HIV-1 |
The kinetic rate constants of binding were estimated for four biochemically relevant molecular systems by a method that uses milestoning theory to combine Brownian dynamics simulations with more detailed molecular dynamics simulations . The rate constants found using this method agreed well with experimentally and theoretically obtained values . We predicted the association rate of a small charged molecule toward both a charged and an uncharged spherical receptor and verified the estimated value with Smoluchowski theory . We also calculated the kon rate constant for superoxide dismutase with its natural substrate , O2− , in a validation of a previous experiment using similar methods but with a number of important improvements . We also calculated the kon for a new system: the N-terminal domain of Troponin C with its natural substrate Ca2+ . The kon calculated for the latter two systems closely resemble experimentally obtained values . This novel multiscale approach is computationally cheaper and more parallelizable when compared to other methods of similar accuracy . We anticipate that this methodology will be useful for predicting kinetic rate constants and for understanding the process of binding between a small molecule and a protein receptor .
Molecular dynamics is a simulation technique that uses Newton’s or Langevin’s equations of motion in combination with a specified molecular bond structure , parametrized force fields , and a starting conformation of atomic positions and velocities in order to propagate the dynamics of atoms within a molecular system . Ensembles of conformations or trajectories can be sampled to estimate thermodynamic or kinetic quantities[20 , 35 , 36] .
All MD simulations were carried out using NAMD 2 . 9[51] . The MD FHPDs were made with the help of MDAnalysis[52] . All calculations were performed on the Gordon supercomputer at the San Diego Supercomputer Center , the Stampede supercomputer at the Texas Advanced Computing Center , and on local machines . MD simulations of the charged and uncharged spherical receptor simulations were prepared using a simple 40 Å x 40 Å x 40 Å TIP3P[53] water box , we placed a Cl- in the center of the box for the charged spherical receptor . Both systems contain approximately 7600 atoms . Na+ and Cl- parameters were obtained from the ions94 library of the AMBER ff03 forcefield[54] . The spherical receptor systems were minimized for 10000 steps to allow the water molecules to relax in relation to each other and to the Cl- . Both systems were then equilibrated for 20 ns at a constant temperature of 300K using the Langevin thermostat and constant pressure using the Langevin piston at 1 atm with a damping coefficient of 5 ps−1 . The Cl- was constrained to a stationary position in the center of the charged spherical receptor system . Following this equilibration , four copies were made of the systems , and a Na+ was placed at the milestones located at 7 Å , 8 Å , 9 Å and 10 Å from the center of the water box in the uncharged system ( Fig 2 ) , and from the Cl- in the charged system . Two additional milestones were also placed at 6Å and 11 Å . Waters clashing with the Na+ were removed . The system was once again allowed to minimize for another 5000 steps to relax the waters around the ions . Then the system was heated in 10 K increments up to 350 K and then reduced back to 300 K at 2 ps intervals each at constant volume . Then , in order to obtain an ensemble distribution , the systems were simulated at constant temperature at 300 K at constant volume for 20 ns . To this point , all ions have been constrained . In order to obtain a FHPD , 900 position/velocity configurations were uniformly chosen between the 2 ns and 20 ns marks in the ensemble simulations . Velocities were reversed , and the trajectories were allowed to propagate backwards . If the trajectory struck another milestone before re-crossing the one it came from , that trajectory was considered part of the FHPD . All members of the FHPD were then allowed to proceed with their velocities in the forward direction . Each transition event was monitored for future milestoning analysis . Reverse simulations were carried out using a special plugin for NAMD 2 . 9[55] , which allows velocities to be reversed at arbitrary timesteps . For comparison with the milestoning results , brute-force MD simulations were run and Smoluchowski theory was used to estimate a β , kon , and MFPT for the spherical receptor systems . All brute-force MD simulations were set up with the same parameters as for milestoning above , except that the system was equilibrated for 40 ns and 10000 frames were sampled between the 20 and 40 ns time . Each of the 10000 simulations were started with the Na+ placed on the 10Å milestone and monitored for a crossing event at either the 6Å or the 11Å milestone . The value β was simply the number that crossed the 6Å milestone out of the total number of simulations . The MFPT was the average amount of time that all the simulations lasted before a crossing event . MD force field ( FF ) parameters for SOD were obtained as a generous gift from Branco et . al . [56] The system was surrounded by a TIP3P[53] water box with 150 mM NaCl solution . The simulation contained approx . 44 , 000 atoms . The SOD system was then equilibrated for 80 ns at a constant temperature of 300 K using the Langevin thermostat and constant pressure using the Langevin piston at 1 atm using a damping coefficient of 5 ps−1 . Following equilibration , ten copies were made of the apo system , and O2− was inserted at eight different milestones ( located at 4Å-11Å in 1Å increments ) from each of the two copper ions in SOD’s two active sites , yielding a total of sixteen different milestones simulated ( Fig 3 ) . Waters clashing with O2− were removed . The solvent molecules in the system were minimized for another 5000 steps to relax around the newly placed ions . Then the system was heated in 10 K increments up to 350 K and then reduced back to 295 K at 2 ps intervals each at constant volume . The protein and O2− atom positions were constrained during the minimizations and heating/cooling . In order to obtain an ensemble distribution , the systems were simulated at a constant temperature of 300 K and constant volume for 200 ns each with an imposed harmonic “spring” force of 300 kcal mol−1 Å−2 that constrained O2− close to a spherical milestone at each system’s proper distance from the SOD active site catalytic copper . In order to obtain a FHPD , 700 position/velocity configurations were uniformly chosen between the 60 ns and 200 ns marks in the ensemble simulations . Velocities were reversed , and the trajectories were allowed to propagate backwards in time . If the trajectory struck another milestone before recrossing the one it came from , that trajectory was considered part of the FHPD . The autoimage function in CPPTraj[57] was used to center the ligand in the waterbox before the reversal stage . All members of the FHPD were then allowed to proceed in the forward direction . Each crossing event was monitored for future analysis . The reversal phases were simulated using a custom plugin for NAMD 2 . 9[55] . FF parameters for TnC were prepared according to the protocol followed by Lindert et . al . [58] The system was surrounded by a TIP3P[53] waterbox with 100 mM KCl solution . The simulation contained approximately 27 , 000 atoms . The TnC system was then equilibrated for 100 ns at a constant temperature of 288 K using the Langevin thermostat and pressure using the Langevin piston at 1 atm using a damping coefficient of 5 ps−1 . Following this equilibration , twelve copies were made of the systems , and the Ca2+ was inserted on the binding side of the TnC site II loop at 1 Å increments from 2 Å to 9 Å from the center of mass of the alpha carbons of residues ASP 65 , ASP 67 , SER 69 , THR 71 , and GLU 76 ( Fig 4 ) . Waters clashing with Ca2+ were removed . The solvent molecules in the system were minimized for another 5000 steps to relax around the newly placed ions . Then the system was heated in 10 K increments up to 350 K and then reduced back to 295 K at 2 ps intervals each at constant volume . The protein and Ca2+ atoms were constrained during the minimizations and heating/cooling cycles . In order to obtain an ensemble distribution , the systems were simulated at a constant temperature of 300 K and constant volume for 100 ns each with an imposed harmonic force of 300 kcal mol−1 Å−2 that constrained Ca2+ close to the spherical surface at each system’s proper distance from the active site center of mass . In order to obtain a FHPD , 700 position/velocity configurations were uniformly chosen between the 30 ns and 100 ns marks in the ensemble simulations . The reversal phase of the TnC system was performed in an identical procedure as the SOD system . All Brownian dynamics simulations were performed using BrownDye[27] with desolvation forces and hydrodynamic interactions activated . All electrostatics calculations were performed using the Poisson-Boltzmann Equation solver APBS[59] . The solvent dielectric was left at the default of 78 , and the permittivity of a vacuum was left at the default of 8 . 854×10−12 C2N−1m−2 . All macromolecular dielectrics were set to 2 , while the dielectrics of Ca2+ and O2− were set to 1 . A 6–12 hard sphere Lennard-Jones interaction was used . Simulations were distributed across 10 to 20 threads on a local computing node . The BrownDye program bd_top was used to prepare all systems for simulation . A phantom atom of zero charge and zero radius was placed at the center of the active sites in order to detect crossings of spherical milestones . The phantom atom has no effect on the dynamics , but is merely a convenient way to detect surface-crossing events . The BrownDye program nam_simulation was used for simulation , and the program compute_rate_constant was used to aid in the calculation of the association rate constants . Trajectories were processed using the BrownDye programs process_trajectories and xyz_trajectory in combination with in-house Python scripts . A PQR file for SOD was prepared from the crystal structure PDB ID: 1CBJ[60] using LEaP[61] and DelEE[62] with the AMBER forcefield[63 , 64] and PROPKA , 47 assigned protonation states at a pH of 7 . 0 . A PQR file for O2− was made by hand , with each oxygen given a partial charge of -0 . 5 and a radius of 1 . 5 Å . APBS[59] was then used to calculate the electrostatic field at 295 K and a NaCl concentration of 150 mM to approximate conditions used during the experimental measurement of kon for SOD[65] . BrownDye was used to prepare and run 1×106 BD simulations at 295 K with the ligand starting from a b-surface at ~61 Å from the SOD center of mass . Based on experimentally determined diffusion coefficient[66] of 1 . 5×10−5 cm2s−1 , a hydrodynamic radius of 1 . 45 Å was used for O2− in the simulations ( See S1 Text ) . We used the Browndye default water viscosity of 1 . 00×10−3 kg m−1s−1 for all BD simulations of SOD . Reactions with both active sites , and also escape events were counted . 1000 configurations of ligand encounters with both active sites ( 12 Å from catalytic copper ) were extracted to make two additional FHPD distributions . 1000 simulations were started from each configuration ( 2×106 total ) . These were allowed to react with a surface further down the site ( 11 Å from the catalytic copper ) react with the surface around the other site ( 12 Å from the other catalytic copper ) or escape to infinity . All reaction and escape events were counted to construct the statistics of the transition kernel K and incubation time vector 〈t〉 . A PQR file for TnC was prepared from the NMR structure 1SPY[67] . Partial charges were assigned according the AMBER forcefield[61] using LEaP [63]and DelEE[62] and PROPKA[64] assigned protonation states at a pH of 7 . 0 . A PQR file for Ca2+ was made by hand , given a charge of 2 . 0 e and an atomic radius of 1 . 14 Å . APBS[59] was then used to calculate the electrostatic field at 288 K and a KCl concentration of 100 mM to approximate conditions used during the experimental measurement of kon and koff for TnC[68] . A hydrodynamic radius of 3 . 0 Å was assigned based on an experimentally determined diffusion coefficient[69] of 6 . 73×10−6 cm2s−1 at 291 K ( See SI S1 Text ) . BD simulations of TnC used an experimentally determined water viscosity of 1 . 138×10−3 kg m−1s−1 at 288 K[70] . BrownDye was used to prepare and run 1×106 BD simulations at 288K with the ligand starting from a b-surface at ~57 Å from the TnC center of mass . Diffusion to the active site surface , and escapes were counted . 1000 configurations of ligand encounters with the active site ( 10 Å from binding site center of mass of residues ASP 65 , ASP 67 , SER 69 , THR 71 , and GLU 76 ) were extracted to make a FHPD distribution . 1000 simulations were started from each configuration ( 1×106 total ) . These were allowed to react with a surface further down the site ( 7 Å from binding site center ) or escape to an infinite distance . All reaction and escape events were counted to construct the milestoning model . For our spherical receptor calculations , we used a dielectric of 92 to mimic the dielectric of TIP3P water[71] , a permittivity of 8 . 854×1012 C2N−1m−2 , and a diffusion coefficient[69] of 1 . 33×10−5 cm2s−1 for Na+ . Although the dielectric of 92 for water is obtained from MD and was not experimentally measured , the spherical receptors were intended more for demonstration purposes rather than physical realism , and a dielectric of 92 was chosen in an attempt to allow the values obtained using Smoluchowski theory to match what we observe in the brute-force and milestoning MD simulations . The rate constants k ( a ) , k ( b ) , and k ( q ) were calculated using Eq 9 for the uncharged spherical receptor and Eq 10 for the charged spherical receptor for the reaction surface , b-surface , and q-surface , respectively . The rate constant k ( a ) is the theoretical model of the spherical receptor association . For comparison , we deduced k ( a ) using only k ( b ) , and k ( q ) by using a transition matrix K obtained from monitoring transitions of the spherical receptor systems in a series of MD simulations . A binding probability β was calculated using Eq 8 . The kon for each spherical receptor system was calculated using Eq 12 . The MFPT represents the mean time taken by a particle started on the b-surface and allowed to diffuse before touching either the reaction surface or the q-surface . The MFPT was calculated using Eq 12 . The values k ( b ) and k ( q ) are obtained using Eq 9 or Eq 10 , depending respectively on the absence of presence of a receptor charge . For each system , the milestoning calculations were performed using custom scripts that used Numpy 1 . 7 , Scipy 0 . 9 . 0 and the GNU Parallel tool[72] .
The total computational cost of all systems simulated in this study for both MD and BD was approximately 65 , 000 CPU hours . Computational costs of each simulated system and simulation regime are listed in Table 4 . The cost of performing all non-simulation calculations was negligible . Table 4 includes all computer time spent on the supercomputer as well as on local machines . The β , kons and error estimates for all systems were well converged and are reported in the SI ( S2–S9 Figs ) .
The kon calculated using milestoning for the uncharged spherical receptor system matches within 3% to the theoretically determined value and 0 . 3% to the brute-force MD value . These estimates are well within the bounds of uncertainty introduced by the milestoning model . As a system that can diffuse freely without forces or solvation shells , it is expected that Smoluchowski theory would yield such a close result to simulation . This similarity to a value obtained using well-established theory is a good validation of our basic methodology . The large difference between the MFPT predicted by theory and the MFPTs predicted by milestoning and brute force MD could be due to a difference between the experimentally measured diffusion coefficient of Na+ , and the diffusion coefficient that is observed in an MD simulation using the AMBER forcefield . The kon calculated using milestoning for the charged spherical receptor system differs by 13% from the kon predicted by Smoluchowski theory and by only 6% from the kon obtained by brute force MD simulation . This difference between the simulation-obtained values and the value obtained by theory is likely due to effects caused by the explicit solvent in our simulations , for which this simple implementation of Smoluchowski theory does not account . Very likely , solvation shells have formed around the Cl- placed in the center of the system , as well as the diffusing Na+ . Solvation shells create unevenness in the potential of mean force and the position-dependent diffusion coefficient of Eq 4 . As such , using Coulomb’s law for the electrostatic potential and a constant diffusion coefficient may not be sufficiently valid assumptions for ions in solution at such close proximity . Previous studies on close NaCl ion pair interactions in dilute solvent show oscillations in the mean force potential of the interionic distance that extend several molecular layers into the solvent[78–80] . Accounting for these factors and using an alternative solution to Eq 4 would likely result in a calculated value much closer to what we obtained using milestoning and the brute force MD . The fact that the milestoning results and the brute-force MD results are so similar supports the validity of the milestoning methodology . Similarly , with the charged receptor , the large difference in the MFPT predicted by theory and the MFPTs predicted by milestoning and brute force MD could be due to a difference between the experimentally measured diffusion coefficient of Na+ , and the diffusion coefficient that would be observed in an MD simulation using the AMBER forcefield . It could also be due to the same effects observed on β caused by the aforementioned solvation shells . SOD is an enzyme found in a wide variety of organisms[73] . It is a homodimer that makes use of a catalytic copper bound in its active site to catalyze the dismutation of the superoxide ion O2− into O2 and H2O2 [65 , 73] . SOD was the subject of many early enzymology experiments[81] and ligand-receptor binding simulations[38 , 82] . The SOD kon estimated using milestoning is within a factor of ~1 . 5 of the experimentally measured kon that this study attempted to emulate . Although this value falls outside the uncertainty bracket calculated for the milestoning model , it is still within the range of kons measured in other studies[73] . The kon we calculated is also close to the value obtained by Luty , et . al . in their seminal study of SOD kinetics[30] . It is well understood that a higher salt concentration slows the rate of O2− binding to SOD[73] . Therefore , the kon measured in this study is likely smaller than the value measured by Luty , et . al . because they simulated MD and BD with a solvent salt concentration of zero . The discrepancy could also be due to differences used by Luty et . al . in their implementations of atomic constraints on the protein , different boundary conditions in the MD phase , and the lack of desolvation forces in the BD phase . While it is not clear how much error is introduced by using an equilibrium distribution across the milestones , our use of a FHPD should , theoretically , provide a more accurate treatment due to its consistency with formal milestoning theory[18 , 19] . The insertion of additional states in the MD region also allowed us to obtain much better sampling of transition events than would be available for a comparable computation time if the MD region had been composed of only a single milestone . The FHPD of SOD at 12Å ( Fig 7 ) indicates that O2− approaches directly from the solvent and does not seem to sample much of the protein surface before entering the active site . Although a kon has already been obtained for this system by Luty et . al . using similar methods , our approach offers a number of key improvements and more closely resembles the experimentally obtained rate constant; both insofar as the conditions that the system was exposed to , as well as the final result . In order to try this milestoning method on a new system , we also calculated the kon of TnC . The troponin complex is a set of proteins that regulates muscle contraction in skeletal and cardiac muscles[67 , 68 , 75] . One of the subunits , TnC is attached to the thin filaments of a muscle fiber , and regulates the binding of Ca2+ to the N-terminal domain of TnC[83] . Ca2+ binding triggers changes within the complex , allowing myosin to latch onto the thin filaments and induce muscle contraction . TnC has been extensively studied due to its critical involvement with heart function and failure , and has been marked as a therapeutic target in heart disease and other disorders[68] . Our method is able to determine the kon to a value that is within a factor of ~5 of the experimentally measured kon that our study attempted to emulate . Although this discrepancy falls outside of both the experimental uncertainty as well as the uncertainty of the milestoning calculation , the value is not unreasonable when compared to kon values measured in other studies[74 , 75] . The FHPD of TnC at 10 Å ( Fig 8 ) indicates that Ca2+ approaches directly from the solvent , probably due to the high desolvation penalty incurred when the highly charged Ca2+ is removed from its aqueous environment . The surface map seems to indicate two close but distinct minima on the FHPD , suggesting that Ca2+ may have two possible routes to binding ( Fig 8 ) . In total , the entire project , including all simulations of all systems analyzed in this study , cost approximately 65 , 000 hours of CPU usage . The vast majority of this computation was spread across hundreds or thousands of cores at any one time due to the highly parallel nature of milestoning . The total length of MD simulation for our systems required anywhere between 100 and 1600 ns of total MD time each with relatively low uncertainty due to the high rate of sampling along the milestones leading to binding . The cost is significantly less per target than brute force MD simulations run in past studies to observe kinetic events while yielding similar or superior resemblance to experiment[4 , 5] , which were indicated to require between 600 and 15000 ns of MD simulation to achieve even just a single binding event , with some simulations never even yielding a binding event . Our multiscale MD-BD-milestoning method offers many advantages; yielding predictive kon estimates for biologically relevant molecular systems within a range of experimental measurements at a cost much less than brute-force MD alone and at accuracy much greater than could be obtained using BD alone . This method also benefits from high parallelism due to the spread of MD computation across multiple states . Given a large number of cores and sufficient CPU hours , the MD portion of the calculation can be completed rapidly . Another advantage of this method is its flexibility , giving the user the ability to adjust the cost-to-accuracy balance by performing additional simulation and adding trajectory samples to increase result convergence . Theoretically , this milestoning framework could be used to investigate any biomolecular association where MD and BD simulation methods can adequately model the process . Estimating the binding kinetics between proteins , DNA , small molecules , or any combination thereof ought to be possible , assuming that sufficient sampling effectively constructs the proper FHPD . The main disadvantage of this method lies in its complexity of concept and implementation ( Fig 9 ) , particularly in the maintenance of large numbers of simultaneous simulations . However , with sufficiently robust software-based automation , the burden of maintaining many parallel instances of simulation , as well as simulation preparation and analysis , can be greatly reduced . Another disadvantage of the milestoning framework is that the simulations are still relatively expensive at this time; requiring a supercomputer or cluster to obtain sufficient sampling within a reasonable time frame , although GPU-based MD could potentially alleviate this burden . We present a new method to estimate kinetic rates . This method uses milestoning to leverage the strengths and minimize the weaknesses of MD and BD , thereby offering an efficient , high-accuracy estimation of kon rate constants . This multiscale method has been successfully used to estimate the kon rate constant for both idealized and realistically sized , biologically relevant systems . Our work demonstrates that milestoning can be used to obtain kinetic quantities of interest with a high resemblance to experiment . We anticipate that this multiscale approach can be used to determine rate constants of interest as well as system-specific binding details that are applicable to drug discovery , biomolecular modeling , and protein-ligand interactions . | We estimated the kon rate constant of four biochemically relevant ligand-receptor systems using milestoning theory . All results closely resemble experimentally and theoretically determined results , indicating that this technique may be applied toward accurate estimation of binding rate constants for additional ligand-receptor systems of biomedical interest . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | Multiscale Estimation of Binding Kinetics Using Brownian Dynamics, Molecular Dynamics and Milestoning |
Mutations in GDAP1 , which encodes protein located in the mitochondrial outer membrane , cause axonal recessive ( AR-CMT2 ) , axonal dominant ( CMT2K ) and demyelinating recessive ( CMT4A ) forms of Charcot-Marie-Tooth ( CMT ) neuropathy . Loss of function recessive mutations in GDAP1 are associated with decreased mitochondrial fission activity , while dominant mutations result in impairment of mitochondrial fusion with increased production of reactive oxygen species and susceptibility to apoptotic stimuli . GDAP1 silencing in vitro reduces Ca2+ inflow through store-operated Ca2+ entry ( SOCE ) upon mobilization of endoplasmic reticulum ( ER ) Ca2+ , likely in association with an abnormal distribution of the mitochondrial network . To investigate the functional consequences of lack of GDAP1 in vivo , we generated a Gdap1 knockout mouse . The affected animals presented abnormal motor behavior starting at the age of 3 months . Electrophysiological and biochemical studies confirmed the axonal nature of the neuropathy whereas histopathological studies over time showed progressive loss of motor neurons ( MNs ) in the anterior horn of the spinal cord and defects in neuromuscular junctions . Analyses of cultured embryonic MNs and adult dorsal root ganglia neurons from affected animals demonstrated large and defective mitochondria , changes in the ER cisternae , reduced acetylation of cytoskeletal α-tubulin and increased autophagy vesicles . Importantly , MNs showed reduced cytosolic calcium and SOCE response . The development and characterization of the GDAP1 neuropathy mice model thus revealed that some of the pathophysiological changes present in axonal recessive form of the GDAP1-related CMT might be the consequence of changes in the mitochondrial network biology and mitochondria–endoplasmic reticulum interaction leading to abnormalities in calcium homeostasis .
Charcot-Marie-Tooth ( CMT ) disease is one of the most frequent inherited neurological disorders and is characterized by either demyelinating or axonal neuropathy of motor and sensory peripheral nerves [1–3] . More than eighty genes are related to the pathogenesis of CMT and other peripheral neuropathies affecting a wide number of biological functions and cellular pathways [4 , 5] ( http://neuromuscular . wustl . edu/time/hmsn . html ) . Mutations in GDAP1 , which maps at human chromosome 8q21 . 1 , are causative for several types of neuropathy and are transmitted through various modes of inheritance including autosomal recessive demyelinating CMT4A with reduced nerve condition velocities ( NCVs ) [6] , autosomal recessive axonal AR-CMT2K with preserved NCVs and abnormal compound motor action potentials ( CMAPs ) [7] , and the less frequent autosomal dominant CMT2K [8 , 9] and recessive intermediate RI-CMT [10] . GDAP1 is a protein of 358 amino acids located at the mitochondrial outer membrane ( MOM ) [11 , 12] and has two glutation S-transferase ( GST ) domains , with a protein interacting α4-α5 loop domain between the two GST domains , and one transmembrane C-terminal domain that is important for the correct anchoring of the protein to the MOM [13] . GDAP1 is mainly expressed in neurons [14] but expression in Schwann cells has also been reported [11] . GDAP1 plays a role in the regulation of mitochondrial dynamics inducing organelle fission [11 , 12] . Whereas loss of function recessive mutations are associated with decreased fission activity , dominant mutations result in impairment of mitochondrial fusion with increased production of reactive oxygen species ( ROS ) and susceptibility to apoptotic stimuli [15] . Through characterization of fibroblasts derived from CMT4A patients , Noack and colleagues [16] observed that GDAP1 is implicated in the control of the cellular glutathione content and mitochondrial activity , supporting the role of GDAP1 in the defense against oxidative stress . We have recently demonstrated that GDAP1 silencing in the neuroblastoma SH-SY5Y cells reduces Ca2+ inflow through store-operated Ca2+ entry ( SOCE ) upon mobilization of endoplasmic reticulum ( ER ) Ca2+ , likely in association with the abnormal distribution of the mitochondrial network [17] . Thus , the pathophysiology of GDAP1-related neuropathies may involve complex cellular interactions between mitochondrial network and ER , oxidative stress and Ca2+ homeostasis , associated with mitochondrial dynamics and distribution . Both demyelinating [6 , 18] and axonal [7 , 19] lesions have been reported in the nerve biopsy from GDAP1-related CMT patients . However , electrophysiological studies indicate that especially in the more severe patients the GDAP1 mutations induce primary axonal neuropathy with secondary demyelination [19 , 20] . Data of mitochondrial morphology in biopsy specimens are scarce , but in those cases when biopsy data were presented no abnormalities were reported [1 , 21] . To investigate the functional consequences of loss of GDAP1 function in vivo , we generated a Gdap1 knockout mouse ( Gdap1-/- ) . Here , we describe characterization of this model and report that the Gdap1-/- mice display neuropathic behavior and symptoms closely similar to those of AR-CMT2K patients . Detailed characterization of the generated in vivo model allowed us to propose a pathogenic mechanism of the disease that involves depletion of cellular calcium associated with changes in the mitochondrial network .
To generate the Gdap1 knockout mice ( Gdap1-/- ) we deleted exon 1 of Gdap1 by homologous recombination . This was achieved by using a targeting vector containing a self-excising loxP-exon 1 Gdap1-FRT-PGK-NeoR-FRT-loxP cassette ( Fig 1A ) , which allows to generate a recombined locus by deletion of the selectable marker Neo . The targeting construct was electroporated into 129sv stem cells and 2 recombinant clones were identified by both PCR and Southern blotting . Gdap1-/- mice were established by breeding heterozygous Gdap1+/flox mice with CMV-Cre deleter mice leading to ubiquitous exon 1 deletion . Gdap1-/- mice are born at expected Mendelian ratios , have normal life-span and are fertile . Western blot analysis of neuronal and non-neuronal tissue lysates of Gdap1+/+ ( WT ) , Gdap1+/- and Gdap1-/- littermates mice confirmed the absence of GDAP1 protein ( Fig 1B ) . The severe recessive form of GDAP1-related CMT starts early in infancy or childhood with weakness and wasting of the feet followed by involvement of the hands leading to pronounced disability . Patients are usually wheelchair bound starting from the second decade of life . Thus , we compared motor behavior between Gdap1-/- and WT mice at the early adult age . We observed abnormal hind-limb clasping reflex in Gdap1-/- mice at the age of 3 months ( Fig 2A , upper panels ) . When observed during locomotion , most animals showed very low position of the body and dragging tail indicating presence of motor deficits ( Fig 2A , lower panels ) . Then we tested motor and coordination behavior by rotarod test in several ages . In order to avoid previous learning we used different animals for each time point . We detected significantly reduced latency to fall in 3 months-old mutant mice . The persistence of this motor behavior was confirmed in older Gdap1-/- mice from 4 to 7 months of age but became more variable and thus non-significant after 9 months of age ( Fig 2B ) . In order to further characterize motor phenotype of Gdap1-/- mice we analyzed their gait behavior at the age of 5 and 12 months ( Fig 2C ) . At both time-points Gdap1-/- mice had shorter stride length ( p<0 . 05 at 5 months and p<0 . 01 at 12 months , Student’s t test ) and at 12 months we also detected more narrow gait angles ( p<0 . 001 , Student’s t test ) ( Fig 2D , upper panels ) . As we observed everted paws in some mutant animals , we also analyzed the plantar print ( Fig 2D , lower panels ) . At the age of 12 months we detected significant decrease of toe spreading and an increased extension of plantar length that suggest a defect on the plantar muscle strength . The observed defects in motor behavior in mice lacking GDAP1 prompted us to further investigate their electrophysiological properties along the sciatic nerve . Therefore , we measured compound muscle action potential ( CMAP ) amplitude ( Fig 2E ) and the motor nerve conduction velocity ( MNCV ) ( Fig 2F ) . In line with the absence of detectable behavioral phenotype , we were not able to detect any electrophysiological differences between Gdap1-/- and WT mice at 2-months of age . Interestingly , at 5-months Gdap1-/- mice showed a significant reduction of CMAP amplitude obtained for distal ( at the ankle ) and proximal ( at the hip ) stimulation , indicative for an axonal neuropathy ( Fig 2E , p<0 . 05 , Student’s t test ) . Concomitantly , at this age , we observed a slight but significant reduction in MNCV for Gdap1-/- mice as compared to WT mice ( Fig 2F , p<0 . 05 , Student’s t test ) . However , we did not measure CMAP and MNCV at later time points to assess the persistence of this effect . Slight reductions in MNCV can be associated with axonal neuropathies [19] . In GDAP1-CMT patients , axonal neuropathy is associated with loss of axons in sural nerve biopsies [19 , 22] . We therefore decided to further investigate the nerve morphology of sciatic nerves from 5-months-old Gdap1-/- and age-matched WT mice , the age at which we detected electrophysiological changes compatible with presence of axonopathy ( Fig 2 ) . Surprisingly , we did not detect any significant reduction in the axonal number correlating with the observed reduction in CMAP amplitude in 5 months-old mice ( S1A and S1B Fig ) . Distribution of axon sizes in both proximal and distal sciatic nerves did not show any differences between Gdap1-/- and WT mice ( S1C Fig ) . Moreover , myelin thickness was preserved in 5-months old Gdap1-/- mice as well . By measuring g-ratio ( ratio of axon diameter versus fiber diameter ) in proximal and distal sciatic nerves of both Gdap1-/- and WT mice we confirmed the presence of normal myelin sheath with no myelination defects or myelin reduction associated with GDAP1 defective nerves ( S1D Fig ) . Axonal neuropathy revealed by our electrophysiological studies in 5-months-old Gdap1-/- mice could be associated with defects in other motor neuron-derived structures affected by the absence of GDAP1 . Thus , we examined motor neuron ( MN ) numbers and morphology in the anterior horn , as well as the neuromuscular junction ( NMJ ) structure in Gdap1-/- mice at different ages . MNs in the anterior horn showed cell lesions such as vacuoles and chromatolysis in Gdap1-/- mice at 5 and 12 months of age ( Fig 3A ) . We found statistically significant difference in the number of healthy MNs between knockout and control mice at 5 and 12 months of age . However , while in control mice the number of healthy MNs per section underwent a progressive reduction over time ( 2-months vs 12-months WT animals , p-value<0 . 05 ) , this process is enhanced in Gdap1-/- mice as indicated by the slopes of neuron loss between 2 and 5-months ( Gdap1-/-: 0 . 71 versus WT: 0 . 333 ) . By contrast , slopes were equal between 5 and 12-months , which suggest that loss of neurons occurs early in the first months of life in the Gdap1-null mice ( Fig 3B ) . These data suggest that MN degeneration in the anterior horn may contribute to the abnormalities revealed by CMAP measurements in 5-months Gdap1-/- mice . Effects of pathological and functional changes in CMT diseases are related to the axonal length , so distal limb muscles are usually the first to be affected . This is the case for patients with GDAP1 mutations , especially those carrying two recessive mutated alleles [23] . Thus , we investigated the proper occupancy and morphology of NMJs in both Gdap1-/- and control mice by staining axonal presynaptic buttons with an antibody against β-III tubulin and the nicotinic acetylcholine receptor with α-bungarotoxin . Occupancy of the NMJs was defined as the percentage of postsynaptic buttons that are completely filled by the nerve terminal ( Fig 3C ) . We did not observe any differences in the NMJ occupancy at 5-months . However , whereas NMJs from 12-month-old control mice were fully occupied the percentage of occupancy in the NMJs from Gdap1-/- was significantly reduced ( Fig 3D ) . In addition , we also observed abnormal tangle-like structures in terminal nerves from knockout animals that were not present in control mice ( Fig 3E ) . We decided to further investigate the consequences of the absence of GDAP1 in primary neuronal cultures . After being cultured for 24 hours , Gdap1-null DRG sensory neurons from 5-months old mice were able to generate neurites but we observed a significant reduction of neurite average length in sensory neurons of Gdap1-/- animals ( Fig 4A ) and in general longer processes in WT versus Gdap1-/- mice ( Fig 4B ) . Embryonic MN cultures derived from both genotypes had similar capacity to generate axons at 24 hours ( length between 200 and 250 μm ) . However , at 48 hours we observed substantial difference between the two genotypes ( Fig 4C ) : whereas axons in cultures generated from the knockout mice have maximal length of around 250 μm , axons from wild-type mice showed continuous growth , some of them reaching 600 μm . We have previously demonstrated that GDAP1 interacts with vesicle transport proteins RAB6B and caytaxin , and with β-III tubulin [17 , 24] . Interaction with β-III tubulin prompted us to investigate the effect of GDAP1 absence on the cytoskeleton of mouse MNs and DRG neurons . Microtubules ( MTs ) are dynamically assembled polymers of α- and β-tubulin that are present in all eukaryotic cells . Post-translational modifications ( PTMs ) that occur on MT are crucial in order to maintain their dynamic stability . Many different PTMs have been reported to occur on tubulin [25] . Among them , tubulin acetylation has been widely related to neurodegenerative diseases [26] , in particular with CMT [27] . Thus , we analyzed lysine-ε-acetylation of α-tubulin in a 24-hr cell culture system . We observed abnormal PTM of microtubules: α-tubulin showed significant decreased acetylation in both neurites of sensory neurons ( Fig 5A ) and MN axons ( Fig 5B ) . These findings suggested a defect of the cytoskeleton associated with the absence of GDAP1 in mitochondria . To get more information about the effect of loss of GDAP1 in the cell biology of neurons we investigated the cell and organelle ultrastructure in 24-hr cell cultures of both WT ( Fig 6A ) and knockout MNs ( Fig 6B–6D ) . For both genotypes our cultured MNs showed neurites and putative axonal prolongations . Control MNs showed irregular and invaginated nuclei , reduced ER cisternae , large number of free ribosomes , well-defined microtubules network , and polarized mitochondrial distribution ( Fig 6A ) . In MNs from Gdap1-/- mice we could distinguish abnormal organelles structures with two major patterns that may be interpreted as consecutive steps of the same pathological process: ( i ) neurons with dispersed round or spheroid mitochondria and apparent increase of mitochondrial metabolism , enlargement and dilatation of the nuclear envelope and perinuclear space , and significant increase number of small and medium-sized vacuoles that may correspond to dilated ER cisternae ( Fig 6B ) ; ( ii ) more severely affected neurons with large and tubular mitochondria with swollen cristae that are oriented along the longitudinal axis , very large vacuoles with invaginations that are originated from the nuclear envelope and from the small-medium vacuoles , phagolysosomes , and some autophagosomes; it seems that ER cisternae are more compacted with reduced internal light ( Fig 6C and 6D ) . These observations suggest progressive cell pathology associated with the lack of GDAP1 in motor neurons . Then , we characterized in detail morphological parameters from more than 900 WT and Gdap1-null mitochondria ( Table 1 ) . There were no differences in the number of mitochondria per cell , surface area and perimeter of mitochondria between both genotypes . Further analysis of mitochondrial shape parameters indicated that whereas Feret diameter and roundness showed no differences , there were significant reduction of circularity and increase of aspect ratio in Gdap1-null mitochondria . It seems that the absence of GDAP1 induces larger mitochondria that may represent the increase of fusion activity of the mitochondrial network . Since GDAP1 mutations have been associated with changes in the dynamics of the mitochondrial network and we detected morphological changes in mitochondria from somas of Gdap1-/- MNs ( Fig 6 ) we decided to characterize mitochondrial phenotype further . We investigated mitochondrial parameters in both proximal ( first 10 μm after neuron body ) and distal ( last 10 μm ) segments of MN axons cultured for both 24 hours and 48 hours by detecting mitochondria as cytochrome c positive organelles ( Fig 7A and 7B ) . At 24 hours WT mice showed homogenous distribution in proximal and distal axonal segments . On the contrary , in Gdap1-null mice the number of mitochondria was significantly increased at 24 hours in the proximal segment ( Fig 7A ) , which could be interpreted as a delay of mitochondrial distribution along the axon in the absence of GDAP1 . Mitochondrial distribution normalized after 48 hours in culture . In parallele , we observed a reduction of the mitochondrial interconnectivity that was evident in the proximal segment after 24 and 48 hours in culture and in the distal segment after 48 hours ( Fig 7B ) . The increased number of mitochondria in the proximal segment at 24 hours and reduction of organelles interconnectivity suggests an anomalous distribution of the mitochondrial network associated with the lack of GDAP1 . We then decided to validate these data in vivo , by analyzing the effect of GDAP1 absence on the structure and number of mitochondria in the proximal and distal regions of sciatic nerve . Whereas mitochondrial morphology was unaffected in Gdap1-/- nerves , the number of mitochondrial particles was significantly increased both proximally and distally ( Fig 7C ) . To know whether such an apparent increase could represent a true increase of the number of mitochondria or may reflect the increase of the mitochondrial mass we determined the number of mtDNA copies in sciatic nerves . We did not detect any significant difference on the mtDNA copy number , and hence mitochondrial mass , between Gdap1-/- and WT nerves ( Fig 7D ) . Motor and sensory neuropathies are frequently a consequence of mitochondrial defects associated with changes in mitochondrial respiratory chain and oxidative phosphorylation , that affect the energetic metabolism of peripheral nerves . To evaluate if similar type of changes are present in Gdap1-/- animals . we performed protein expression analysis on selected mitochondrial proteins at the age of 5 months . Extracts from skin , skeletal muscle , liver , brain , cerebellum , spinal cord and peripheral nerves were prepared and printed onto Reverse Phase Protein Microarrays ( RPPM ) to quantify the expression of relevant enzymes of glycolysis ( GAPDH , PKM2 , LDHA ) , oxidative phosphorylation [OXPHOS , Complex I ( NDUFS3 ) , Complex II ( SDHB ) , Complex III ( Core 2 ) , Complex IV ( COXI , COXII and COXIV ) and Complex V ( β-F1-ATPase ) ] , mitochondrial dynamics ( MFN1 and MFN2 ) , β-oxidation of fatty acids ( HADHA ) , oxidative stress ( SOD2 and catalase ) and the cytoskeleton ( β-actin , used as loading control ) using validated antibodies ( S2 Fig ) . For most of the analyzed tissues ( liver , skeletal muscle , brain , skin , and spinal cord ) no relevant differences were observed in the expression of the proteins studied ( Fig 8A , ) . Interestingly , cerebellum , involved in balance and motor behavior , and especially peripheral nerves ( sciatic and axillary nerves ) showed significant differences in protein expression ( Fig 8B and 8C ) . Specifically , the cerebellum of Gdap1-/- mice revealed a diminished expression of several enzymes of glycolysis , OXPHOS and mitochondrial dynamics when compared to WT mice , suggesting a limited energy provision in this tissue ( Fig 8B ) . Even more significant changes were observed in expression of all markers of glycolysis , OXPHOS and mitochondrial dynamics as well as catalase of oxidative stress but not HADHA of β-oxidation in peripheral nerves of Gdap1-/- mice when compared to WT mice ( Fig 8C ) , which were confirmed by Western blot experiments in nerves ( S3 Fig ) . This strongly suggests that a bioenergetic compromise , because of dysfunctional mitochondria , may contribute to a dysfunction of peripheral nerves and most likely cerebellum in Gdap1-/- mice . Intriguingly , differences on peripheral nerves were more significant in females , which may be related to the animal hormone environment . In SH-SY5Y cells we have observed that GDAP1 depletion decreases SOCE activity and impairs Ca2+ entry in mitochondria following activation of store-operated calcium entry [17] . This might be related to the altered distribution of the mitochondrial network within the cell associated with abnormal movement of mitochondria along the cytoskeleton towards the ER and subplasmalemmal microdomains in neuroblastoma cells having residual GDAP1 expression . With our knockout mouse model we are now able to analyze the consequences of a complete lack of GDAP1 on Ca2+ levels in the cytoplasm in neurons and the effect on SOCE in vivo . We first studied cytoplasmic calcium ( [Ca2+]cyt ) transients triggered by emptying of ER-Ca2+ with thapsigargin ( TG ) in a Ca2+-free medium and the effect of Ca2+ readmission on SOCE activity in 24 hour-cultured motor neurons . Interestingly , neurons lacking GDAP1 showed significant lower [Ca2+]cyt than WT neurons before and after ER-Ca2+ emptying with TG ( Fig 9A ) . Both WT and Gdap1-/- cells responded to 2 mM Ca2+ by increasing [Ca2+]cyt , which suggest that SOCE mechanism is working . However , SOCE response was smaller with lower [Ca2+]cyt peak in Gdap1-/- neurons , which may reflect the reduction of Ca2+ in resting conditions ( Fig 9B ) . In addition , such reduced Ca2+ levels was also associated with slow entrance velocity of Ca2+ into the neuron ( Fig 9C ) . Low resting levels of [Ca2+]cyt and slow entry of Ca2+ after store Ca2+ depletion suggest that lack of GDAP1 induces a defect on the proper maintenance of intracellular Ca2+ concentration and homeostasis . To know if such a low cytoplasmic Ca2+ concentration may be due to reduced calcium stores in cells , we treated motor neurons with ionomycin , a strong ionophore that releases calcium from cellular stores . Neurons defective for GDAP1 showed significant lower increase of [Ca2+]cyt than WT neurons , which might affect ER-Ca2+ release or reduction of Ca2+ stores ( Fig 9D and 9E ) . These findings suggest that GDAP1 is necessary to maintain neuronal Ca2+ homeostasis . To know more on Ca2+ management within the ER , we further investigate the effect of lack of GDAP1 in the ER-Ca2+ physiology by measuring the calcium-handling proteins calreticulin , IP3R and BIP in both mouse sciatic nerve and lumbar spinal cord ( S4 Fig ) . Whereas Gdap1-/- mice did not showed any change of the expression levels of calreticulin and BIP , the absence of GDAP1 induced upregulation of IP3R . This change was significant in the sciatic nerve at the age of 5 months and spinal cord at the age of 12 months . Such an upregulation might be interpreted as a specific response to the [Ca2+]cyt depletion in MNs .
Mitochondria have a pivotal role in the pathogenesis of neurodegeneration [28] . Besides the classical role of defective oxidative phosphorylation in neurological disease there is increasing evidence of the relevance of mitochondrial trafficking and transport , inter-organelle communication—especially with the ER—and mitochondrial dynamics and quality control [28 , 29] . Peripheral nerve pathologies are observed in mitochondrial disorders and vice versa CMT has been associated with mutations in genes implicated in mitochondrial biology e . g MFN2 and GDAP1 [1] . Mutations in the GDAP1 gene cause both demyelinating and axonal CMT neuropathies . Most patients are homozygous or compound heterozygous for the mutation and express a severe early-onset of autosomal recessive CMT4A [6 , 18] or AR-CMT2K [7 , 19] phenotypes respectively . Yet milder axonal dominant variant ( CMT2K ) has also been reported [8 , 9 , 30] , which suggest different pathogenic consequences of the GDAP1 mutant protein and mechanisms of the neuropathy . GDAP1 is mainly expressed in neurons [14] , but expression in Schwann cells has also been reported [11] . Thus , a major question regarding the pathophysiology of GDAP1-related CMT is the primary injury target in the nerve , the axon or the Schwann cell . In order to gain insight into the cellular pathogenesis of GDAP1 defects and disease pathophysiology we have generated and characterized a knockout mouse model of recessive forms ( CMT4A/2K ) of GDAP1-related peripheral neuropathy . We observed multiple differences in motor behavior of Gdap1-/- mice . Knockout animals showed significant reduced running time on the rotarod starting at the age of 3 months , suggesting presence of defects in the balance and motor coordination . This difference disappeared at later time-points ( >9 months ) , likely as a consequence of the onset of age-related coordination impairments at control mice or because of the compensatory effects in CNS of the Gdap1 gene paralogue Gdap1l1 ( S5 Fig ) [31] . Gait analysis also showed changes in the motor behavior in 5- and 12-months animals as measured by the plantar footprint including stride length and angle . Similar features , even having earlier onset , have also been observed in mouse models of neurological disease involving motor neurons such amyotrophic lateral sclerosis [32 , 33] . Thus , the observed motor behavior deficits suggested that lack of GDAP1 leads to a peripheral neuropathy phenotype . Reduced CMAPs amplitude in both distal and proximal nerve segments from 5-months-old mice strongly suggested the axonal nature of the neuropathy . Unexpectedly , normal number of axons and axon diameter distribution revealed no evidence of morphological axonopathy as observed in patients [19 , 22 , 23] or signs of demyelination . However , physiological damages were revealed by proteomic studies of the energetic metabolism in peripheral nerves . Limited energy supply and reduction of mitochondrial mitofusins under the Gdap1-/- background further suggested the existence of an axonal defect associated with mitochondrial abnormal function . Such functional-morphological contradictions may reflect specific response to the lack of GDAP1 in mice that could be more subtle than in human beings or reflect differences between both species related to diseased nerve length and/or natural history [23] . In contrast with the unaltered morphology of nerves , Gdap1-/- mice showed evidence of pathological changes in both motor neuron somas and the NMJs . This phenotype may be interpreted in two ways: ( i ) changes in motor neurons might be either a compensatory response to the abnormal function of nerves that are not altered enough to show pathological changes , or lack of GDAP1 in mice may be expressed primarily as a motor neuron disease; or ( ii ) both proximal lesions at neuronal somas and distal changes at NMJs may be the consequence of mitochondrial dysfunction in dendrites and distal synapsis where mitochondria are more present . GDAP1 is located at the MOM and has originally been previously show to play a role in mitochondrial dynamics [11 , 12 , 34] . The increase in mitochondrial density observed in both proximal and distal axons of Gdap1-/- mice can be caused by various changes in mitochondrial biology: alteration in mitochondrial fission-fusion balance , increase of the total mitochondrial mass or mitochondrial transport defects . Since we observed that in sciatic nerves of Gdap1-/- mice the total mitochondrial mass is not different between WT and knockout mice , we favor the hypothesis of a defective mitochondrial dynamics . Interestingly , an increase in the number of mitochondria in peripheral axons , especially in those with diameters smaller than 3 . 5 μm , has also been observed in Mfn2 transgenic mice which is a mitochondrial fusion-related protein that is also located in the MOM [35] . In addition , we also observed that the loss of GDAP1 is associated with lower interconnectivity of organelles found in primary culture of MNs , reduced axon growth capacity in both MNs and DRG sensory neurons , and a reduction of microtubule acetylation . Together , defects in the maintenance of the mitochondrial network and cytoskeleton along with abnormal energy metabolism might explain the electrophysiological changes in axons with preservation of the nerve architecture . More recent experimental data suggest a protective role of GDAP1 against oxidative stress related to intracellular levels of glutathione [16 , 36] . Lopez Del Amo et al . have validated the Drosophila GDAP1 ortholog ( Gdap1 ) [36] . In this model , Gdap1 RNAi produces progressive aggregation of the mitochondria , and eventually the presence of large elongated mitochondria in the fly thorax muscle , and at the retina mitochondria have larger size and tend to lose their peripheral localization . Moreover , we have shown that depletion of GDAP1 in the SH-SY5Y cells decreases SOCE activity and impairs SOCE-driven Ca2+ uptake in mitochondria , which may be linked to abnormal distribution of the mitochondria network [17] . In addition , GDAP1 is also present in mitochondrial-associated membranes ( MAMs ) [17] suggesting its function in the interaction between mitochondria and ER as it has also been proposed for MFN2 , the gene mutated in CMT type 2A [37] . Thus , available experimental data indicate that GDAP1 may link mitochondrial dynamics , ER and Ca2+ homeostasis regulation . In spite of these studies , the function of GDAP1 is not well established yet . Most of the recessive mutations predict truncated proteins with loss of the C-terminal domain that anchors GDAP1 to the MOM . Missense mutations may not affect the proper protein location but rather impair mitochondrial fusion and cause mitochondrial damage [15] . Our studies on Ca2+ homeostasis and SOCE activation in Gdap1-/- cultured MN showed reduced [Ca2+]cyt in both resting status and after pharmacological-mediated calcium depletion with thapsigargin or ionomycin . These findings suggest that the low levels of [Ca2+]cyt associated with the lack of GDAP1 could represent partial depletion of cellular calcium stores and/or release defects . Reduced Ca2+ stores and/or abnormal ER-Ca2+ release may impair proper stored-operated Ca2+ influx as well , which in turn may affect Ca2+ signaling homeostasis . These data in primary mouse MN and those we previously obtained in gene-silenced human neuroblastoma cells [17 , 38] suggest that GDAP1 plays a relevant role on the [Ca2+]cyt maintenance that can be , at least partially , mediated by SOCE . Thus , GDAP1 might mediate interactions between mitochondria and ER by positioning the mitochondrial network properly . The observation of large , dispersed and non-polarized mitochondria within cultured MNs in electronic microscopy experiments favors our hypothesis . Functional mitochondria are required to maintain SOCE activity [31 , 39] . The absence of mitochondria close to the Ca2+ microdomains may reduce the formation of SOCE channels between ER and plasmatic membrane and may reduce the Ca2+ entry operated by ER Ca2+ release . Calcium ions themselves control a wide number of cellular functions underlying cell signaling events . Abnormal calcium signaling has been associated with neurodegeneration [40] and increased [Ca2+] in the ER is observed in neurodegenerative diseases [41] but the pathological effects of reduced [Ca2+] in the ER is a less known phenomenon . Chronic depletion of cellular Ca2+ , either in ER or cytosol or both , affects the fluctuations and signaling of calcium ions and may induce pathological changes in mitochondria and ER and their physiological interaction which in turn may lead to the unfolding protein response ( UPR ) and activate the ER-stress-induced apoptosis pathway [42] . Difficulties to maintain the Ca2+ concentrations in cellular compartments and the abnormal cellular response may explain the progressive phenotype and neuropathology observed in our Gdap1-/- mouse model . Recently , Niemann et al . [43] reported an age-related hypomyelinating peripheral neuropathy in a Gdap1 knockout mouse model generated by deletion of exon 5 . These knockout mice developed a late-onset neuropathy with reduced nerve conduction velocity and changes in the nerve myelination . Morphometric studies revealed hypomyelination in 19-months-old knockout mice with no detectable axonal loss . In addition , sciatic nerve injury of knockout animals at the age of 2 months led to a less effective remyelination as compared with WT animals . The authors also found reduction in the CMAP amplitudes that may suggest an axonal component in the neuropathology phenotype . In contrast , our Gdap1-/- mouse model expressed an earlier onset in neuropathy-related signs as indicated by motor behavior abnormalities at the age of 3 months and electrophysiological changes in 5-months-old animals ( S1 Table ) . Since the same technological approach was used to generate knockout animals , the major genetic difference is the ablation of exon 1 in our case versus exon 5 . In both cases GDAP1 protein is absent on western blot analyses . However , we do not have sufficient information about consequences on transcription and RNA biology . These authors propose that mutations in Gdap1 lead to mild , persistent oxidative stress in the peripheral nervous system , which may be compensated in the central nervous system ( CNS ) by translocation of GDAP1L1 from the cytosol to mitochondria . We have studied Gdap1l1 expression in our Gdap1-null mouse , but we did not observe differences in brain , spinal cord and DRG between WT and Gdap1-null mice . ( S5 Fig , S1 Text ) . However , the strong expression of Gdap1l1 in brain of Gdap1-null mice , as in WT animals , may protect the brain structures from any pathological effect caused by the lack of GDAP1 . In conclusion , our data indicate that the lack of GDAP1 in mice lead to abnormalities of calcium homeostasis and changes in the axonal and neuronal physiology of the mitochondrial network and the ER . These results provide new insight into neuropathy caused by the loss of function GDAP1 mutations . The effects of GDAP1 recessive missense and dominant mutations on calcium signaling and the interaction with mitochondria and ER require further studies .
Vector construction and targeted knockout strategy was designed together with genOway ( Lyon , France ) , where mice were generated . The genomic region of murine Gdap1 locus was isolated from a 129Sv library/Pas , called rTgV , developed at genOway . The BAC clone collection rTgV was screened by PCR . The first pair of primers ( sense: 5`-CAG GGG AAC ATA ATC TGT GAG GAG GC-3 '; antisense: 5'-TCA CTA CTG GTG GTT CTT GTC AGC GC-3' ) amplifies a genomic fragment 409 bp gene GDAP1 upstream of exon 1 . The second pair of primers ( sense: 5'-GCG ACG GAA AAG CTC TAC CCT TAC C-3'; antisense: 5'-ACT GCA GTA GCA CTT GAG TGG CAG G-3' ) amplified a pb 553 of exon 2 and intron 2 Gdap1 genomic fragment . This molecular design led to the identification of a single BAC clone ( # 2956A2 ) , renamed clone FPA1-rTgV . DNA sequencing showed the Gdap1 homology regions of interest in that specific clone , which was used subsequently to isolate both FPA1-Long arm ( 5 . 7 kb ) as FPA1-Short arm ( 1 . 8 kb ) of target vectors . The target vector , FPA1-HR , contained two inserted loxP sites flanking exon 1 , the neomycin positive selection gene flanked by FRT sites and the presence of diphtheria toxin A ( DTA ) as a negative selection marker . Robust PCR screening strategy and Southern blot for detection of homologous recombination were designed . The FPA1-HR 129SvPas construct was transfected into ES cells according to standard electroporation procedures . Positive selection was started 48 h after electroporation , by adding 200 mg/ml G418 ( Life Technologies , Inc . ) . 179 resistant clones were isolated and amplified in 96-well plates . ES cell clones were screened by PCR to verify homologous recombination at the 3 'end of Gdap1 locus: sense 5'-GCC ACT CTC CAG ATG TTG AAA GGA G-3'; antisense 5'-TCA CTA CTG GTG GTT CTT GTC AGC GC-3' . a product of approximately 2 . 1 kb was expected . Of 179 clones tested , 3' homologous recombination was observed in 8 clones , which were verified by Southern blot analysis of 3' and 5' recombination events in Gdap1 locus . Two recombinant clones were microinjected into C57BL/6 blastocysts , giving rise to five male chimeras with significant contribution of ES cells ( positive agouti color ) were characterized . These male mice were bred with C57BL/6 Flp deleter female mice in order to cause cleavage of the germline neomycin selection cassette . Genotyping by PCR and Southern blot of pups derived from F1 breeding allowed identification of 6 heterozigous Gdap1 floxed mice heterozygous ( Gdap1+/flox ) . To generate a germline deletion of exon 1 female mice Gdap1+/flox mated with male C57BL/6 expressing CMV-Cre recombinase . Such a breeding allowed us the generation of heterozygous constitutive Gdap1 knock-out mice ( Gdap1+/- mice ) . Heterozygotes were matched to obtain homozygous Gdap1-l- and Gdap1+/+ littermate mice that were subsequently amplified by intercrossing for experimental work . All mice were maintained at 21 ± 2° C in cycles of 12 h light/dark with food and water ad libitum . Mitochondrial DNA copy number quantification was based on the mouse tissue-adapted ND1/ND4-quantification method [48] . Briefly , we quantified the mitochondrial 16S gene , starting at position mt2469 . The following primer/probe combinations were used for real-time PCR with LightCycler 480 Probes Master Mix ( Roche ) . Custom primer sequence for 16S: forward primer: 5' AAT GGT TCG TTT GTT CAA CGA TT 3' , reverse primer: 5' AGA AAC CGA CCT GGA TTG CTC 3' , and probe: FAM-5' AAG TCC TAC GTG ATC TGA GTT 3'-MGB ( Roche ) . Genomic DNA of a nuclear housekeeping gene , ANG1 ( Assay AppliedBiosystems: Mm00833184_s1 ) was also quantified using the same amount of DNA input . Relative mitochondrial DNA copy number was determined by comparing 16S gene to the nuclear endogenous control gene , also normalized to genomic DNA from blood . Wild-type and Gdap1-/- 5-month-old mice were sacrificed and their lumbar DRGs dissected out and collected in L15 media . Ganglia were incubated with 0 . 2% collagenase ( 30 min at 4°C and 1 h at 37°C; GIBCO ) followed by 0 . 05% trypsin ( 30 min at 37°C; GIBCO ) and 1% DNase ( 5 min at RT; Sigma ) treatments , and further dissociated by passing several times through a Pasteur pipette . After washing in F-12 medium , cells were diluted in F-12 medium supplemented with 200 mM glutamine , 60 ng/ml progesterone , 16 μg/ml putrescine , 400 ng/ml L-thyroxine , 38 ng/ml sodium selenite , and 340 ng/ml triiodothyronine , 35% Albumax II , 10 μg/ml penicillin , 10 μg/ml streptomycin and 25 μg/ml amphotericin B . Neurons were plated at a density of 2000 cells/cm2 on 0 . 05% poly-DL-ornitine ( overnight at room temperature; Sigma-Aldrich ) and 20 μg/ml laminin ( 4 h at 37°C; Sigma-Aldrich ) precoated 13 mm glass coverslips and maintained in culture at 37°C in a humidified incubator under 5% CO2 . For immunolabelings , DRG cultures on coverslips were fixed in 4% paraformaldehyde ( PFA ) in PBS for 20 min , followed by 3 rinses in PBS , and an incubation with 3% horse serum and 0 . 2% Triton X-100 in PBS ( blocking buffer ) for 30 min at room temperature . Coverslips were then incubated for overnight at 4°C with monoclonal rabbit anti-β-III tubulin ( 1:1000; Sigma-Aldrich ) . After several rinses in PBS , coverslips were incubated for 1 h at room temperature , with fluorochrome-labeled secondary antibody alexa-488 ( 1:400; Sigma-Aldrich ) . Finally , coverslips were rinsed with PBS again and mounted in DAPI-fluoromount G . For quantification of various parameters of neurite outgrowth , fluorescent images from a Leica microscope were taken , and morphological measurements were performed using the plugin Neuron J from the ImageJ software . A neurite was defined as a process that measures at least the length of one cell body and stains positively for neurofilament protein . The length of the longest neurite from the cell body to its distal end were traced and measured for each neuron in order to calculate the average length ( 300 cells were measured at each of 3 independent experiments ) . MN cultures were prepared from 13 . 5 embryonic day ( E13 . 5 ) mouse spinal cord as described previously [49–51] , but with minor modifications . Briefly , mouse embryo spinal cords were dissected and the dorsal half removed . Ventral spinal cords were dissociated mechanically after trypsin treatment ( 0 . 025% trypsin in HBSS ) , and collected afterwards under a 4% bovine serum albumin cushion . The largest cells were isolated by centrifugation ( 10 min at 520 g ) using iodixanol density gradient purification . The collected cells were finally suspended in a tube containing MN complete medium: Neurobasal ( Life technologies ) supplemented with B27 ( Life technologies ) , 2% horse serum , 1x glutamax ( Life technologies ) , and a cocktail of recombinant neurotrofins: 1 ng/mL BDNF; 10 ng/mL GDNF , 10 ng/mL CNTF , and 10 ng/mL HGF ( PreProtech ) . 10 μM AraC ( Sigma-Aldrich ) was added to the culture medium to limit the growth of non-neuronal cells . Isolated MNs were plated on poly-D-lysine/laminin-coated surfaces as described previously [52] , and grown in a 5% CO2 incubator at 37°C . For long-term experiments , media was changed every 2–3 days . Cultured MNs were clearly identified by immunofluorescence using SMI-32 ( Covance ) and HB9 ( Millipore ) antibodies or by morphological criteria . For quantitative analyses , isolated neurons plated on glass coverslips were fixed 24 and 48 h after plating with freshly prepared PHEM ( 60 mM PIPES , 25 mM HEPES , 5 mM EGTA , 1 mM MgCl2 , pH 7 . 4 ) , 4% PFA for 20 min at 37°C , washed with PBS and then activated with NH4Cl for 5 min . Cells were then blocked for 45 min at RT prior to the double-staining for mouse anti-acetylated α-tubulin ( Ac-tub ) and rabbit anti-β-III tubulin ( β-III tub ) ( Sigma-Aldrich ) . The blocking solution consisted of 10% fetal bovine serum ( FBS ) and 0 . 1% Triton X-100 in PBS . As secondary antibodies we used an anti-mouse-Alexa488 and an anti-rabbit-Alexa633 . Cells were finally counterstained with dapi for nuclei visualization and kept at 4°C for 48h prior to image capture . For confocal z-axis stacks , stacks of 10 images separated by 0 . 2 μm along the z-axis were acquired using the appropriate color channel on the detector assembly of a DMI 6000 microscope equipped with a Leica-TSC-SP8 laser scanning confocal imaging system and with an 63x 1 . 4 oil immersion objective . Z-stacks at each point were collapsed to maximum projections and 3D reconstruction and volume rendering of the stacks were performed with the appropriate plugins of ImageJ ( N . I . H . , USA ) . The acetylated α-tubulin was measured with ImageJ software using stacks ( 6 slides , 0 . 21 μm each one ) with the maximum projection . Results were expressed as the acetylated α-tubulin fluorescence intensity through the neurite or axon . Cytosolic Ca2+ imaging with Fura-2 was performed as described by [56] . MNs were isolated from 13 . 5 days embryos from each genotype as described previously . Cells were plated onto 24 mm round coverslips and , after 24 h , loaded with Fura-2AM by incubation in 30 mM D-glucose Ca2+-free HCSS with 5 μM Fura-2AM and 50 μM pluronic F . 127 acid ( Sigma-Aldrich ) , for 30 min at 37°C , and rinsed with HCSS , 2 mM CaCl2 , for 30 min . Regions of interest ( ROIs ) were selected by morphology taking into account only motorneurons . Fluorescence ( emission 510 nm ) ratio of Ca2+-free ( F380 ) to Ca2+-bound probe ( F340 ) was analyzed using Metafluor for Leica developed by Metamorph ( Universal Imaging ) . Analysis of SOCE was done using a standard protocol , depleting ER-Ca2+ using 5 μM of TG and inducing SOCE with 2 mM of CaCl2 . The analysis of total Ca2+ content was carried out using 5 μM of Ionomycin ( Sigma-Aldrich ) in Ca2+-free HCSS plus 1 mM EGTA . Each experiment was plotted and calibrated in terms of [Ca2+] . Calculations and statistical analysis were performed using Excel ( Microsoft Corporation , Redmond , WA ) , Statgraphic statistical software and Prism software ( GraphPad Software , Inc . , San Diego , CA ) . The specific test applied in each case is indicated in the figure legend or the text . Differences were considered statistically significant when p<0 . 05 , with a confidence limit of 95% . The experimental research performed in animals has been approved by the Bioethics Committee of the Consejo Superior de Investigaciones Científicas ( CSIC ) and the Animal Experimentation Ethics Committee of the Centro de Investigación Príncipe Felipe ( CIPF ) . | Charcot-Marie-Tooth ( CMT ) disease is an inherited motor and sensory peripheral neuropathy . Mutations in the GDAP1 gene cause either an axonapathy or an myelinopathy that can be transmitted recessively or dominantly to offspring . GDAP1 is located in the mitochondrial outer membrane and seems to participate in the mitochondrial network dynamics . To investigate the biological and functional consequences of lack of GDAP1 and to gain insight into the pathophysiology of the GDAP1-related neuropathies we have generated a Gdap1 knockout mouse . Characterization of this model revealed that the absence of GDAP1 induces a peripheral neuropathy with loss of motor neurons and abnormal neuromuscular junctions . We also observed defects in embryonic motor neurons and adult dorsal root ganglia sensory neurons derived from affected animals . Specifically , cultured motor neurons showed large and abnormal mitochondria , dilated perinuclear space and endoplasmic reticulum , changes in acetylation of cytoskeletal α-tubulin and calcium depletion . We propose that pathophysiology of GDAP1-associated recessive CMT neuropathy may be the consequence of abnormal calcium homeostasis and changes in the mitochondrial network biology and mitochondria–endoplasmic reticulum interactions . Our findings may be also relevant to understand the role of GDAP1 in relation to other neuropathy-related mitochondrial proteins such as mitofusin 2 . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Lack of GDAP1 Induces Neuronal Calcium and Mitochondrial Defects in a Knockout Mouse Model of Charcot-Marie-Tooth Neuropathy |
Neurotransmitter release depends on the fusion of secretory vesicles with the plasma membrane and the release of their contents . The final fusion step displays higher-order Ca2+ dependence , but also upstream steps depend on Ca2+ . After deletion of the Ca2+ sensor for fast release – synaptotagmin-1 – slower Ca2+-dependent release components persist . These findings have provoked working models involving parallel releasable vesicle pools ( Parallel Pool Models , PPM ) driven by alternative Ca2+ sensors for release , but no slow release sensor acting on a parallel vesicle pool has been identified . We here propose a Sequential Pool Model ( SPM ) , assuming a novel Ca2+-dependent action: a Ca2+-dependent catalyst that accelerates both forward and reverse priming reactions . While both models account for fast fusion from the Readily-Releasable Pool ( RRP ) under control of synaptotagmin-1 , the origins of slow release differ . In the SPM the slow release component is attributed to the Ca2+-dependent refilling of the RRP from a Non-Releasable upstream Pool ( NRP ) , whereas the PPM attributes slow release to a separate slowly-releasable vesicle pool . Using numerical integration we compared model predictions to data from mouse chromaffin cells . Like the PPM , the SPM explains biphasic release , Ca2+-dependence and pool sizes in mouse chromaffin cells . In addition , the SPM accounts for the rapid recovery of the fast component after strong stimulation , where the PPM fails . The SPM also predicts the simultaneous changes in release rate and amplitude seen when mutating the SNARE-complex . Finally , it can account for the loss of fast- and the persistence of slow release in the synaptotagmin-1 knockout by assuming that the RRP is depleted , leading to slow and Ca2+-dependent fusion from the NRP . We conclude that the elusive ‘alternative Ca2+ sensor’ for slow release might be the upstream priming catalyst , and that a sequential model effectively explains Ca2+-dependent properties of secretion without assuming parallel pools or sensors .
Neurotransmitter release and synaptic transmission depend on the fusion of secretory vesicles with the plasma membrane by exocytosis , and the ensuing release of the contained neurotransmitter molecules . Exocytosis itself is the conclusion of a number of steps , which starts by the generation of the vesicle and its filling with neurotransmitter , and continues with the transport of the vesicle to the plasma membrane , its physical attachment to the membrane ( docking ) , the attainment of fusion competence ( priming ) , and ends with its fusion as the result of the arrival of a Ca2+ signal . The essential nature of Ca2+ for the final step of neurotransmitter release has been known since the pioneering work of Bernard Katz [1] , whereas the high Ca2+-cooperativity of this step was demonstrated by Dodge and Rahamimoff [2] . Ca2+ uncaging made it possible to describe this cooperativity quantitatively and to derive mathematical models for the Ca2+ triggering step [3]–[6] . Later studies showed that at least one upstream replenishment step , probably vesicle priming , is also Ca2+-dependent [7]–[9] ( for a review , see [10] ) . However , the arrangement of the different Ca2+-dependent steps with respect to each other is currently less than clear . Initially , most working models assumed Ca2+-dependent vesicle priming and neurotransmitter release through a sequential pathway with one release sensor [11] , [12] . Later on , to account for the observation of kinetically distinct ( i . e . fast and slow ) release phases , models incorporated different releasable vesicle populations , or pools [13]–[16] . These pools deviated from each other either in terms of molecular composition or localization with respect to Ca2+ channels . The notion of parallel releasable vesicle pools was reinforced when deletion of the Ca2+ sensor synaptotagmin ( syt ) ( -1 or -2 ) was found to eliminate fast release , while Ca2+-dependent slow release components remained [17]–[19] . This led to the suggestion of multiple sensors in parallel either controlling a single [19] , [20] or different vesicle pools [17]: in the absence of syt , a second release sensor would drive fusion . Despite the fact that parallel pathways have been the working model for more than a decade , molecular correlates of the slow release pathway are still missing: syt-1 , syt-2 and syt-9 are now widely accepted to be Ca2+ sensors for fast release [21] , and detailed roles for many other proteins in fast release have been identified in vivo and reconstituted in vitro , including SNAREs , Munc13/CAPS and Munc18 [22] , but no similar set of proteins dedicated to slow release is known ( but see Discussion ) . Therefore , it remains important to consider alternatives and even return to sequential models to investigate if , with novel assumptions , such models suffice to explain the Ca2+-dependent properties of regulated secretion . Based on properties derived from molecular perturbation studies , we propose a sequential model , where we make novel assumptions regarding the nature of the Ca2+-dependent priming steps . Recent studies concluded that initial assembly of SNARE-complexes is catalyzed by Ca2+-sensitive molecules such as Munc13 [23] , [24] and coincides with priming [25] . Even though the word ‘catalyzed’ is often used loosely , we here decided to explore a model with a Ca2+-dependent catalytic mechanism in the strict sense , i . e . a mechanism , which increases both forward and backward rates by reducing the energy level of the transition state . In addition , to account for data obtained in syt-1 knockout cells ( see below ) , we suggest that the Ca2+ sensor for fast release ensures the steady-state population of the Readily Releasable Pool ( the RRP ) at rest , either by preventing SNARE-dependent fusion in the absence and synchronizing fusion in the presence of Ca2+ [26]–[30] , or by lowering the free energy of the RRP in the Ca2+-unbound state [31] . By mathematical modeling and comparison to older and new data we here show that these assumptions suffice to explain the Ca2+-dependence of secretion from chromaffin cells . Moreover , our model can account for recovery of the RRP after strong stimulation and deliver parsimonious explanations for the observed effects of SNARE or syt mutations without the need to evoke additional parallel pathways or release sensors . We suggest that a fresh look at sequential models is of value to understand kinetic diversity in secretory systems .
In order to explain the complex Ca2+-dependent properties of regulated exocytosis , we investigated a mathematical model with sequentially arranged pools ( Sequential Pool Model , SPM ) , separated by Ca2+-dependent steps [11] ( Fig . 1A ) . Our model assumes that some vesicles reside in a Non-Releasable Pool ( NRP ) from which they can undergo reversible priming to a Readily-Releasable Pool ( RRP ) . From there , vesicles can irreversibly converge to the fused state ( F ) . Vesicles enter the NRP by recruitment from a much larger Depot Pool ( Fig . 1A ) . This recruitment is Ca2+-dependent , driven by a Michaelis-Menten Ca2+-dependent forward process ( k1 ) with a Kd of 2 . 3 µM , as in previous models of the adrenal chromaffin cell [16] , [32] . At rest , the supply of vesicles to the RRP from the NRP can be slow , because the demand is low . However , during repetitive stimulation , RRP vesicles may soon be depleted unless Ca2+ speeds up their re-supply , as observed experimentally [9] , [32] . Here , we investigated the possibility that the process that enhances priming ( i . e . transition from NRP to RRP ) is Ca2+-dependent catalysis . A catalyst lowers the activation energy barrier of a reaction ( Fig . 1A ) , thereby simultaneously increasing both forward ( k2 ) and backward ( k−2 ) rates ( see Materials and Methods ) . Thus , unlike the selective increase of the forward rate , a catalyst speeds the transition without affecting the proportion of NRP- to RRP-vesicles in equilibrium . This complies with data in the adrenal chromaffin cells , as shown below . The Ca2+ trigger for fusion is modeled as a sequential sensor binding 3 Ca2+ ions , followed by a Ca2+-independent , but very fast , conversion to the Fused state [16] ( rate constant k4 , not shown in Fig . 1A; see Fig . 2A ) . In thermodynamic terms , the Ca2+ trigger affects the energy barrier for fusion . This barrier is high in the absence and low in the presence of Ca2+ , resulting in strongly Ca2+-dependent fusion rates ( Fig . 1A ) . We modeled release from chromaffin cells using the SPM to investigate whether previously obtained experimental data could be explained by the sequential Ca2+-dependent actions we suggest . For comparison , we simulated the Parallel Pool Model ( PPM ) , where this had not already been done in the literature [16] , [32] ( Fig . 1B ) . In the PPM , parallel fusion of two releasable pools , the Slowly-Releasable Pool ( SRP ) and the ( fast ) Readily-Releasable Pool ( RRP ) , is assumed . The Ca2+ sensors for RRP and SRP fusion are separate sequential and cooperative sensors each binding three Ca2+ ions ( see Materials and Methods ) . Like in the SPM , supply of vesicles from the larger Depot Pool is a Ca2+-dependent Michaelis-Menten process [16] . The parameters of this step as well as the ones for the Ca2+-binding and fusion from the RRP were identical in the two models . Each model was solved in the steady-state at the beginning of each simulation , and then allowed to evolve by numerical integration , driven by the Ca2+ signal characteristic for each stimulation protocol . The cumulative release was calculated , from which release rates and pool sizes were determined by fitting a sum of exponential functions . Parameters were taken or estimated from published data , as described in Materials and Methods , and are listed in Table 1 ( SPM ) and Table 2 ( PPM ) , respectively . Pioneering work [6] , [16] , [32] using Ca2+ uncaging and cellular capacitance recordings resulted in the identification of fast and slow release components in adrenal chromaffin cells . After parameter estimation , we simulated both models ( SPM , PPM ) using an abrupt Ca2+-step from 0 . 5 µM to 25 µM ( Fig . 2 ) . Both models displayed biphasic capacitance responses , in agreement with published data . The fastest phase is referred to as the ‘fast burst’ of release and has a rate of ∼50 s−1 under these conditions . The subsequent slow burst of release is approximately 10-fold slower . Following these two components , release persists in the so-called sustained phase , which is nearly linear . In the ‘classical’ PPM , the fast burst component is caused by the fusion of vesicles from the RRP , whereas the slow burst component owes itself to the fusion of SRP-vesicles ( Fig . 2B ) . Finally , when both RRP and SRP have been depleted , newly recruited vesicles ( from the Depot Pool ) fuse continuously as long as the intracellular Ca2+ concentration ( [Ca2+]i ) remains high , giving rise to the sustained component of release . A particularity of the PPM is that fusion during the sustained phase is almost entirely through the SRP , since the SRP-to-RRP conversion is relatively slow and Ca2+-independent ( Fig . 2B , middle panel: Fs gives the fusion through the slow pathway , Fr gives the fusion through the fast ( rapid ) pathway ) [33] . In the sequential model ( SPM ) , the fast burst component corresponds mainly to the rapid depletion of RRP-vesicles , whereas the slow burst component corresponds mainly to NRP-vesicles that mature to the RRP-state before fusing ( Fig . 2A ) . This maturation step is slower than fusion from the RRP and it is Ca2+-dependent , driven by the Ca2+-dependent catalyst , giving rise to the slow burst phase . It should be noted that the identification of the fast component with the RRP and the slow component with the NRP is only approximate , because the size and kinetics of the two phases depend on the entire system ( Materials and Methods ) . Therefore , to identify fast and slow burst components we fitted simulated traces with a sum of exponentials in order to compare these values to the ones obtained from experiments in the same way . Finally , when both the NRP and RRP are empty , sustained release is driven by the upstream Ca2+-dependent reaction from the Depot Pool ( k1 , Fig . 1 ) . Other than in the PPM , these vesicles also fuse through the RRP , which is the only releasable state in the SPM . Since both models could reproduce biphasic burst release and the sustained component , we next compared our new model with data that had previously been fitted with the PPM [16] . Systematic variation of the Ca2+ signal driving the release model allowed the determination of amplitudes and rate constants of fast and slow burst components as a function of pre- and post-flash Ca2+ concentrations ( data points in Fig . 3A , 3B , and 3C , from [16] ) . Experimentally it was found that the time constants depended on the Ca2+ levels reached after Ca2+ uncaging ( Fig . 3A ) , whereas the number of vesicles released in each phase was relatively independent of post-flash [Ca2+]i ( data points in Fig . 3B , from [32] ) . Our model accounts well for the kinetic data ( model simulations are shown as lines in Fig . 3; for examples of simulated capacitance traces see insert in Fig . 3B ) . In further agreement with the SPM , both slow and fast burst components were augmented with moderate increases of resting ( pre-flash ) [Ca2+]i below 0 . 7 µM , owing to the Ca2+-sensitive supply rate ( k1 in Fig . 1 ) from the Depot ( Fig . 3C , model simulation are lines , data are points ) , while at higher pre-flash [Ca2+]i the size of both components decreased , due to the partial depletion of the underlying pools ( NRP , RRP ) through release before the uncaging event . However , the relative proportion of fast and slow components remained fairly constant ( Fig . 3C ) . The invariance in fast-to-slow amplitude when changing the Ca2+ concentration originally served as an argument for a direct slow fusion pathway in the PPM , since it seemed inconsistent with a Ca2+-dependent interconversion between fast and slow vesicles . However , we show here that when modeling Ca2+-dependent priming as a catalytic process , the invariance is maintained in a sequential model ( lines in Figs . 3C and 3B ) , except for very low post-flash Ca2+ concentrations ( Fig . 3B ) , where the distinction between fast and slow burst components becomes experimentally challenging . Another stimulation paradigm probed the recovery of the fast burst component in chromaffin cells [32] . The stimulation paradigm combined depolarizations and Ca2+ uncaging to allow a selective depletion of the fast component without affecting the slow component . It was found that the recovery of the fast component occurred at the loss of the slow component ( data points in Fig . 3D , from [32] ) . This behavior is also an inherent feature of our sequential model: the refilling of fast ( RRP ) vesicles after selective depletion occurred at the cost of NRP vesicles , which are the major source of the slow burst ( lines in Fig . 3D ) . In the above , we simulated a model where only a single Ca2+ bound to the priming catalyst ( cooperativity one ) . We also constructed a version of the model with a cooperativity of two ( Materials and Methods ) . Fitting this model to the data also resulted in a satisfactory fit ( Fig . S1 , Table S1 ) . The fit to the fast and slow burst components as a function of preflash [Ca2+]i was a little better ( Fig . S1C ) , but the fit to the fast burst fraction as a function of postflash [Ca2+]i was a little worse ( Fig . S1B ) . Thus , the data do not allow a clear conclusion as to whether the catalyst has a cooperativity of one or two . Therefore , we continued exploring the simplest model with cooperativity one . We conclude that the now classical data of Thomas Voets on fast and slow release phases and fusion kinetics in mouse chromaffin cells can all be satisfactorily fit by our sequential model ( SPM ) , as well – as previously shown [16] – by the PPM . One fundamental difference between the PPM and the SPM is the recovery behavior of the RRP after its depletion . In the PPM , the refilling of the RRP from the SRP is Ca2+-independent . This Ca2+-independence was concluded from the observation of parallel enhancement of the fast and slow burst by increasing steady state [Ca2+]i in the sub-micromolar range before stimulation [16] ( Fig . 3C ) . However , as shown above , such a behavior is also inherent to the SPM , as long as the Ca2+-dependent acceleration of priming is catalytic . Therefore , the two models can be distinguished by probing the recovery of the RRP under conditions of elevated Ca2+: according to the SPM , its recovery should be sped up , whereas the PPM predicts Ca2+-independent - and thus slower - recovery . Recently , we used a dual-uncaging protocol to investigate the recovery of the RRP after Ca2+-uncaging had emptied both the fast and the slow burst components [34] . Applying a second Ca2+ uncaging flash at variable inter-stimulus intervals ( ISI ) while measuring Ca2+-relaxation ( Fig . 4A , top panel ) allowed estimating the RRP recovery ( points in Fig . 4D ) . Note the difference in approach to the previous selective depletion of the RRP under conditions where Ca2+ presumably relaxed to baseline much faster ( Fig . 3D ) . We found that already at 8 s the fast component had recovered to 18 . 1±4 . 2% of the initial value , and at 22 s recovery was at 35 . 3±4 . 9% ( symbols in Fig . 4D , note that all lines in Fig . 4 represent simulations , not data ) . This behavior is inconsistent with the PPM ( broken line in Fig . 4D ) , which featured a full ∼10 s delay , before the first recovery of the fast component was visible , owing to the lack of Ca2+-accelerated refilling of the RRP . This is also clearly appreciated from the simulations of the capacitance increases ( Fig . 4A , B ) , which lack a fast component in the PPM at these time points . Looking at the evolution of the pools in the PPM ( Fig . 4C ) it is clear that the RRP does not start to recover appreciably within the first 10 s . In contrast , the SPM in fact predicted the faster recovery of the fast component without the need to adjust parameters ( i . e . the same parameters were used here as in Fig . 2 and 3 ) ( Fig . 4D ) . On the simulation of single traces , the fast burst is clearly seen at short interstimulus intervals ( Fig . 4A , B ) . In spite of the quicker recovery of the fast component in the SMP , its recovery still lagged behind the recovery of the slow component ( Fig . 4B and Fig . S2 ) . The faster recovery of the slow pool is a fundamental feature of neurotransmission also found in the Calyx of Held synapse [35] . The faster recovery in the SPM is a potentially physiologically important property , which allows secretory cells to regain potency for fast release quickly after stimulation . Clearly , the SPM is more consistent with this type of experiment . The recovery ‘pause’ in the PPM due to the Ca2+-independent SRP-to-RRP conversion is a fundamental feature of that model and it would therefore not be possible to account for this behavior while still accounting for the data in Fig . 3 by mere adjustment of the model's parameters . We therefore sought to modify the PPM in such a way that this important property could be reproduced by a model with parallel fusion pathways by speeding the recovery of the RRP in a Ca2+-dependent manner . As described above , this cannot be achieved by an increase in the SRP-to-RRP forward conversion rate alone , as this would change steady-state pool sizes ( inconsistent with Fig . 3 ) . Therefore we made use of our idea of a Ca2+-dependent catalyst , which is used in the SPM , and incorporated catalysis in the SRP-to-RRP inter-conversion , so that once again forward and reverse rates were Ca2+-dependent ( Fig . S3A ) . Indeed , such a model ( PPM+cat ) was also able to account for the general Ca2+-dependence of release ( Fig . S3A–C , Table S2 ) while the catalysis of the SRP-RRP inter-conversion sped the recovery of the fast component at elevated Ca2+ to a closer agreement with the experimental data ( Fig . S3D ) . In conclusion , the classical PPM failed in reproducing the faster recovery of the fast component after strong stimulation , while the SPM or the modified PPM+cat , which both feature Ca2+-dependent catalysis in RRP replenishment , matched the data more closely . In our opinion this still argues for the SPM rather than the PPM+cat , because the SPM is simpler , has fewer parameters and does not require a second fusion pathway , while it can account for all experiments investigated . Thus far , the SPM is in line with experimental data describing vesicle pools , fusion kinetics and RRP recovery . The recovery after strong stimulation particularly favors the SPM over the PPM . Another difference between the models is in the predicted effect of a change in the fast release rate . In the PPM , release rates and amplitudes are independent . Therefore , a change in any one of these parameters is expected to leave all others ( almost ) unaffected . In contrast , in the SPM model , the amplitudes and release rates are all functions of the elementary rate constants k2 , k−2 and k3 ( see Materials and Methods ) . Therefore , the time constants and amplitudes of fast and slow burst release are interdependent ( due to their individual dependence on k2 , k−2 and k3 ) and a change in k3 is expected to affect both fast and slow release . This is demonstrated in Fig . 5A where we have calculated the time constants of fast and slow burst release while varying the final fusion rate k3 ( lines are model simulations ) . Moreover , the relative contribution of fast and slow release depends on the time constant of fast release ( Fig . 5B ) , reducing k3 will therefore also change the ratio of fast to slow release . It is commonly assumed that one of the last events before vesicle fusion is the assembly of the final interaction ‘layer’ ( layer +8 ) in the SNARE complex . In the SPM this molecular event is primarily reflected by the rate constant k3 ( although it might affect k2 as well ) . Mutagenesis studies of layer +8 in synaptobrevin-2 previously performed by us [25] indeed resulted in the identification of an increase in the fast time constant ( data points in Fig . 5A ) . In the PPM one might expect that the release rates of both fast and slow vesicles are decreased if both fusion reactions required the C-terminal interaction of synaptobrevin-2 . However , changes in the proportion of fast to slow burst release are not expected . In contrast to this we found that mutation of layer +8 resulted in a shift of the relative contribution of fast and slow release; as fast release became slower , the amplitude of the fast component decreased while the amplitude of the slow component increased ( [25]; data points in Fig . 5B ) . With the SPM we can explain the connection between these effects in the framework of a simple model ( solid lines in Fig . 5 ) : C-terminal SNARE destabilization reduces k3 ( illustrated by the purple arrow in Fig . 5Ci ) , resulting in an increased time constant for fast release ( τfast , Fig . 5A ) and a simultaneous change in fast and slow contribution to release ( Fig . 5B ) . The PPM cannot account for the simultaneous change of release rates and amplitudes by a single parameter change , because pool sizes and release rates are independent . Therefore , changes in the fast release rate are predicted to leave the relative contribution of fast and slow release unchanged ( dashed lines in Fig . 5B are simulations of the PPM ) , which is not easily reconciled with the experiment . The PPM can still mathematically describe the data , but for this it is necessary to assume that several rates are changed by the single point mutation , including one that changes the relative sizes of the RRP and SRP ( purple arrows in Fig . 5Cii ) . Hence , while both models fit the data , the SPM delivers a straightforward explanation for the observed relationship of fast release rates and amplitudes in general and relates the main effect of C-terminal SNARE mutation to a single rate constant ( k3 ) in particular . Thus , a sequential model featuring an upstream Ca2+-dependent catalyst with a downstream Ca2+ sensor predicts the effects of SNARE mutations . One argument for parallel pool or parallel sensor models has been the observation that after deletion of the fast Ca2+ sensor in various systems slow release components persist . This is also the case in mouse adrenal chromaffin cells , where deletion of syt-1 leads to ablation of the fast burst component , whereas the slow burst component and the sustained component both remain [17] , [31] , [36] . Clearly , a Parallel Pool Model , where the RRP would utilize syt-1 and the SRP a separate Ca2+ sensor ( even though no such molecule has been found yet ) is an easy and reasonable way to account for this finding . How can one account for it in a sequential model ? Recent in vitro and in vivo experiments have emphasized the function of syts and/or complexins as clamps on release [26]–[30] . Therefore , one idea is that these proteins – probably working in concert – are able to arrest SNARE complex assembly . Ca2+-binding to syt would release the clamp and allow the SNARE complex to complete its assembly , leading to rapid membrane fusion . In our model , the clamp would generate the final energy barrier for fusion ( Fig . 1A ) , which could be removed by the binding of three Ca2+ ions to syt-1 ( Fig . 2A ) . In this scenario one can model the syt-1 null data by removing the final fusion barrier , making RRP vesicles fuse with the maximal rate ( k4 ) ( Fig . 6A , top panel ) . This fast – and now Ca2+-independent – fusion from the RRP leads to chronic pool depletion . As a result of the missing clamp , the NRP vesicles in effect become releasable , with the transition from NRP to RRP being rate limiting for fusion . This transition is sped up by the Ca2+-dependent catalyst ( Fig . 1 ) , resulting in Ca2+-dependent , but much slower , release rates . We simulated our model under these conditions driven by a Ca2+ uncaging event ( Fig . 6A , bottom panels ) . Indeed , our model predicted a missing fast component of release , and the persistence of both slow and sustained release components [17] , [31] , [36] . The slow component was driven by the sequential transition of vesicles from the NRP to the RRP to the F-state , following an increase in [Ca2+]i . The slow component of release was somewhat smaller in the syt-1 null than in the WT case , because of partial depletion of the NRP at rest ( comp . Fig . 6A and Fig . 2A ) . Because of the lack of the last energy barrier , ongoing release at rest depleted the RRP ( and the NRP partially ) and led to an increase in the spontaneous release rate from 1 . 7 fF/s to 6 . 9 fF/s . Another way of accounting for the selective loss of the fast burst component in the SPM model is to assume that syt-1 stabilizes the RRP state , by lowering its free energy . This is in line with previous data showing that syt-1 overexpression increases the RRP/SRP ratio [31] . We simulated this by increasing k−2 by a factor of 10 ( Fig . 6B ) . Thermodynamically , this corresponds to increasing the energy state of the RRP , plus the subsequent energy barrier . Simulating this situation resulted in only a very small fast burst , followed by a slow burst and a sustained component ( Fig . 6B ) . Under these circumstances , the NRP was a bit larger than in the WT case ( Fig . 2A ) , leading to a slightly larger slow burst . In this version of the model , the properties of the Ca2+ triggering step remained unchanged . This might appear unrealistic for a simulation of the syt-1 null , because syt-1 is usually supposed to provide the Ca2+ binding sites . However , since most vesicles are present in the NRP , and the NRP-to-RRP transition is rate limiting for the majority of release , the properties of the Ca2+ sensor can in fact be changed simultaneously without noticeably changing the kinetics of overall release . In other words , when the majority of vesicles are present in the NRP , the model is pretty insensitive to the details of the RRP-to-F conversion . We conclude that the SPM offers at least two possibilities for qualitatively accounting for the selective loss of the fast burst component , as observed in the syt-1 null , without assuming parallel release sensors or vesicle pools: either by assuming that the RRP is emptied because of spontaneous fusion , or by assuming that the RRP is destabilized , so that most vesicles reside in the NRP instead .
The Parallel Pool Model ( PPM ) assumed that Ca2+-dependent priming ( refilling of the SRP ) indirectly led to refilling of the RRP through a Ca2+-independent process ( Fig . 1B ) . A Ca2+-independent SRP-to-RRP inter-conversion was assumed , because this could account for the parallel increase in SRP and RRP size with increasing Ca2+ concentrations at rest ( Fig . 3C ) . If merely the forward rate were Ca2+-dependent , then the RRP-to-SRP ratio would change with [Ca2+]i , which was inconsistent with the experiment . However , the slow , Ca2+-independent RRP refilling reaction made it impossible to account for the slow burst of release by fusion through the RRP state alone . Instead , it was assumed that the SRP-vesicles could fuse directly through a separate pathway , which in turn necessitated a separate Ca2+ sensor for this pool . This arrangement seemed to be confirmed when deletion of syt-1 led to a specific deletion of the fast burst of release [17] . We solved the problem of the invariant slow-to-fast release ratio in a different way: by assuming that the NRP-to-RRP conversion is driven by a Ca2+-dependent catalyst . This in effect makes both the forward and the reverse reactions Ca2+-dependent . Moreover , because the catalyst acts on the transition state ( Fig . 1A and Materials and Methods ) , it follows immediately that the Ca2+-dependence of forward and reverse rates are identical , ensuring the invariance of the NRP/RRP ratio . The catalytic process then speeds up refilling of the RRP to the degree that fusing the upstream NRP vesicles after transit through the RRP accounts for the slow burst of release in uncaging experiments , while this transition is slow at rest . This removes the need to assume a separate release pathway for those vesicles . Thus , in our model the ‘slow’ vesicles are no longer releasable , which is why we renamed this state Non-Releasable ( NRP ) . Moreover , the Ca2+-dependent NRP-to-RRP conversion speeds up RRP refilling , which makes it possible to explain the faster recovery kinetics upon strong stimulation ( Fig . 4 ) that could not be explained in the PPM . Thus , modeling the Ca2+-dependence of priming as a catalytic process explains a range of phenomena . Ca2+-dependent refilling of the RRP is a physiologically important mechanism , which ensures that the same signal that causes release from the RRP also speeds up refilling . We also managed to modify the PPM by the incorporation of a Ca2+-dependent , catalytic acceleration of RRP refilling ( PPM+cat , Fig . S3 ) , which could account for the data in a parallel model with additional parameters . This emphasizes the advantage of modeling RRP replenishment as a catalytic process independent of whether release commences in parallel or not . Moreover , catalysis accelerates priming under high-use conditions while strictly preventing ‘overfilling’ of the RRP , which is characteristic of models with non-catalytic Ca2+-dependent priming [12] . Since vesicle priming coincides with SNARE-complex formation [25] , [37] , the molecular counterpart of the catalyst should be sought amongst Ca2+-dependent proteins which have been shown to facilitate SNARE-complex assembly; likely candidates are Munc18 , Munc13 and CAPS proteins [23] , [38] . Indeed , it was recently suggested that Munc13-1 can reduce the energy barrier for SNARE-complex formation [24] , and stimulates both opening and closing of syntaxin [22] implying catalytic action . Furthermore , Munc13/CAPS are Ca2+-dependent proteins known to affect vesicle priming [38]–[42] . Reinterpreting these proteins as catalysts for vesicle priming means that they have to be reclassified as enzymes , which might appear provocative or unusual . But it is not hard to see how proteins interacting with the SNAREs might provide an alternative pathway for assembly ( and disassembly ) , by providing a surface stabilizing an intermediate conformation , and such mechanisms are frequently discussed in the literature . Opening and closing of syntaxin might well be stimulated by the action of a catalyst , which stabilizes an intermediate configuration . Another catalyzed step could be the initial formation of the ternary SNARE-complex , which involves the interaction of a very short , transiently formed alpha-helix in synaptobrevin with the SNAP-25:syntaxin dimer [43] , [44] . Stabilization of the alpha-helix is a candidate mechanism for the catalysis of vesicle priming . It is not necessary for the validity of our model that a protein can be found that only acts catalytically without inducing pool size effects . Instead , the catalyst may be a molecule on the plasma membrane , the association to which is necessary for priming , but which is sped up by Ca2+ . But – unlike our simple assumption of sufficient abundance – its number may be limited , thus limiting the number of RRP vesicles . This might play a role in those cell types ( e . g . neurons ) where the RRP size appears to be limited by a fixed number of ‘release sites’ . We introduced a specific interpretation of the release sensor , as a clamp that can be lifted by Ca2+ ( Fig . 1 ) . This idea aligns with recent findings and ideas that the fusion trigger consists of syt-1 and complexin in combination [27] , [45] , [46] , even though those proteins might have additional upstream functions as well [36] , [47] . Complexin and syt-1 appear to arrest SNARE-complex assembly , effectively setting up an energy barrier for fusion , which is removed when Ca2+ binds to syt-1 . In order to explain the syt-1 null phenotype we made two separate assumptions . In the first , we assumed that the fusion barrier was removed . This caused depletion of the RRP and effectively converted the NRP into a releasable pool , because the downstream barrier was removed . Fusion in this ‘barrier-less’ model resulted in a ‘slow burst’ followed by a normal sustained component , which qualitatively matches data from syt-1 null chromaffin cells . This follows immediately from the model itself , since fusion from the NRP is the source of the slow burst in the wildtype case . In the other simulation of the syt-1 null , we assumed that syt-1 was responsible for stabilizing vesicles in the RRP , and in its absence the RRP would be nearly empty . As a result , the slow burst of release dominates after Ca2+ uncaging . The two models for the syt-1 null are distinguishable based on the size of the slow burst , which is larger in the second model , and the frequency of spontaneous release , which is increased in the first and decreased in the second case . The larger slow burst in the second model fits somewhat better with published data from the syt1 null [17] , [31] , [36] , however it should be noted that syt-1 most likely has several functions in secretion [47] , which could cause secondary changes in release amplitude . Therefore , we conclude that both models qualitatively account for the syt-1 null data . In both versions of our syt-1 null model , the ‘slow Ca2+ sensor’ driving release in the syt-1 null is the priming catalyst , which is arranged upstream of RRP and moonlights as a Ca2+ sensor for release when the NRP-to-RRP transition becomes rate limiting , as it has previously been suggested [48] . This introduces a specific interpretation of the syt-1 null . In most cases , the slow release component has been assumed to either originate from a separate releasable pool ( the SRP in the case of the chromaffin cell ) , or from an alternative sensor , which only gains access to the release machinery in the absence of syt-1 . Our interpretation would explain why the ‘slow Ca2+ sensor’ for release – though heavily searched for – thus far could not be identified: because it is also important for fast release and the two sensors are not additive . Investigations of syt-7 , which has slower Ca2+ ( un- ) binding kinetics than syt-1 [49] indeed showed that syt-7 also affected fast release in chromaffin cells [50] while its deletion in central neurons was apparently without effect [51] . Thus syt-7 does not seem to fit the description of a parallel sensor in the chromaffin cell , but the situation might be different in some neurons , including the zebrafish neuromuscular junction [52] . A recent study implicated Doc2α in asynchronous release in central neurons , but it cannot be the slow Ca2+ sensor itself , because asynchronous release was still Ca2+-dependent when Doc2α was mutated to be Ca2+-insensitive [53] . Thus , a testable prediction of our model is that the elusive secondary sensor should be found among Ca2+-dependent proteins driving priming . The alternative hypothesis of another syt-like molecule , taking the place of syt-1 or syt-2 in their absence , is also possible , but additional Ca2+-dependent priming reactions need to be assumed to account for the enhanced recovery of the RRP during elevated Ca2+ ( Fig . 4D , Fig . S3 ) , which in our model are both direct consequences of the priming catalyst . Our work documents that a sequential model can describe salient wildtype and mutant phenotypes in adrenal chromaffin cells . By this , we do not mean to imply that parallel models are inherently wrong , or even unlikely , and future experiments might revive the need for such models . However , at a point in time where it seems that molecular manipulations and experiments more often than not lead to the suggestion of separate vesicle pools or sensors , we think it is important to point out that small changes in the assumptions behind sequential models might lead to similar phenotypes . Specifically , we have shown that the assumptions of catalytic action of the priming machinery , and stabilization of the RRP by syt-1 , remove the necessity of assuming two fusogenic vesicle pools in the adrenal chromaffin cell . It is of note that recent experiments in cultured hippocampal neurons and in the Calyx of Held synapse led to the conclusion that the vesicles fusing during the slow phase of release are recruited to the fastest releasable pool [54] , [55] , in agreement with our model . When considering neurons , other factors will contribute to kinetic diversity of release . Chief amongst these is the variable distance between RRP vesicles and Ca2+ channels , which might lead to kinetically distinct release phases [56] . The experiments we analyzed here were mostly performed using a spatially homogeneous Ca2+ signal ( uncaging ) , which makes it easy to derive pool behavior . Future efforts will be needed to understand whether a sequential model when combined with spatial heterogeneity has explanatory power in the world of neurons .
To model the release from chromaffin cells we set up the reaction rate equations , which consist of a set of ordinary differential equations describing the temporal change in the population of each state ( see below ) . We followed the suggestion by Thomas Voets [16] to simplify the scheme by assuming that the size of the Depot is so large , that it does not effectively decrease by the release ( infinite Depot ) . We used the same parameters determined by Thomas Voets to describe the Ca2+ sensitivity of vesicle supply from the Depot . We also used the parameters of the release sensor found for the RRP vesicles ( three site binding model [16] ) to describe the release from the RRP state in our model . The full secretion model is shown in Fig . 2A and the values of all parameters can be found in Table 1 . The steady state values were solved analytically under the assumption of constant NRP , RRP , RRPCa , RRPCa2 and RRPCa3 ( and Depot ) sizes at rest , where RRPCan represents RRP vesicles with n Ca2+ bound . We suggest that the transition between the NRP and RRP states is catalyzed by a Ca2+-dependent catalyst . Therefore , the rate constants k2 and k−2 are functions of the Ca2+ concentration . A catalyst acts by increasing the rate of a reaction through the lowering of the activation energy barrier ( Fig . 1 ) . Yet a catalyst does not change the equilibrium of a reaction as it accelerates both the forward and the reverse rates simultaneously . Both reactions also occur without catalysis , but with low rates ( k20 and k−20 ) . We suggest that the overall rate constants k2 and k−2 are functions of the un-catalyzed ( Ca2+-independent ) rate constants k20 and k−20 and functions of the faster , catalyzed ( Ca2+-dependent ) reaction rates k2cat and k−2cat . The relative contribution of the catalyzed rate in turn depends on the probability ( g ( Ca2+ ) ) of the catalyst to have bound the required number of Ca2+ ions in order to become activated . For simplicity , we assumed that the supply of catalyst is not limited ( this could for instance mean that each vesicle bears a catalyst ) . The overall rate constants k2 and k−2 take the following form: ( 1 . 1 ) ( 1 . 2 ) We assume instantaneous binding of Ca2+ to the catalyst , so that the catalyst is in equilibrium with Ca2+ at all times . In general , a catalyst might bind n Ca2+ ions in order to be activated:The overall dissociation constant is given by: ( 1 . 3 ) For an individual Ca2+-binding step: ( 1 . 4 ) The fraction of activated ( bearing the correct number n of Ca2+ ions ) to total catalyst ( g ) is ( 1 . 5 ) Experimentally , the relative concentration of the active catalyst is difficult to access , whereas the Ca2+-levels are well defined . g ( Ca2+ ) can be rearranged: ( 1 . 6 ) ( 1 . 7 ) ( 1 . 8 ) For simplicity , it will be assumed that n Ca2+ can bind independently and with identical dissociation constants: ( 1 . 9 ) Then g ( Ca2+ ) can be further simplified: ( 1 . 10 ) ( 1 . 11 ) In most of our work we assumed the simplest case of a Ca2+-cooperativity of one ( n = 1 ) . Then equation ( 1 . 11 ) takes the simplified form: ( 1 . 12 ) Since a catalyst does not interfere with the free energy of RRP and NRP ( same in the presence and absence of Ca2+ ) , the following relationship holds true: ( 1 . 13 ) which implies that ( 1 . 14 ) Thus , it is sufficient to have information about the relative sizes of the NRP and RRP at equilibrium ( depend mostly on the k20 to k−20 ratio ) , the uncatalyzed re-supply rate of vesicles ( k20 ) and the asymptotic rate of NRP-to-RRP conversion at high Ca2+-levels to calculate all four rates . We estimated the KD and the k2cat values of the catalyst from the previous study by Thomas Voets ( data points in Fig . 3A [16] ) . The k2cat corresponds to the asymptotic rate of the slow component at high Ca2+ concentrations . As can be seen from the data points in Fig . 3A , this corresponds roughly to a rate constant of around 20 s−1 . According to equation ( 1 . 12 ) , the KD for a cooperativity of n = 1 corresponds to the Ca2+ concentration , where g ( Ca2+ ) = 0 . 5 and the slow component has reached 50% of its asymptotic rate ( a value of k20+0 . 5*k2cat , where we can assume that k20 is negligible ) . As can be seen from the data points in Fig . 3A , this corresponds to a value of roughly 100 µM . Now , k20 was calculated by using equation ( 1 . 1 ) and investing the observation that at resting Ca2+ concentrations ( 500 nM ) , the rate k2 is equal to 0 . 12 s−1 [32] . Finally , k−20 and k−2cat were calculated using equations ( 1 . 2 ) and ( 1 . 14 ) by investing the observations that k−2 at resting Ca2+ concentration is 0 . 1 s−1 and at steady state the k2/k−2 ratio is 1 . 2/1 [32] . All parameters can be found in Table 1 . The kinetic equations for the model presented in Fig . 2A take the following form: ( 2 . 1 ) ( 2 . 2 ) ( 2 . 3 ) ( 2 . 4 ) ( 2 . 5 ) ( 2 . 6 ) where k1Max = 55 fF/s is a constant because of the assumption of constant Depot size [16] . The system of differential equations was integrated numerically using a fifth order Runge Kutta method with Cash-Karp coefficients and adaptive step size . For this purpose , we used a custom made macro in IGOR Pro ( version 6 . 22A , WaveMetrics Inc . ) . The procedure of Runge Kutta was adapted from Numerical Recipes [57] . The sizes and time constants of the fast and slow release components were determined by fitting a sum of exponentials to 5 s of simulated cumulative release . The sustained release that is typically observed after the burst phase of secretion was approximated by a line . The fit function used had the following form: ( 3 . 1 ) A0 was fixed to the baseline value immediately before the onset of the Ca2+ step and t0 was set to the inflection point of the cumulative release after stimulus onset . All other parameters were free . A1 and τ1 correspond to the amplitude and time constant of the fast , and A2 and τ2 to the size and time constant of the slow burst component , respectively . A3 is the slope of the sustained component and t is time . The rate constants displayed in Fig . 3A are the inverse of the time constants . The temporal change of the two kinetic components during recovery of the fast component of release was characterized and compared to the experimental values presented in [32] . In the experiments , Voets and colleagues used a paradigm of voltage depolarizations that selectively depleted the fast component . Subsequently , the sizes of both the fast and the slow component were probed by a flash experiment ( which releases both components ) at varying inter-stimulus intervals [32] . Unfortunately , the precise Ca2+ signal at the site of the vesicle in response to the depolarization is not known . In order to test whether our model in principle would allow for a selective depletion of the fast- without major depletion of the slow component , we resorted to a simplified stimulus and probed the effect of a Ca2+ step from 0 . 5 µM to 25 µM on the NR and R states . In order to mimic the experimental conditions of the previous study we were looking for a point in time after onset of the stimulus , where the majority of fast vesicles were depleted , whereas the majority of the slow component's amplitude was still left intact , which was satisfied at the condition shown in the top panel of Figure 3D . Ca2+ was then stepped back to resting ( 0 . 5 µM ) levels and the system was allowed to re-equilibrate for various inter-stimulus intervals before a second step to 25 µM Ca2+ was applied . Then , the sizes and time constants of the two components were again obtained by the fitting of exponentials ( see above ) . For the comparison between the PPM and the SPM , the data presented in [25] were evaluated . We made use of the fact that that pre- and post-flash Ca2+ levels were similar in all conditions , allowing us to solve the model's differential equations under the assumption of constant k2 , k−2 and k3 ( Fig . 2A ) . Therefore these values are merely functions of the genetic alteration . Since we are only interested in the kinetics of burst release , we further disregarded refilling , arriving at the simple model:The kinetic equations can be written in matrix form: ( 4 . 1 ) Only the fused vesicles ( F ) contribute to the cell's increase in membrane capacitance ( CM ) during secretion . To arrive at the initial conditions we assumed a fixed amount of vesicles in the burst phase ( the sum of [NRP] and [RRP] ) , Vtot . We further assumed that the NRP and RRP states are in equilibrium and that the fusion rate is essentially zero ( k3 = 0 ) before stimulation , such that: ( 4 . 2 ) ( 4 . 3 ) ( 4 . 4 ) After solving the differential equations analytically one can express the cellular capacitance change ( ΔCM , which is equal to F ( t ) when calculating in capacitance units ) in terms of the kinetic rate constants k2 , k−2 and k3 ( same rate constants as depicted in Fig . 1 ) : ( 4 . 5 ) with ( 4 . 6 ) Accordingly , the size of the fast and slow release components and their time constants are: ( 4 . 7 ) ( 4 . 8 ) ( 4 . 9 ) ( 4 . 10 ) In Figure 5C , the equation ( 4 . 5 ) was used to fit the total burst ( 1st second ) of release from chromaffin cells expressing wildtype and mutant synaptobrevin-2 , using a custom made macro in Igor Pro 6 . 22A ( Wave Metrics ) . The onset of the capacitance increase was determined manually . The data of the measured time constants and release amplitudes following mutation of synaptobrevin-2 shown in Figure 5 were taken from Table S1 in [25] . The lines in Figure 5 represent model predictions with constant k2 and k−2 while varying k3 . The values of k2 and k−2 were determined by a global fit of the model to the data sets shown in Figure 5 . The behavior of the Parallel Pool Model was probed by simulating responses to Ca2+ uncaging ( from 0 . 5 µM to 25 µM ) for 5 s while varying the fusion rate ( kRRP = 1/τRRP ) in the following model: ( 4 . 11 ) The cumulative release was subsequently fit using the above routine . The starting values were chosen to be [RRP]0 = 61 . 3 fF and [SRP]0 = 52 . 3 fF . | The release of neurotransmitter involves the rapid Ca2+-dependent fusion of vesicles with the plasma membrane . Kinetic heterogeneity is ubiquitous in secretory systems , with fast phases of release on the millisecond time scale being followed by slower phases . In the absence of synaptotagmin-1 – the Ca2+sensor for fast fusion – the fast phase of release is absent , while slower phases remain . To account for this , mathematical models incorporated several releasable vesicle pools with separate Ca2+ sensors . However , there is no clear evidence for parallel release pathways . We suggest a sequential model for Ca2+-dependent neurotransmitter release in adrenal chromaffin cells . We assume only a single releasable vesicle pool , and a Ca2+-dependent catalytic refilling process from a limited upstream vesicle pool . This model can produce kinetic heterogeneity and does better than the previous Parallel Pool Model in predicting the Ca2+-dependence of releasable pool refilling and the consequences of SNARE-protein mutation . It further accounts for the release in the absence of synaptotagmin-1 by assuming that the releasable vesicle pool is depleted , leading to slow and Ca2+-dependent fusion from the upstream pool , but through the same release pathway . Thus , we suggest that the elusive ‘alternative Ca2+ sensor’ is an upstream priming protein , rather than a parallel Ca2+ sensor . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2013 | A Sequential Vesicle Pool Model with a Single Release Sensor and a Ca2+-Dependent Priming Catalyst Effectively Explains Ca2+-Dependent Properties of Neurosecretion |
Bacterial genomes typically consist of a single chromosome and , optionally , one or more plasmids . But whole-genome sequencing reveals about ten per-cent of them to be multipartite , with additional replicons which by size and indispensability are considered secondary chromosomes . This raises the questions of how their replication and partition is managed without compromising genome stability and of how such genomes arose . Vibrio cholerae , with a 1 Mb replicon in addition to its 3 Mb chromosome , is the only species for which maintenance of a multipartite genome has been investigated . In this study we have explored the more complex genome of Burkholderia cenocepacia ( strain J2315 ) . It comprises an extra replicon ( c2 ) of 3 . 21 Mb , comparable in size to the3 . 87Mb main chromosome ( c1 ) , another extra replicon ( c3 ) of 0 . 87 Mb and a plasmid of 0 . 09 Mb . The replication origin of c1 is typically chromosomal and those of c2 and c3 are plasmid-like; all are replicated bidirectionally . Fluorescence microscopy of tagged origins indicates that all initiate replication at mid-cell and segregate towards the cell quarter positions sequentially , c1-c2-p1/c3 . c2 segregation is as well-phased with the cell cycle as c1 , implying that this plasmid-like origin has become subject to regulation not typical of plasmids; in contrast , c3 segregates more randomly through the cycle . Disruption of individual Par systems by deletion of parAB or by addition of parS sites showed each Par system to govern the positioning of its own replicon only . Inactivation of c1 , c2 and c3 Par systems not only reduced growth rate , generated anucleate cells and compromised viability but influenced processes beyond replicon partition , notably regulation of replication , chromosome condensation and cell size determination . In particular , the absence of the c1 ParA protein altered replication of all three chromosomes , suggesting that the partition system of the main chromosome is a major participant in the choreography of the cell cycle .
The long-held view that bacteria carry the essential part of their genomes on a single chromosome blurred about 25 years ago , when the species Rhodobacter sphaeroides was found to carry certain essential genes on a large replicon distinct from the main chromosome [1] . The size and essentiality of this replicon qualified it as a chromosome , albeit a secondary one . Many bacterial genomes have since proven to be multipartite—about 10% of those sequenced and notably those of pathogenic and metabolically versatile species . For example , all Vibrio species carry one secondary chromosome [2 , 3] and all Burkholderia species have at least one and typically two [4] . They are thought to have arisen by transfer of essential genes to coresident low-copy number plasmids , which thereupon grew through further recombination events . Whether the split-genome arrangements resulting from such events persisted by conferring selective advantage is speculative , but it is reasonable to view expansion of secondary chromosomes as a means of incorporating large numbers of beneficial genes without unduly disturbing the regulation and organization of essential genes on the main chromosome . Our aim here is to determine how the maintenance of one principal and two secondary chromosomes is accommodated within the cell cycle of the beta-proteobacterium Burkholderia cenocepacia J2315 , an opportunistic pathogen of sufferers from cystic fibrosis . ( We use the term "secondary chromosome" for convenience , and deal with the nomenclature of such replicons in the Discussion . ) The size of secondary chromosomes , which can approach that of the main chromosome , makes them potentially problematic . First , the replication control systems of secondary chromosomes resemble those of low-copy number plasmids . Replication of such plasmids has been seen to lack the close coupling with the cell cycle shown by the main chromosome [5–8]; because it occupies only a brief period it is regulated only to ensure that it precedes cell division [9] . Enlargement of such a replicon requires its replication to occupy a large fraction of the cell cycle and so to be initiated early enough not to delay cell division or to risk DNA cleavage by the closing septum . Thus for large secondary chromosomes to be stable , either the system that regulates initiation of plasmid replication must be augmented in some way , or such chromosomes must develop only from plasmids , yet to be identified , that naturally replicate early in the cycle . Second , the large size of secondary chromosomes risks confusion or entanglement with the main chromosome during the mitotic segregation ( partition ) that follows their replication . Low-copy number plasmids and most chromosomes assure efficient partition through assembly of a partition complex based on binding of a specific ParB protein to a cluster of parS sites near the replication origin; poleward movement of the complex on each replicon copy is then mediated by the cognate ParA ATPase . ParABS systems provide the specificity needed to distinguish between coresident chromosomes [10] but whether they alone can assure orderly movement of bulky chromosome copies is unproven . In this they might be aided through regulatory linkage with replication control . Although replication and Par-mediated segregation operate independently in plasmids , recent work has shown the Bacillus subtilis and Vibrio cholerae partition proteins to regulate replication-initiator activity as well as chromosome segregation [11–13] . Of the many split-genome cell cycles that might be studied , only that of V . cholerae has been in depth . The Vc genome comprises chromosomes of 2 . 9Mb ( Chr1 ) and 1Mb ( Chr2 ) . Although the Chr2 origin is plasmid-like , it is replicated in phase with the cycle , but later than Chr1 , such that Chr2 and Chr1 replication terminate at about the same time [14] . The question of how the cell manages partition of two bulky replicons has been settled in the case of V . cholerae by adoption of distinct segregation patterns . The Chr1 origin appears to be tethered at one end of the cell so that segregation consists of moving one origin copy to the far pole where it in turn is fixed , as seen also in Caulobacter crescentus , Agrobacterium tumefaciens , and Sinorhizobium meliloti [15 , 16] , while the Chr2 origin is centrally located in new-born cells whence its copies segregate to the quarter positions prior to division [17] , as typified by low copy number plasmids and the single chromosomes of bacterial models such as B . subtilis and Pseudomonas aeruginosa [18 , 19] . Genomes of the Burkholderia group are divided still further . That of the B . cenocepacia J2315 reference strain comprises the principal chromosome ( c1 , 3 . 9 Mb ) , two secondary chromosomes ( c2 , 3 . 2 Mb , and c3 , 0 . 9 Mb ) and a plasmid ( p1 , 0 . 09 Mb ) [20] . How are these replicons replicated and segregated without confusion and without perturbing the cell cycle ? Each carries a parABS system which displays non-overlapping specificity in stabilization of plasmids in E . coli [10] and in formation of partition complexes in vitro [21] . Nevertheless , the role of the Par systems in their mother organism has not been defined , nor is it known whether they are brought into play independently or in concert via a cell cycle master regulator . We have examined these issues by characterizing the replication mode of each chromosome , by analyzing the number and positioning of each replicon's ori-par region with respect to the cell cycle , and by assessing the consequences for partition and growth of inactivating each ParABS system .
To analyze replication of the genome we first characterized the origin regions and determined base-pair frequency gradients throughout each chromosome . The replication origins had been provisionally located from the similarity of their genetic context to that of known origins and from GC-skew minima [10] , and we extended this analysis to substantiate the prior indications . Application of the programme Ori-finder 1 . 0 ( Tubic; [22] ) , designed to identify origins on the basis of DnaA-box density and sequence disparity , confirmed these general locations , as depicted in Fig 1A . The skewed distributions of KOPS sites ( Fig 1B ) , which bind FtsK to facilitate terminus segregation [23] , are largely consistent with these ori positions . Fig 1C shows for each replicon the outline arrangement of elements characteristic of origin regions . In the main chromosome , c1 , these features—DnaA boxes , short sequence repeats , IHF site , AT-rich block—are sufficiently dispersed to leave the precise position of the replication origin uncertain; the geography of this region is examined in further detail in the Supplementary Information ( S1 Fig ) . The origin regions of the secondary chromosomes , c2 and c3 , each include an ORF with strong similarity to repA genes that specify classic replication regulators of low copy-number plasmids , as well as typical clusters of 19–21 bp iterons to which these regulators bind , leaving no doubt that these replicons originated as plasmids ( for detail , see S2 Fig and S1 Table ) . Chemical assay of the amount of DNA in exponentially growing Bcen J2315showed it to be 1 . 6 genome equivalents per average cell ( S2A Table ) . To verify the replicon copy numbers and to examine the basic replication parameters of the Bcen genome , we initially used quantitative Southern blot hybridization to measure the relative abundance of replication origins and termini . The data showed the relative concentrations of the chromosomal ter sequences to be close to unity ( S4 Fig ) , which together with the DNA assay above confirmed that the basic copy number of each chromosome is one per cell; the higher value for the number of the plasmid ter sequences , 1 . 3–1 . 4 , results from most plasmids completing replication during the first half of the cycle ( see below ) . The ori/ter ratios obtained from these data should allow us to confirm experimentally the strong indication from GC-disparity analyses ( Fig 1; ref . 10 ) that replication of each chromosome is bi-directional , as well as to determine how long it takes ( C period ) . However , anomalous hybridization behaviour of certain probes led us to use an alternative method , direct determination of base-pair frequency by sequencing . DNAs purified from Nel13 cells growing exponentially in LB ( doubling time 72 mins ) , as well as from cells incubated in stationary phase for 8 hours to allow replication to terminate , were processed and subjected to deep sequencing ( see Materials & Methods ) . Read-number data were binned and plotted as a function of chromosome position ( Fig 2A ) . Scatter arising from variation in amplification and sequencing efficiency was minimized by normalizing the raw data ( first row ) with respect to the corresponding reads from stationary-phase cells ( second row ) , yielding the relatively tidy curves shown in the third row . The base-pair frequency gradients show replication of the main chromosome , c1 , to proceed from the predicted origin region in both directions and to terminate diametrically opposite . Replication of c2 is essentially the same , although the fully symmetric pattern observed with c1 is not seen; origin-proximal sequences in the left arm are slightly less abundant than those on the right . Possible explanations for this asymmetry are occasional unidirectional , clockwise replication from the c2 origin or delayed initiation of the left arm . The replication mode of c3 is more complex . It appears to be predominantly bidirectional with a difference in frequency of the origin-flanking sequences implying sporadic unidirectional or delayed replication as suggested for c2 . But in addition , the steepness and extent of the two gradients differ from each other . The base pair frequency minimum is displaced anticlockwise , such that the left and right replichores occupy , respectively , about 60% and 40% of the c3 replicon . This further asymmetry can be provisionally interpreted as resulting from relatively slow movement of the clockwise fork and from absence of a strong replication terminator opposite the origin , such that the anticlockwise fork terminates within the right chromosome half . The bp frequency gradients were used to calculate the replication time , C , of each chromosome ( strictly , each chromosome arm ) , from ori / ter = 2C/τ , ( C = τ [log ( ori/ter ) / log2]; [27] ) where τ = culture doubling time . Although the normalized frequency curves had enabled us to discern the overall replication pattern of c1 and c2 , they proved unsuitable for quantitative purposes owing to the residual concavity in the corresponding stationary-phase curves used as references; this presumably reflects failure of some cells to terminate c1 and c2 replication even long after cessation of net growth . Accordingly , we estimated ori / ter ratios from the raw data plots ( see legend ) , and the C values derived from them are shown in Fig 2B . For c1 and c2 these are 58 and 51 ( average ) minutes respectively , representing a replication speed of about 33 kb per minute , comparable to that of E . coli growing at the same rate ( ~36kb/min at τ = 72 min [28] ) . In the case of c3 the essentially flat stationary-phase curve allowed normalization without distortion of ori / ter ratios . The data scatter and the shallow gradients render definition of the replication terminus approximate and prevent a similarly simple estimate of C . If the bp gradients indeed result from fast and slow forks meeting within the left half of c3 , C would effectively be set by the slow fork , at about 16 minutes . These times for the duration of c1 and c2 replication are readily accommodated within the ~75 minute cell cycle but cannot tell us when in the cycle replication is initiated . Direct observation of the positioning and separation of the replicated ori regions , although not a direct indicator of initiation , should enable us to address this question . To visualize the origin regions of the replicons in Bcen we used the binding of fluorescent derivatives of the native ParB proteins to their cognate parS clusters adjacent to each origin . We saw no indication that the ParB fusion proteins of c1 and c2 interfere with indigenous wt ParB function; indeed they showed partition activity equivalent to that of the native ParB proteins ( S5A and S5B Fig ) , and none of the abnormalities stemming from deletion of par loci or provision of excess parS ( see below ) . The fluorescent ParBc3 derivative was defective in partition function ( S5C Fig ) as well as aberrantly localized in the presence of wt ParBc3: to visualize the c3 origin region we used the phage P1 ParB-parS system of Li & Austin [29] Exponential-phase cultures of Nel13 derivatives , each carrying a pMLBAD plasmid from which one of the parB::fps is expressed , were sampled for microscopic observation . In nearly all cells , whether grown in MGCC , LB or SOB medium , the three chromosomes were seen as a single centrally-located focus or as two foci positioned roughly symmetrically about the midpoint ( Fig 3A and 3C ) , in agreement with the Southern hybridization results ( S4 Fig ) . To estimate cell cycle parameters we examined the distribution of ori-proximal foci as a function of the length of cells growing at 30°C in MGCC with doubling times of ~110minutes ( equivalent to ~76 mins at 37°C ) ( Fig 3B ) . The range of cell length over which replicated origins begin segregation is delimited , observationally , by the first appearance of two-focus cells and the end of the one-focus cell cluster , as arrowed . How these segregation events are placed within the cell cycle cannot be directly estimated from these data , but can be inferred on the assumption that the approximately two-fold range of newborn cell sizes reported for E . coli and B . subtilis [30–32] applies also to Bcen and read from the abscissa as 1 . 2–2 . 4 μm . The ori regions of c1 and c2 segregate within the length limits of 1 . 5- ~2 . 4 μm and 1 . 6- ~2 . 6 μm respectively . This behaviour indicates that segregation of c1 and c2 oris begins early in the cell cycle and that it occurs within a range of cell lengths , ~1 μm , similar to the length range of newborn cells , implying close coupling with the cell cycle . It notable also that segregation of the plasmid-like origin of c2 is as tightly coupled as that of the c1 chromosomal origin . Segregation of c3 origins is not seen until cells are 0 . 3–0 . 4 μm longer than the first to show c1 and c2 segregation , implying cell cycle phasing . However , its longer segregation range , 1 . 6 μm ( 1 . 9–3 . 5 ) , suggests that any coupling of c3 segregation to the cycle is less strict than for the larger replicons . Plasmid p1segregation is first seen in cells ~0 . 2 μm shorter than the first to segregate c3; its length-at-segregation range , 1 . 5 μm ( 1 . 8–3 . 3 ) , is similar to that of c3 and also indicates relatively loose coupling to the cycle . These data are uninformative as to the time at which replication is initiated , but an indication of whether segregation follows initiation immediately or after a delay can be gleaned from DNA/cell data . Assuming that initiation at the c1 origin occurs at a fixed point in the cycle , as is the case for those bacterial chromosomes studied [14 , 33 , 34] , and that initiation at the c2 origin is similarly phased with the cycle , as our data suggest ( Fig 3 , and see Discussion ) , we can calculate that replication of c1 and c2 is initiated shortly before or at the end of the preceding cycle , as derived in S2B Table , implying that a significant interval probably separates initiation and visible segregation of the c1 and c2 origins seen here . Replotting of the 2-focus data as a fraction of cell length and as interfocal distances ( Fig 3A ) confirms two aspects of segregation behaviour: the widths of focus distribution are more restricted for c1 and c2 than for c3 and p1 , implying more precise positioning of the former two , and the average distance moved towards the poles follows the order in which partition was initiated , indicating an order of segregation ages , c1 < c2 <p1≤ c3 . The difference in distributions of c1 and c2 two-focus cells over the length range within which segregation occurs ( 1 . 5–2 . 6 μm , see above ) was compatible with the order above but for c1 and c2 was at the limit of statistical significance ( S6 Fig ) . Because it appeared possible that experimental variability arising from our use of independent cultures for visualizing each replicon could have reduced the reliability of comparative segregation times , we again measured focus positions , this time using comparison at the single cell level with pairs of replicons marked at their origins . The results shown in Fig 4A confirm that segregation of the c1 origin generally precedes that of c2 , the c2 origin that of c3 , and the p1 origin also , but more narrowly , that of c3; examples of the cells observed are shown in Fig 4B . These data also confirm the relative average destinations of segregated origins shown in Fig 3A . If the segregation order is correct it should be reflected in the relative frequency of focus combinations . The tabulation of cells in each focus category ( Fig 4C ) bears this out . Doubling of c1 foci generally precedes that of c2 , c2 nearly always precedes c3 , and on average p1 also precedes c3 . These data can be used to estimate average cell age at segregation for each replicon ( S2C Table ) . The order is not absolute , however . In particular , in a minority of the cells the c2 origins segregated before the c1 origins , behaviour which is concealed in the focus distributions . Possible explanations include a looser regulation of c2 initiation such that it occasionally precedes that of c1 and , more likely , an occasional prolonging of the c1 initiation-segregation interval . Having defined in outline the main features of the Bcen cell cycle , and knowing that the ParABS systems of each replicon can act independently and specifically to partition plasmids in E . coli [10 , 21] , we next asked whether the Par systems also behave this way in their natural host , where a possible involvement in other processes might influence the coordination of segregation . Mutation of chromosomal ParABS systems do not only impair segregation but also affect replication , DNA compaction , cell division and viability [19 , 35–39] through direct , functional interaction of Par proteins with the regulators of these processes , e . g . the initiator , DnaA [11 , 13] , the condensin , SMC [40 , 41] , the division inhibitor , MipZ [42] . We explored the range of roles that the Bcen Par systems play by observing the effects of nullifying each system on growth , morphology , replication and partition . ParAB function was disrupted in two ways: by deletion of parA or parAB from each replicon , and by introduction of excess parS sites to deplete ParB available to the chromosomal parSs ( see Materials & Methods ) . The deletion in the ΔAc1mutant is polar on parB , reducing the ParB protein level to <5%that of wild type ( S7 Fig ) and rendering this strain phenotypically ParAB-minus . Excess parS sites were introduced either singly or as the natural cluster ( for c2 , c3 and p1 ) on the vector pMMBΔ ( 10–15 copies per cell ) . To obtain reproducible growth of and focus formation by mutant cells it was necessary to use growth media other than the MGCC used so far , as noted in the figure legends .
Our aim in undertaking this study has been to understand how bacteria with split genomes organize the maintenance of their multiple replicons within the cell cycle—how they time replication and segregation of plasmid-like chromosomes to avoid division delays and how they programme partition to avoid entanglement . For B . cenocepacia J2315 , the results obtained point to a basic strategy—successive activation of the replication and segregation of each origin from a single locale , the cell midpoint ( Fig 8 ) . This maintenance mode resembles that of the only other multipartite genome for which these issues have been studied , that of V . cholerae , insofar as segregation of replicon copies is staggered through the cycle , but differs from it in that the resting origins of V . cholerae Chr1 and Chr2 are physically distant from each other , at the cell pole and midpoint respectively . The Bcen succession proceeds from segregation of the primary chromosome early in the cycle to that of secondary chromosomec2 shortly afterwards and then of c3 and the plasmid p1 at later , less well-defined cell ages ( Figs 3 and 4 ) . Notably , the immediate destinations of the replicated origins follow this order of segregation ( Fig 3A ) , and only at around the time of cell division do all origins assume the midcell location seen in single-focus cells . Time-lapse monitoring of ori and ter movement will be needed to define this repositioning in more detail . Although we have presented the temporal and spatial aspects of multi-replicon maintenance as separate , for Bcen they may be intimately related . In a newborn daughter cell where all four ori-par regions are close together at midcell , attempts at simultaneous segregation could be self-defeating . The ParA proteins of all four systems are of the Walker-box ATPase type , like those of the F and P1 plasmids and of bacterial chromosomes which appear to use the nucleoid surface to modulate transitions in ParA conformation essential to the partition mechanism [46–49] . It is unclear whether two partition processes of this type can simultaneously use overlapping nucleoid patches and move their origins over them . Staggering partition of the three large replicons should help avoid such scenarios of physical interference , thus improving partition efficiency . Participation of the c1 Par system in this temporal separation is suggested by the apparent delay in c1initiation timing and altered c3 bp frequency gradients in the ΔparAc1strain ( Fig 6 ) . The precedents for functional interaction of ParA proteins with DnaA [11 , 13] and the presence of DnaA-box clusters in the Bcen chromosomal origins lends credence to this proposal . A subsidiary aim of this study was to verify that the specificity of action which each Par system had manifested previously in E . coli and in vitro [10 , 21] applied also in the systems' native cells . The importance of this verification was underlined by our discovery of overlapping specificities in other Burkholderia species [21] . The defects in origin positioning that appeared only in cells carrying the cognate parA ( B ) deletion ( Fig 5 ) confirmed that this was so . It is unlikely that the par systems alone are primarily responsible for timing the segregation of replicated origins . Rather , it is through their regulatory role in initiation , indicated by changes of bp frequency gradients in the ΔAc1mutant ( Fig 6 ) that they could contribute to replication timing . The distinction between global effects on initiation and a specific role in partition is mirrored at the temporal level—our preliminary estimate of initiation age suggests an interval amounting to ~15% of the cycle between initiation and origin segregation ( S2B Table ) . Even if future work proves this accurate , it applies only to c1 and c2 , the replicons that contribute significantly to the genome mass on which the calculation is based . We have no results that bear on whether newly replicated c3 and p1 origins remain colocalized or cohered for a period before segregation . We do not know , for example , whether clustering of p1 siblings , a phenomenon held responsible for sub-copy number focus numbers of several E . coli plasmids [50] , artificially prolongs the apparent age at segregation of p1 seen in Fig 3 . It is reasonable to question the term "chromosome" as a title for large secondary replicons . In the case of Bcen c2 and c3 the issue is not settled . First , their complement of essential genes is very limited [20] , suggesting that acquiring a few of them ensured that the replicon became indispensable and removed any selective pressure to accumulate more . Moreover , none of the acquired core genes is a constant feature of secondary chromosomes , as would be expected if the replicons were chromosomal in the eukaryotic sense . Second , and more importantly in the present context , the replication control systems resemble those of plasmids with a specific initiation regulator and iteron-like binding sites rather than that of primary chromosomes , for which the near-universal DnaA acts as the main regulator . Likewise , the parABS partition systems are variable and specific rather than based on the "universal centromere" [51] of the main chromosome . To reflect these characteristics , as well as necessity for cell viability and a GC content close to that of the chromosome , the term "chromid" has been proposed as a replacement for the often ambiguous labels—secondary chromosome , megaplasmid , etc—in use till now [52] . This sensible proposal appears at first sight applicable to c2 and c3 . Nevertheless , certain criteria were not taken into account in the definition of chromids . The asymmetry of KOPS distribution centred on a dif site ( Fig 1 ) is characteristic of chromosomes . Likewise , linkage to the cell cycle can reasonably be considered a chromosomal attribute , and has been in the case of V . cholerae Chr2 [14] . As pointed out above , a replicon size comparable to that of the main chromosome obliges cycle-phased replication , and in this sense is a chromosomal characteristic . Our data ( Figs 3 and 4 ) indicate that in general , segregation , and presumably the prior replication , of c2 are as well phased with the cell cycle as they are for c1 . On this basis we propose that the c2 replicon of Bcen J2315 qualifies as a chromosome . The c3 replicon , on the other hand , does not , since its segregation is only loosely timed with respect to the cycle . Moreover , its essentiality is unclear . Agnoli et al [53 , 54] obtained from many isolates , representing16 species of the 17 in the B . cepacia complex , derivatives cured of c3 whose growth properties were essentially unchanged , demonstrating that Burkholderia c3 replicons are in general neither chromosomes nor chromids but simply plasmids . However , B . cenocepacia J2315 was not among these species . The status of its c3 replicon has still to be determined . Whether or not synchronization of replication with the cell cycle justifies elevation of c2 to chromosome status , the question of how such coupling came about is important . The simplest answer would be that an inherently synchronized plasmid was the progenitor of the present-day c2; the view that plasmids replicate at random through the cell cycle is based on experiments performed on only a few E . coli plasmids [6–9] , which might not be representative of the plasmid universe . Alternatively , a synchronizing host-plasmid interaction might have been selected once a c2 ancestor , already essential , had expanded to a size problematic for the cell cycle [55] . A further possibility is linkage to replication of the c1 chromosome via a common regulator . Inspection of the c1 and c2 ori regions ( S1 and S2 Figs ) provides no obvious evidence for shared or overlapping regulatory processes ( apart from the purely speculative roles of the clustered 7mers ) . A synchrony element of this type has recently been discovered in the V . cholerae genome [56] . A 70bp sequence situated 0 . 8kb from the origin on one 1 . 5kb arm of the primary chromosome ( Chr1 ) was found to modify the replication regulator protein RctB of Chr2 in such a way as to stimulate Chr2 replication . Doubling of the 70bp element by replication was proposed to trigger Chr2 initiation , thus bringing Chr2 replication timing under the ultimate control of DnaA and coordinating it with the cycle . The possibility that an analogous mechanism links c1 and c2 replication in Bcen is worth exploring , although the observation that c2 foci occasionally double before c1 ( Fig 4 ) suggests that such a mechanism is not an absolute requirement . Although the c2 and c3 replicons appear to have evolved beyond the simple plasmid state , some of our data betray persistence of typical plasmid-like behaviour . Most striking is the asymmetry seen in the left and right arm bp frequency profiles ( Fig 2 ) . We favour the idea that this asymmetry reflects frequent failure of bidirectionality , such that c2 and c3 occasionally replicate by reverting to the unidirectional replication mode that presumably characterized their plasmid ancestors [57] . This observation has an important corollary—that a crucial component of the transition of a replicon from low-copy number plasmid to chromosome lifestyle is acquisition of the ability to replicate bidirectionally , thus halving the replication time and allowing the progressively expanding replicon to be replicated within the cell cycle . The alternative of rephasing initiation to allow it on not-yet terminated replicons , as seen in rapidly-growing E . coli , might not be compatible with an essentially iteron-based replication control system . Apart from the demonstrated necessity of each replicon's ParABS system for its own partition ( Fig 5 ) , the appearance of several phenotypes specific to one or other of the disturbed Par systems suggests wider involvement in cell processes . Perhaps the clearest evidence for this is the altered replication of c1 and c3 in the mutant lacking the c1 ParA and ParB proteins ( Fig 6 ) . It suggests that in Bcen , as in B . subtilis and V . cholerae , the main chromosomal Par system helps regulate initiation . A further abnormality hinting at an expanded role for this system is displayed by cultures of cells in which c1 ParB is depleted by deletion of ParA or by parSc1 sequestration . About 5% of the cells form a triplet , one of whose terminal cells decondenses its nucleoid , elongates and eventually bursts , while the nucleoids of the two normal-sized partners are mis-positioned and show some lesser degree of compaction anomaly ( S8 Fig ) . The mis-segregation and decompaction are reminiscent of the failure to load the condensin SMC at the B . subtilis and S . pneumoniaer eplication origins upon depletion of their ParB proteins [40 , 41 , 58] , and suggests that the c1 Par system functions likewise in Bcen , a Gram-negative species . Deletion of the c2 parAB operon also resulted in a specific phenotype , the reduction in average length and width of cells to about 70% of normal dimensions , and thus to an average cytoplasmic volume about one-third that of wild-type . Such contraction of the space available to the nucleoid might increase segregation deficiency beyond that specifically due to failure of c2 partition and contribute to the high level of anucleate cells generated in ΔparABc2 cultures ( Fig 7B ) . Participation of parABSc2 in regulating c2 initiation , analogous to that reported for the Chr2 chromosome of V . cholerae [39 , 59] , might also contribute . ParSc2 interference does not produce the phenotype , implying that ParAc2 influences the mechanisms governing cell size or division . Disruption of all chromosomal Par systems retarded growth , but loss of c3 Par function was particularly severe . The parAB deletion abolished growth on LB medium and imposed the longest colony-appearance delay on SOB medium , while the full parSc3 locus provoked a delay twice as long as the next most severe ( parSc1; Fig 7B ) as well as high cell fragility . These observations suggest a specific effect of parABSc3 on cell physiology . However , our data do not allow us at present to distinguish clearly between direct implication of parABS in host processes and inhibition of these processes by toxin-antitoxin system activation following c3 mis-segregation . All Bcen replicons carry toxin-antidote modules [54] , and several of those in c3 appear important for the stability of their own replicons [53] . How these might account for the growth deficiencies seen here remains to be analyzed . Such TA mechanisms could underlie the dramatic loss of cell integrity that follows failure of Chr2 segregation in V . cholerae [59] . Defining the roles of the c2 and c3 Par systems in cell growth and morphology , as well as exploring their involvement in the cell cycle , is one of two major tracks towards elucidation of genome management in Bcen indicated by this study . The other is investigation of the mechanisms that enable the c1 Par system to act specifically in partition of its own chromosome and generally in regulating initiation of replication of all three chromosomes . Probing these aspects should throw light on the reciprocal adaptations that enabled ancestral cells and progenitor plasmids to evolve towards the multipartite genome states we now observe .
E . coli strain DH10B [60] was used as the primary transformation recipient for plasmid construction , and the dam dcm strain SCS110 ( Stratagene ) for propagating plasmids destined for Bcn . The basic Burkholderia isolate is B . cenocepacia J2315 , genomovar III of the ET12 lineage , used as the UK cystic fibrosis reference strain [20] . The antibiotic-sensitive derivative , Nel13 , was used for most experiments; instances of J2315 use are noted . Nel13 was obtained by deletion from J2315 c1 of a mexAB-oprM locus ( mex1 ) that encodes an RND efflux pump ( see ref . 61 , where the strain is called ΔMex1 ) . Deletion of the par loci was carried out by allele replacement: cells transformed with suicide vectors carrying the desired deletion were screened for loss of the integrated-then-excised vector by antibiotic sensitivity and of parA ( B ) by PCR ( see S3 Table for details ) . The same approach was used to insert the parS site of phage P1near the c3 origin , as described [61] , yielding strain Nel35 . Plasmids used to produce fluorescent fusion derivatives of ParB proteins were constructed by first inserting the GFP and mCherry coding sequences ( gfp and chfprespectively ) , tailed at their 5' ends by an NdeI site , into the SmaI site downstream of paraBAD in pMLBAD [62] . The parB genes of the four Bcen replicons were then amplified using primers with EcoRI and NdeI ends , enabling in-frame fusion to gfp and chfp in the pMLBAD vectors; the parBp1 gene had been mutated to remove the internal parSp1 site . For marking the parSP1 site in Nel35 , the gfp::parBP1fragment cut from pALA2705 with BsrBI and HindIII was inserted between the SmaI and HindIII sites of pMLBAD , to make pDAG825 . Plasmids expressing tandem parB::fp genes were made by inserting NheI-HindIII fragments carrying one between the XbaI and HindIII sites in plasmids carrying the other , giving pDAG845 ( parBc1::chfp-parBc2::gfp ) , pDAG846 ( gfp::parBP1-parBc2::chfp ) , and pDAG847 ( gfp::parBP1-parBp1::chfp ) . Plasmids for providing excess parS sites were made as described [21] , by replacing the EcoRI-MluI and ApaI-HpaI fragments of lacI in pMMB206 [63] with fragments carrying single parS sites and parS clusters respectively . The control plasmid , pMMBΔ , is deleted of the EcoRI-MluI lacI fragment , which inhibits Bcen growth . Media used were MGCC , composed of M9 salts ( 0 . 42M Na2HPO4 , 0 . 22M KH2PO4 , 0 . 009M NaCl , 0 . 018M NH4Cl , 1mM MgSO4 , 0 . 1mM CaCl2 ) , 3 . 4mM Na3citrate , 0 . 1% glucose , 0 . 2% Casamino acids , 0 . 04% tryptophan; MglyC , being MGCC with glycerol substituted for glucose; Luria-Bertani ( LB ) medium ( 1% NaCl version ) ; and SOB . As an anti-contamination measure , media were routinely supplemented with gentamicin at10μg/ml , a concentration which does not affect Bcen growth . Antibiotics for selecting entry of plasmids into Nel13 and J2315 were used at , respectively , ( μg/ml ) chloramphenicol 40 , 80; trimethoprim 200 , 600; tetracycline 200 , 400 . Cultures were grown with aeration at 37°C or , for fluorescence microscopy , at 30°C . Growth rate in liquid medium was determined by periodic measurement of the OD600 of samples from cultures grown exponentially at OD < 0 . 2 for at least three generations . Bcen culture doubling times were observed to be less reproducible than those of E . coli , and showed day-to-day variation in all media , regardless of pre-culture history , number of generations in exponential phase or presence of antibiotics . Doubling times ( minutes ± standard deviation ) of Nel13-based strains were: SOC—60 ± 5 ( 37° ) , LB—108 ± 8 ( 30° ) , 75 ± 4 ( 37° ) ; MGCC—110 ± 13 ( 30° ) , 76 ± 5 ( 37° ) ; MglyC—144 ± 9 ( 30° ) , 91 ± 13 ( 37° ) . Doubling times of J2315-based strains were 3–4 minutes longer in all media . The colony-appearance assay consisted of spreading cultured or transformed cells on solid LB or SOB medium , incubating them at 37°C and counting the colonies two or three times per day; the time at which colony number reached its maximum was taken as the colony-appearance time for purposes of comparison . Viability was estimated by applying cells from dilutions of SOB cultures to LB agar in two ways—by spreading using glass beads , and as a 10μl drop—followed by incubation at 37°C; colony counts per OD unit were calculated . Samples of 10μl taken from exponential SOB cultures at OD600 ~0 . 2 were applied to polylysine-coated slides and allowed to dry at room temperature . After three rinses in M9 salts and drying in air , the cells were fixed with a drop of methanol and allowed to dry , then covered with 5μl 2μg/ml DAPI and a cover slip and viewed by phase-contrast and fluorescence microscopy using a DAPI broad filter . In most experiments , MGCC medium was inoculated from freshly-grown colonies at a concentration which ensured cells were still growing exponentially following overnight incubation at 30°C . Cells from the overnight cultures were diluted to OD600 0 . 05 in 25ml MGCC and incubated to OD600 ~0 . 10 . ParB::FP production was then induced by addition of arabinose . Arabinose concentrations and induction periods appropriate for optimum signal:background ratio were determined empirically , according to whether ParB::FP was used for origin marking , whether one ParB::FP was being produced or two simultaneously , and whether normal or disrupted Par function prevailed in the cells observed . Induction was usually arrested by addition of further glucose or by resuspension in MGCC , but occasional omission of this step proved not to be detrimental . Cells in which the c3 Par system was disrupted ( ΔparABc3 and extra parSc3 ) grew erratically in MGCC but reproducibly in SOB; the latter medium was used in this case . Microscopy Culture samples were centrifuged and the cells resuspended in ~1/30 volume of medium . 1–2 μl was then applied to a 1% agarose-M9 salts layer on a microscope slide , spread by application of a coverslip and viewed under oil-immersion by phase contrast and epi-fluorescence microscopy . Microscopy was carried out in two laboratories: that of Dr J . Errington ( Centre for Bacterial cell Biology , Newcastle-upon-Tyne; Fig 3 ) and that of the authors' institute ( Figs 3–6 ) . At the former , cells were observed with a Zeiss Axiovert M200 microscope equipped with a 300W Xenon lamp and a Zeiss Plan-Neofluar 100x/1 . 30 objective . The filters used were: Chroma 49002 ET-EGFP ( exciter ET470/40X , dochroic T495LP , emitter ET525/50M ) and 49008 ET-mCherry ( ET560/40X; T585LP; ET630/75M ) , Schott UV GG385 and Schott IR KG5 . Images were captured with a Photometrics Coolsnap HQ monochrome camera and analyzed using MetaMorph V . 6 . 2r6 . At the latter , cells were observed using a 100x oil-immersion objective ( Plan Achromat , 1 . 4 NA; Olympus ) the equivalent was an Olympus X81 wide-field inverted microscope equipped with an Olympus phase-contrast 100x/1 . 4 objective . The light source was a monochromator ( Polychrome V; Till Photonics GmbH ) with a 150W Xenon lamp used with a 15nm bandwidth . For two-colour experiments , multiband dichroic mirrors ( Chroma BGR 69002 ) were used , and specific single-band emission filters ( GFP 520/40 , mCherry 632/60 ) were mounted on a motorized wheel . Images were captured with a Roper Coolsnap 2 camera and processed using Metamorph and ImageJ software . Upon deposition on slides , Bcen cells tend to amass to form large groups in which cell dimensions cannot be accurately determined; accordingly we included only isolated cells or those in small groups in analyses of focus position . Cell poles were located by contrast difference in the grey-scale and the line connecting them was drawn using the ImageJ Straight function . Cells with septal constrictions were considered to be unitary unless septation was estimated to be more than two-thirds complete or the nascent cells were not aligned . Focus positions were determined as local fluorescence maxima at least two-fold higher than background using the plot-profile function of ImageJ . Distances were obtained from pixel size ( 64 . 5 nm ) . The initial pole-focus distance was determined by perpendicular projection to the line ( Straight ) from the focus maximum closest to a pole; second focus position was measured relative to the same pole . Cultures of Nel13and its ΔparA1c1 derivative FBP7 were harvested after three generations of balanced exponential growth ( 72 and 136 minute doubling times respectively ) to OD600 ~0 . 2 and , for Nel13 , after 8 hours of incubation in stationary phase . DNA was purified as described and subjected to high-throughput sequencing by the Imagif Platform ( Gif-sur-Yvette ) using Illumina technology . Base-pair frequencies from the read density profiles were binned ( 10kb—c1 and c2 , 1kb—c3 ) to generate base-pair gradients for the three chromosomes . | Unlike higher organisms , bacteria typically carry their genetic information on a single chromosome . But in a few bacterial families the genome includes one to three additional chromosome-like DNA molecules . Because these families are rich in pathogenic and environmentally versatile species , it is important to understand how their split genomes evolved and how their maintenance is managed without confusion . We find that mitotic segregation ( partition ) of all three chromosomes of the cystic fibrosis type strain , Burkholderia cenocepacia J2315 , proceeds from mid-cell to cell quarter positions , but that it occurs in a sequential manner , from largest chromosome to smallest . Positioning of each chromosome is specified solely by its own partition proteins . Nevertheless , the partition system of the largest chromosome appears also to play a global role in the cell cycle , by modulating the timing of initiation of replication . In addition , disrupting the partition systems of all three chromosomes induced specific cell abnormalities . Hence , although such bacteria are governed mainly by the largest , housekeeping chromosome , all the Par systems have insinuated themselves into cell cycle regulation to become indispensable for normal growth . Exploration of the underlying mechanisms should allow us to understand their full importance to bacterial life . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"bacteriology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"cell",
"cycle",
"and",
"cell",
"division",
"plasmids",
"cell",
"processes",
"pathogens",
"microbiology",
"vibrio",
"plasmid",
"construction",
"vibrio",
"cholerae",
"dna",
"replication",
"genetic",
"elements",
"bacterial",
"genetics",
"forms",
"of",
"dna",
"dna",
"construction",
"dna",
"molecular",
"biology",
"techniques",
"microbial",
"genetics",
"bacteria",
"bacterial",
"pathogens",
"microbial",
"genomics",
"research",
"and",
"analysis",
"methods",
"bacterial",
"genomics",
"chromosome",
"biology",
"medical",
"microbiology",
"microbial",
"pathogens",
"molecular",
"biology",
"burkholderia",
"cenocepacia",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"burkholderia",
"genomics",
"mobile",
"genetic",
"elements",
"organisms",
"chromosomes"
]
| 2016 | Orderly Replication and Segregation of the Four Replicons of Burkholderia cenocepacia J2315 |
Kaposi’s sarcoma ( KS ) , a highly disseminated tumor of hyperproliferative spindle endothelial cells , is the most common AIDS-associated malignancy caused by infection of Kaposi’s sarcoma-associated herpesvirus ( KSHV ) . KSHV-encoded viral interferon regulatory factor 1 ( vIRF1 ) is a viral oncogene but its role in KSHV-induced tumor invasiveness and motility remains unknown . Here , we report that vIRF1 promotes endothelial cell migration , invasion and proliferation by down-regulating miR-218-5p to relieve its suppression of downstream targets high mobility group box 2 ( HMGB2 ) and cytidine/uridine monophosphate kinase 1 ( CMPK1 ) . Mechanistically , vIRF1 inhibits p53 function to increase the expression of DNA methyltransferase 1 ( DNMT1 ) and DNA methylation of the promoter of pre-miR-218-1 , a precursor of miR-218-5p , and increases the expression of a long non-coding RNA OIP5 antisense RNA 1 ( lnc-OIP5-AS1 ) , which acts as a competing endogenous RNA ( ceRNA ) of miR-218-5p to inhibit its function and reduce its stability . Moreover , lnc-OIP5-AS1 increases DNA methylation of the pre-miR-218-1 promoter . Finally , deletion of vIRF1 from the KSHV genome reduces the level of lnc-OIP5-AS1 , increases the level of miR-218-5p , and inhibits KSHV-induced invasion . Together , these results define a novel complex lnc-OIP5-AS1/miR-218-5p network hijacked by vIRF1 to promote invasiveness and motility of KSHV-induced tumors .
Kaposi’s sarcoma-associated herpesvirus ( KSHV ) , also known as human herpesvirus 8 ( HHV-8 ) , is a double-stranded DNA virus , which belongs to γ-herpesvirus . KSHV was initially identified in an AIDS-associated Kaposi’s sarcoma ( AIDS-KS ) lesion , and has since been strongly linked to Kaposi’s sarcoma ( KS ) , primary effusion lymphoma ( PEL ) , a subset of multicentric Castleman’s disease ( MCD ) , and KSHV-associated inflammatory cytokine syndrome ( KICS ) [1] . Like other herpesviruses , the life cycle of KSHV consists of two phases , latent and lytic phases , both of which contribute to KSHV-induced pathogenesis , tumorigenesis and angiogenesis [2 , 3] . KSHV genome contains over 90 open reading frames [4] , some of which are homologous to human genes . To establish a successful persistent infection , KSHV encodes these homologous proteins to regulate cell growth , immune response , inflammatory response and apoptosis , and thus escape the immune antiviral response of host cells [5] . Moreover , these homologous proteins are also in favor of KSHV-induced tumorigenesis . For examples viral interferon regulatory factors ( vIRFs ) [6] , viral interleukin-6 ( vIL-6 ) [7] , viral G protein-coupled receptor ( vGPCR ) [8] , viral Bcl-2 ( vBcl-2 ) [9] , viral FLICE inhibitory protein ( vFLIP ) [10] and viral cyclin ( vCyclin ) [11] have been shown to be pro-oncogenic or promote tumorigenesis . The cellular IRFs ( IRFs 1~9 ) are a family of cellular transcription proteins that regulate the expression of interferon and interferon-stimulating genes ( ISGs ) in innate immune response , among which IRF3 and IRF7 play key roles in the induction and secretion of type I interferon [12] . vIRF1 ( 449 amino acids ) , as one of the KSHV vIRFs ( vIRF1 to vIRF4 ) , is encoded by KSHV ORF-K9 , which has 26 . 6% and 26 . 2% of protein homology to cellular IRF3 and IRF7 , respectively [13] . vIRF1 has been shown to compete with IRF3 to interact with CBP/p300 coactivators by blocking the formation of CBP/p300-IRF3 complexes , thereby inhibiting IRF3-mediated transcription and signal transduction of type I interferon [14] . However , vIRF1 could not block IRF-7-mediated transactivation [14] . In the other hand , vIRF1 represses tumor suppressor gene p53 phosphorylation , leading to an increase of p53 ubiquitination by reducing ATM kinase activity [15]; vIRF1 could also directly bind to p53 and effectively inhibit p53-mediated apoptosis by reducing its acetylation and inhibiting the transcription of p53 activation [16 , 17] . In addition , vIRF1 restrains TGF-beta signaling via direct interaction with Smads ( Smad3 and Smad4 ) to disturb Smad3/Smad4 complexes from binding to DNA and suppresses IRF-1-induced CD95/CD95L signaling-mediated apoptosis [18 , 19] . As the first identified oncogenic protein encoded by KSHV , vIRF1 has been reported to transform mouse embryonic fibroblasts ( NIH3T3 ) cells [6] , however , its role in KSHV-induced tumor invasiveness and motility and its underlying mechanism remains totally unclear . Less than 2% of the human genome encodes protein-coding genes , while the vast majority of the genome is transcribed as non-coding RNAs [20] . Based on the size , non-coding RNAs could be vaguely divided into three groups: microRNAs ( miRNAs ) , long non-coding RNAs ( lncRNAs ) , and circular RNAs ( circRNAs ) [21] . Many miRNAs ( ~22 nucleotides in length ) have been well-characterized and shown to repress gene expression by inhibiting the translation or destabilization of mRNA transcript via binding to mRNA sequences [22] . LncRNAs ( >200 nucleotides in length ) have indispensable roles in diverse biological processes , including chromatin remodeling , X chromosome inactivation , genomic imprinting , nuclear transport , transcription , RNA splicing and translation [23–25] . A growing volume of literatures support the notion that both lncRNAs and miRNAs could function as tumor suppressors or oncogenes involved in the regulation of cell proliferation , metastasis , apoptosis , and invasion [24 , 26 , 27] . More interestingly , emerging evidence indicates that numerous lncRNAs might act as competing endogenous RNAs ( ceRNAs ) that competitively bind miRNAs , hence exerting influence on posttranscriptional regulation [28] . Recently , several oncogenic viruses have been shown to encode lncRNAs and are thought to participate in enhancing viral replication , promoting oncogenesis and contributing to pathogenesis [29–31] . KSHV encodes an lncRNA , known as polyadenylated nuclear RNA ( PAN RNA ) . PAN is multifunctional , regulating KSHV replication , viral and host gene expression , and immune responses [32–35] . However , whether cellular lncRNAs are involved in the progression of KS is still unknown . In the present work , we aimed to elucidate the role of vIRF1 in cell migration , invasion and proliferation . We found that vIRF1 promoted cell migration , invasion and proliferation by epigenetically silencing miR-218-5p and activating lncRNA-OIP5-AS1 transcription . Further , we uncovered that the crosstalk between miR-218-5p and lnc-OIP5-AS1 contributed to vIRF1-induced cell motility and proliferation via increasing HMGB2 and CMPK1 expression . Our novel findings illustrated a critical role of vIRF1 in the invasiveness , motility and development of KS tumor .
Previous works showed that vIRF1 , as a homologue of cellular IRFs , disrupted immune antiviral response of host cells and contributed to KSHV-induced tumorigenesis [5] . However , its role on tumor invasiveness and motility remains unclear . To determine whether vIRF1 had a role in cell motility , we transduced HUVECs with lentiviral vIRF1 at a MOI of 2 . vIRF1-transduced HUVECs showed a vIRF1 mRNA expression level similar to that of KSHV-infected HUVECs ( S1 Fig , Fig 1A and 1B ) . We then examined the effect of vIRF1 on cell migration and invasion . In transwell migration and Matrigel invasion assays , overexpression of vIRF1 enhanced cell migration and invasion ( Fig 1C , 1D and 1E ) . In plate colony formation assay , vIRF1 clearly enhanced cell proliferation ( Fig 1F and 1G ) . To assess the mechanism mediating vIRF1 promotion of cell motility and proliferation , we performed microarray-based miRNA expression profiling and identified a set of miRNAs that were differentially expressed between vIRF1- and pHAGE-transduced HUVECs ( GEO accession number GSE119034 ) . As a known tumor suppressor [36] , miR-218-5p was significantly down-regulated in vIRF1- transduced cells , and hence was selected for further validation by qRT-PCR . As shown in Fig 1H and 1I , downregulation of miR-218-5p was observed in both vIRF1-transduced and KSHV-infected HUVECs . Then , we sought to determine whether the downregulation of miR-218-5p might contribute to vIRF1 promotion of cell motility , and proliferation . As expected , overexpression of miR-218-5p in vIRF1-transduced HUVECs reversed vIRF1-enhanced cell migration and invasion ( S2 Fig , Fig 1J and 1K ) as well as cell proliferation ( Fig 1L ) . Next , we conducted mass spectrometry analysis to investigate the direct targets of miR-218-5p . As shown in Table 1 , there were a series of proteins that were up-regulated by > 1 . 5 folds in cells overexpressing vIRF1 . Using bioinformatics analysis , we predicted four proteins that might be the potential targets of miR-218-5p , and hence chose them for further luciferase reporter assay . We confirmed that miR-218-5p decreased the 3’UTR reporter activities of both high mobility group box 2 ( HMGB2 ) and cytidine/uridine monophosphate kinase 1 ( CMPK1 ) ( Fig 2A ) , which was further shown in a dose-dependent fashion ( Fig 2B ) . Indeed , overexpression of miR-218-5p suppressed the levels of HMGB2 and CMPK1 proteins in a dose-dependent manner ( Fig 2C ) . Conversely , blocking the miR-218-5p function with a specific inhibitor elevated the expression levels of HMGB2 and CMPK1 proteins in a dose-dependent manner ( Fig 2D ) . To further confirm that miR-218-5p directly targeted HMGB2 and CMPK1 , we performed mutagenesis with miR-218-5p ( Fig 2E and 2F ) . The mutant mimic did not have any effect on the 3’UTR reporter activities of HMGB2 and CMPK1 ( Fig 2G ) , and the levels of HMGB2 and CMPK1 proteins ( Fig 2H ) . Moreover , the mRNA and protein levels of HMGB2 and CMPK1 were significantly up-regulated in cells expressing vIRF1 or infected by KSHV ( Fig 3A–3D ) . In IHC staining , there were more HMGB2- and CMPK1-postive cells in KS lesions than in normal skin tissues ( S3 Fig and Fig 3E ) . Previous studies have shown that HMGB2 and CMPK1 are abundantly up-regulated in various malignant tumors , and are closely associated with tumor development and poor prognosis [37–46] . To determine if upregulation of HMGB2 and CMPK1 was necessary for vIRF1-induced cell motility , and proliferation , we silenced HMGB2 and CMPK1 expression in vIRF1-transduced HUVECs with a mixture of siRNAs , respectively ( S4 Fig ) , and observed diminished vIRF1-induced cell migration , invasion and proliferation ( Fig 4A–4I ) . Moreover , knock-down of HMGB2 and CMPK1 also inhibited KSHV-induced cell migration and invasion ( Fig 4J–4L ) . MiR-218-5p is expressed from two separate loci , pre-miR-218-1 and pre-miR-218-2 , which are co-expressed with their host genes SLIT2 and SLIT3 , respectively [47] . The expression of miR-218-5p depends on the promoter activity of its host genes , and hypermethylation of the promoter inhibits miR-218-5p expression [48] . Therefore , we examined the expression of SLIT2 and SLIT3 . The expression of SLIT2 mRNA was low in both vIRF1-transduced and KSHV-infected cells while no expression of SLIT3 was detected in HUVECs ( Fig 5A and 5B ) . These results indicated that vIRF1 might suppress miR-218-5p by reducing the expression of primary form pre-miR-218-1 and its host gene SLIT2 by epigenetically silencing their promoter . Indeed , the expression of pre-miR-218-1 was significantly suppressed by both vIRF1 and KSHV infection ( Fig 5C and 5D ) . Consistent with these results , methylation-specific PCR showed that the promoter of pre-miR-218-1 was more hypermethylated in vIRF1-expressing cells and KSHV-infected cells than normal cells ( Fig 5E and 5F ) . Moreover , treatment with a potent inhibitor of DNA methylation , 5-aza , not only blocked vIRF1 suppression of miR-218-5p ( Fig 5G ) , but also decreased the expression of its targets HMGB2 and CMPK1 ( Fig 5H ) . These results revealed that vIRF1 silencing of miR-218-5p expression was due to DNA methylation on its promoter . DNA methyltransferase 1 ( DNMT1 ) mediates DNA methylation and has been reported to cause miR-218-5p silencing [49] . We found that DNMT1 was up-regulated by 1 . 62-fold in vIRF1-transduced HUVECs ( Table 1 ) . Moreover , Western-blotting confirmed that DNMT1 protein was up-regulated in vIRF1-expressing cells and KSHV-infected cells ( Fig 6A and 6B ) . Knock-down of DNMT1 expression with specific siRNAs ( siDNMT1 ) reduced the hypermethylation of the promoter of pre-miR-218-1 in vIRF1 expressing cells ( Fig 6C ) , and enhanced the expression of SLIT2 , pre-miR-218-1 and miR-218-5p ( Fig 6D ) . Meanwhile , vIRF1 induced expression of HMGB2 and CMPK1 was also abolished following DNMT1 inhibition ( Fig 6E ) . vIRF1 binds to p53 and represses p53-dependent transcription and apoptosis [16] . p53 transcriptionally suppresses the DNMT1 promoter by interacting with specificity protein 1 ( Sp1 ) and forma complex [50] . Based on these studies , we sought to elucidate whether vIRF1 inhibition of p53-dependent transcription was responsible for vIRF1-induced DNMT1 up-regulation and therefore was involved in the inhibition of miR-218-5p . Indeed , overexpression of p53 in vIRF1-transduced HUVECs reduced the expression levels of DNMT1 , HMGB2 and CMPK1 ( Fig 6F ) , as well as hypermethylation of the promoter of pre-miR-218-1 in vIRF1-expressing cells ( Fig 6G ) , hence causing an increase of both pre-miR-218-1 and miR-218-5p expression in vIRF1-infected HUVECs ( Fig 6H ) . These results indicated that vIRF1-induced miR-218-5p inhibition via aberrant DNA methylation at the pre-miR-218-1 promoter by inhibiting p53 to cause DNMT1 upregulation . Numerous studies have shown that lncRNAs can act as ceRNAs to regulate the functions of miRNAs . To identify lncRNAs which may serve as ceRNAs and interact with miR-218-5p , we utilized online software programs starbase v2 . 0 ( http://starbase . sysu . edu . cn/ ) and LncBase Predicted v . 2 ( http://carolina . imis . athena-innovation . gr/diana_tools/web/index . php ? r=lncbasev2%2Findex-predicted ) to search for lncRNAs that have complementary base pairing with miR-218-5p . Considering the abundance in the cytoplasm , and high score of predicted binding sites , we identified lncRNA OIP5 antisense RNA 1 ( lncRNA-OIP5-AS1 ) as a potential candidate . We found that there were four putative miR-218-5p-binding sites in lnc-OIP5-AS1 ( Fig 7A ) and that lnc-OIP5-AS1 was indeed up-regulated in both vIRF1-transduced and KSHV-infected HUVECs ( Fig 7B and 7C ) . We also found that vIRF1 was capable of activating the luciferase activity of lnc-OIP5-AS1 promoter ( S5 Fig ) . We generated four luciferase reporter constructs , each of which contains only one putative miR-218-5p-binding site . Of these , miRNA-218-5p mimics reduced the luciferase activities of lnc-OIP5-AS1 ( S3 ) and lnc-OIP5-AS1 ( S4 ) reporters ( Fig 7D ) in a dose-dependent fashion ( S6 Fig ) . In contrast , the miR-218-5p mutant mimic lacking the seed sequence did not reduce the luciferase activities of both lnc-OIP5-AS1 ( S3 ) and lnc-OIP5-AS1 ( S4 ) reporters ( Fig 7E ) . To confirm the physical interaction between miR-218-5p and lnc-OIP5-AS1 , we performed RNA pull-down experiments . Biotin-labeled mimics were incubated with HUVECs lysates , isolated with streptavidin agarose beads and then analyzed by RT-qPCR . We observed that lnc-OIP5-AS1 enriched miR-218-5p but not the miR-218 mutant ( Fig 7F ) . These results indicated that lnc-OIP5-AS1 could directly bind miR-218-5p . To determine whether lnc-OIP5-AS1 could act as a ceRNA to abrogate the function of miR-218-5p by releasing its binding to the targeted transcripts , we knocked down lnc-OIP5-AS1 with specific Smart Silencer in HUVECs and performed an RNA immunoprecipitation ( RIP ) assay on Ago2 ( S7 Fig ) . We found that silencing of lnc-OIP5-AS1 in HUVECs increased the HMGB2 and CMPK1 transcripts in the Ago2 complex ( Fig 7G ) . Furthermore , overexpression of lnc-OIP5-AS1 ( S3 ) and lnc-OIP5-AS1 ( S4 ) fragments abolished the inhibition of HMGB2 and CMPK1 3’UTR reporter activities by miR-218-5p ( Fig 7H ) . These results demonstrated the sequestration of miR-218-5p by lnc-OIP5-AS1 , which relieved the inhibition of the HMGB2/CMPK1 transcripts by miR-218-5p . Intriguingly , overexpression of miR-218-5p significantly reduced the level of vIRF1-induced lnc-OIP5-AS1 ( Fig 7I ) . Conversely , inhibition of lnc-OIP5-AS1 reversed vIRF1 inhibition of the expression of both pre-miR-218-1 and miR-218-5p ( Fig 7J ) . We further performed knock down of Dicer to prevent maturation of miR-218-5p from pre-miR-218-1 and then examined the effect of silencing lnc-OIP5-AS1 on miR-218-5p stability ( S8 Fig ) . We found that suppression of lnc-OIP5-AS1 reduced the degradation of miR-218-5p ( Fig 7K ) . As a result , silencing of lnc-OIP5-AS1 attenuated vIRF1-induced DNMT1 , HMGB2 and CMPK1 expression ( Fig 7L ) . These results indicated that lnc-OIP5-AS1 could not only inhibit the function of miR-218-5p by acting as a ceRNA but also reduce the level of miR-218-5p by direct binding to induce miR-218-5p degradation . Consistent with these results , silencing of lnc-OIP5-AS1 inhibited vIRF1-induced cell migration , invasion and proliferation ( Fig 7M and 7N ) . To further dissect the functions of vIRF1 in the context of KSHV genome , we constructed a KSHV mutant with ORF-K9 deleted using a two-step red recombination system as previously described [51–53] . Positive colonies were screened and verified by PCR ( S9A and S9B Fig ) . Restriction analysis showed that the RGB-K9-mutant had a band shift of about 1 . 3 kb compared to the wild-type RGB-BAC16 , indicating that the K9 mutant bacmid was successfully generated ( S9C Fig ) . The RGB-K9-mutant was transfected into iSLK cells and selected to obtain stable producer iSLK cells . As expected , we did not detect the expression of vIRF1 in iSLK-RGB-K9 mutant cells , while the levels of vIRF4 and ORF57 had minimal changes ( S9D Fig ) . Similarly , HUVECs infected by the mutant virus had no vIRF1 expression ( S9E Fig ) but had minimal changes in the levels of vIRF4 and ORF57 ( S9F Fig ) . As expected , expression levels of phosphorylated p53 , acetylated p53 , and p21 were increased in vIRF1_mutant cells compared to those of KSHV_WT virus-infected HUVECs ( S9G Fig ) . Because HUVECs transduced with lentiviral vIRF1 at a MOI of 2 showed a vIRF1 mRNA expression level similar to that of wild type KSHV-infected HUVECs ( S1 Fig ) , we transduced vIRF1_mutant cells with 2 MOI of lentiviral vIRF1 . We found that loss of vIRF1 not only reduced cell migration and invasion ( Fig 8A ) but also decreased the level of hypermethylation in the pre-miR-218-1 promoter ( Fig 8B ) . Importantly , complementation with vIRF1 in vIRF1_mut-infected HUVECs was sufficient to rescue cell migration and invasion induced by KSHV ( Fig 8A ) , and reverse the level of hypermethylation in the pre-miR-218-1 promoter induced by KSHV ( Fig 8B ) . We also observed a significant decrease of lnc-OIP5-AS1 expression , and an increase of miR-218-5p and pre-miR-218-1 expression in the mutant cells compared to both KSHV_wild-type infected cells and vIRF1-transduced mutant cells ( Fig 8C ) . Furthermore , deletion of vIRF1 reduced the expression of DNMT1 , HMGB2 and CMPK1 while complementation with vIRF1 was sufficient to rescue the expression levels of these proteins ( Fig 8D ) . Meanwhile , inhibition of miR-218-5p with a specific inhibitor in both mutant cells and vIRF1-transduced mutant cells increased cell migration and invasion ( Fig 8E ) . Similar increase in cell migration and invasion was also observed in the mutant and vIRF1-transduced mutant cells following overexpression of lnc-OIP5-AS1 ( S3 ) and lnc-OIP5-AS1 ( S4 ) fragments ( Fig 8F ) . Taken together , these results demonstrated that vIRF1 mediated KSHV-induced cell migration , and invasion by down-regulating miR-218-5p and up-regulating lnc-OIP5-AS1 .
KSHV K9/vIRF1 was initially characterized as an early lytic gene . However , subsequent studies have shown that it is also expressed during viral latency . vIRF1 has two transcription start sites , one is distal to the AUG , which is active during latency in PEL , and another is a more proximal site , which is induced upon lytic reactivation [54] . Hence , vIRF1 might have a dual modes of expression during latent and lytic replication [55–57] . Furthermore , K9/vIRF1 mRNA is expressed in all KS tumors ( total 21 KS clinical biopsies ) and preferentially transcribed during latent infection of either endothelial/mesenchymal lineage cells , which strengthens the role of K9/vIRF1 in KS tumorigenesis [58] . In the current study , we found that vIRF1 promoted endothelial cell migration and invasion , as well as proliferation . Further , deletion of vIRF1 from the KSHV genome reduced KSHV-induced cell migration , invasion and proliferation . However , we could not assess the expression level of the endogenous vIRF1 protein because there is currently no vIRF1 antibody available . Despite the limitation , this work still revealed a novel role of vIRF1 in cell migration , invasion and proliferation , which is an important part of KS pathogenesis , particularly in the invasiveness and dissemination of KS tumors . MiR-218-5p , a vertebrate-specific intronic miRNA co-regulated with its host genes SLIT2/SLIT3 , functions as a tumor suppressor by modulating multiple pathways [36] . It is downregulated in numerous human cancers , such as colorectal , prostate , pancreatic , gastric , and thyroid cancers [47 , 49 , 59–63] . The mechanism of silencing miR-218-5p and its host genes , SLIT2/SLIT3 is through promoter hypermethylation . For instance , human papillomavirus type 16 oncogene E6 reduces the level of miR-218-5p and SLIT2 through promoter hypermethylation [64] . However , significance of miR-218-5p in the development of KS remains undefined . In this study , we revealed that both KSHV infection and vIRF1 expression reduced the level of miR-218-5p at least in part by silencing of pre-miR-218-1/SLIT2 via promoter hypermethylation . Further , we demonstrated that vIRF1 increased DNMT1 expression by inhibiting p53 transcriptional activity , leading to a higher level of DNA methylation of the pre-miR-218-1 promoter . Lnc-OIP5-AS1 located at chromosome 15q15 . 1 , known as cyrano , is ~8 , 000 nucleotides in length and abundant in the cytoplasm . It was originally characterized in zebrafish and displayed crucial effects in embryonic nervous system development [65] . It was also reported to play a vital role in embryonic stem cells ( ESCs ) self-renewal maintenance [66] . With regard to its role in cancer , lnc-OIP5-AS1 exhibits multifaceted and complex features . For example , lnc-OIP5-AS1 is shown to be a tumor suppressor and inhibit HeLa cells proliferation by interacting with the RBP HuR to reduce HuR’s availability for binding target mRNAs , or associating with GAK mRNA to impair GAK mRNA stability [67 , 68] . On the contrary , lnc-OIP5-AS1 can exert oncogenic functions in several other cancers . It was consistently up-regulated in renal cell carcinoma , glioblastoma , and gastric cancer [69–71] . Silencing of lnc-OIP5-AS1 repressed YAP-Notch signaling pathway activity leading to decrease of glioma cells’ proliferation , migration in vitro and tumor formation in vivo [72] . Moreover , lnc-OIP5-AS1 was highly expressed in lung adenocarcinoma tissues and cells , and the loss of lnc-OIP5-AS1 inhibited lung adenocarcinoma cell proliferation , migration and invasion [73] . In our report , we found that both KSHV infection and vIRF1 expression increased the expression of lnc-OIP5-AS1 in endothelial cells . Silencing of lnc-OIP5-AS1 suppressed cell migration , invasion and proliferation . Intriguingly , we found that vIRF1 activated the transcription of lnc-OIP5-AS1 , however , the precise mechanism remains unknown . The cross-regulatory interactions between lncRNAs and miRNAs have been recognized to regulate their downstream targets of either lncRNAs or miRNAs [74 , 75] . Several miRNAs including miR-7 [66] , miR-410 [76] , miR-424 [67] , and miR-448 [73] have been identified to interact with lnc-OIP5-AS1 . On the other hand , miR-218-5p has been reported to interact with lnc-MALAT1 , participating in choriocarcinoma growth [77] . In the current study , we revealed the crosstalk between miR-218-5p and lnc-OIP5-AS1 , confirmed a direct interaction between miR-218-5p and lnc-OIP5-AS1 , and unearthed the fateful consequences of this interaction . We showed that lnc-OIP5-AS1 functioned as a ceRNA and sequestered miR-218-5p to relieve its binding and targeting of HMGB2 and CMPK1 transcripts . Further , lnc-OIP5-AS1 could inhibit miR-218-5p expression through regulating miR-218-5p stability . Once entering the RNA-induced silencing complex ( RISC ) , miRNAs become extremely stable due to the protection of both ends of miRNAs by AGO proteins from 3ʹ–5ʹ exoribonucleases-mediated degradation . Therefore , we speculated that lnc-OIP5-AS1 might block miR-218-5p from loading onto AGO proteins , and hence accelerate its degradation . Interestingly , by an unclear mechanism , lnc-OIP5-AS1 also increased DNMT1 expression to promote DNA methylation of the pre-miR-218-1 promoter , leading to decreased level of miR-218-1 . On other hand , the lnc-OIP5-AS1/miR-218-5p interaction also resulted in miR-218-5p suppression of lnc-OIP5-AS1 expression albeit the precise mechanism is unknown . In conclusion , our study revealed that vIRF1 promoted cell migration , invasion and proliferation by a p53- and lnc-OIP5-AS1-mediated down-regulation of miR-218-5p , leading to increased expression levels of its target genes HMGB2 and CMPK1 ( Fig 8G ) . This process was mediated by the complex crosstalk between miR-218-5p and lnc-OIP5-AS1 . These novel findings extend the cross-regulatory network of cellular lncRNAs and miRNAs involved in the pathogenesis of KS .
The clinical section of the research was reviewed and ethically approved by the Institutional Ethics Committee of the First Affiliated Hospital of Nanjing Medical University ( Nanjing , China; Study protocol # 2015-SR-116 ) . Written informed consent was obtained from all participants , and all samples were anonymized . All participants were adults . The iSLK cells were cultured in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 1% penicillin-streptomycin , 1 μg/ml puromycin and 250 μg/ml G418 . The established iSLK-RGB-BAC16 and iSLK-RGB-K9 mutant cells were cultivated in DMEM supplemented with 10% fetal bovine serum ( FBS ) , 1 μg/ml puromycin , 250 μg/ml G418 , and 1 . 2 mg/ml hygromycin B [53] . HEK293T and continuous cell lines human umbilical vein endothelial cells ( catalog #CRL-1730 , ATCC , Manassas , VA , USA ) were maintained as previously described [78] . The latter were only used for plate colony formation assay to evaluate the ability of cell proliferation . Primary human umbilical vein endothelial cells ( HUVECs ) , which were used for all assays except for luciferase and plate colony formation assays , were isolated and cultured as previously delineated [79] . Flag-vIRF1 was cloned by inserting the coding sequences into plasmid pHAGE-CMV-MCSIzsGreen as previously described [78 , 80] . The respective sequences of HMGB2 3’UTR , CMPK1 3’UTR and lncRNA-OIP5-AS1 fragments containing putative miR-218-5p binding sites ( S1: 365–1167; S2: 3078–3539; S3: 4293–4701 and S4: 7775–8169 ) were amplified by PCR and inserted into the pGL3-Control plasmid ( Promega , Madison , WI , USA ) , respectively . The pCMV6-Entry-C-Myc-p53 construct and pCMV6-Entry-C-Myc construct were provided by ORIGENE ( Beijing , China ) . The DNA fragment of lnc-OIP5-AS1 covering –1500 bp to 0 bp of the transcription start site was amplified and subcloned into pGL3-Basic plasmid ( Promega , Madison , WI , USA ) . HUVECs were transfected using the Effectence transfection reagent ( Qiagen , Suzhou , Jiangsu , China ) , while HEK293T cells were transfected using the Lipofectamine 2000 Reagent ( Invitrogen , Carlsbad , CA , USA ) . 5-aza ( Decitabine ) , a potent inhibitor of DNA methylation , was from Selleck Chemicals ( Shanghai , China ) . siRNAs were synthesized from Genepharma ( Suzhou , China ) , the sequences of siRNAs are listed in S1 Table . LncRNA Smart Silencer was obtained from RiboBio ( Guangzhou , China ) . Antibodies against KSHV LANA , HMGB2 , CMPK1 , DNMT1 and Dicer were from Abcam ( Cambridge , MA , USA ) . Anti-Flag was obtained from Cell Signaling Technologies ( Beijing , China ) . Anti-Myc , anti-α-Tubulin , and anti-GAPDH were from Santa Cruz Biotechnology ( Dallas , TX , USA ) . Anti-rabbit immunoglobulin G ( IgG ) , anti-mouse IgG , anti-phosphorylated p53 , anti-acetylated p53 , anti-p53 , and anti-p21 were purchased from Beyotime Institute of Biotechnology ( Nantong , Jiangsu , China ) . Western-blotting analysis was conducted as previously described [81 , 82] . In this study , all Western blotting results were independently repeated at least three times unless otherwise stated . Cell migration , invasion and colony formation assays were executed as previously described [83–85] . Luciferase reporter assay was conducted using the Promega dual-luciferase reporter assay system according to the previous study [86] . Methylation-specific PCR ( MS-PCR ) was adopted using DNA Bisulfite conversion kit ( TIANGEN BIOTECH , Beijing , China ) and Methylation-specific PCR kit ( TIANGEN BIOTECH , Beijing , China ) according to the manufacturer’s instructions . MS-PCR primers were designed as previously described [87] . HUVECs were collected , washed , and re-suspended with lysis buffer ( Thermo Fisher Scientific , Waltham , America ) . After incubating for 5 min , the lysates were precleared by centrifugation at 14 , 000 rpm for 10 minutes , and then were added to streptaviden magnetic beads ( Thermo Fisher Scientific , Waltham , America ) , which were incubated with Biotin-labeled miR-218-5p , miR-218-5p mut 2 , or Neg . Ctrl ( Genepharma , Suzhou , China ) for 4 hours . The bound RNAs in the pull-down material were quantified by qRT-PCR . HUVECs were transfected with lnc-OIP5-AS1 Smart Silencer or its Neg . Ctrl for 48 h , and used for RIP experiments with an anti-Ago2 antibody ( MERCK , Darmstadt , Germany ) and the Magna RIPTM RNA-Binding Protein Immunoprecipitation Kit ( MERCK , Darmstadt , Germany ) , according to the manufacturer’s instructions . The levels of lnc-OIP5 AS1 , HMGB2 or CMPK1 were examined by qRT-PCR . A KSHV mutant with ORF K9 deleted was constructed as described in previous studies [52 , 88] . In brief , using the bacterial artificial chromosome ( BAC ) technology and the Escherichia coli Red recombination system , together with PCR , restriction digestion , and sequencing for strict quality control , a KSHV ORF K9 mutant ( called RGB-K9-mutant ) was constructed by removing K9 coding sequence ( CDS ) from the wild-type recombinant KSHV RGB-BAC16 [53] . RGB-BAC16 and RGB-K9 mutant DNA were transfected into iSLK cells and selected using 1 μg/ml puromycin , 250 μg/ml G418 , and 1 . 2 mg/ml hygromycin B for 3 weeks to establish stable viral producer cell lines , iSLK-RGB-BAC16 and iSLK-RGB-K9 mutant cells . To produce virus stocks for infection , iSLK-RGB-BAC16 and iSLK-RGB-K9 mutant cells were plated at 30 to 40% confluence and induced with both Doxycycline ( Dox ) ( 1 μg/ml ) and sodium butyrate ( NaB ) ( 1 mM ) . After induction for 4 or 5 d , the supernatant was harvested , centrifuged , filtered , and concentrated by ultracentrifugation ( 25 , 000 g at 4°C for 3 h ) using SW32 Ti rotor ( Beckman Coulter Inc , USA ) . The pellet was resuspended , supplemented with 8 μg/mL polybrene and then incubated with 105 HUVECs in a 6-well plate for 4 h . The primers for construction and identification of K9 mutant bacmid were designed as previously described [89] and the sequences of the primers could be found in S2 Table . RNA was extracted using RNA Isolator Total RNA Extraction Reagent ( Vazyme Biotech Co . , Ltd , Nanjing , China ) from cells . Total RNA was reverse transcription by HiScript Q RT SuperMix ( Vazyme Biotech Co . , Ltd , Nanjing , China ) . Real time quantity PCR was performed by AceQ qPCR SYBR Green Master Mix ( Vazyme Biotech Co . , Ltd , Nanjing , China ) . The sequences of the primers for PCR could be found in S3 Table . The extraction of genome DNA was performed by using TIANamp Genomic DNA Kit ( TIANGEN BIOTECH , Beijing , China ) according to the user’s guide . Briefly , cells were trypsinized , and neutralized by 20% FBS DMEM . The suspension was centrifuged and the supernatant was discarded . Mass spectrometry analysis was adopted according to the previous study [86] . The KS clinical specimens were kindly offered by Jiangsu Province Hospital . All samples were anonymized and all participants are provided with informed consent . IHC was carried out as previously described with specific antibodies [85 , 90] . All data are appeared as the means ± SD with at least three replications . Statistical analysis was on account of Student’s t-test and the criterion for statistical significance was adopted as P values of < 0 . 05 . Microarray data have been submitted and can be accessed by GEO accession number GSE119034 . | Kaposi's sarcoma-associated herpesvirus ( KSHV ) infection caused Kaposi’s sarcoma ( KS ) , a highly disseminated tumor that frequently occurs in patients with AIDS . KSHV-encoded viral interferon regulatory factor 1 ( vIRF1 ) is an oncogenic protein , which has been shown to be vital in KSHV evasion of innate antiviral response and induction of tumorigenesis but its role in KS tumor invasiveness and motility remains unclear . A growing volume of literatures has proposed that lncRNAs could function as tumor suppressors or oncogenes , and numerous lncRNAs might act as competing endogenous RNAs ( ceRNAs ) that competitively bind to microRNAs ( miRNAs ) , hence exerting influence on posttranscriptional regulation . However , whether cellular lncRNAs are involved in the progression of KS is still unknown . Here , we revealed a previously undefined role of vIRF1 in cell motility and proliferation , and described the cross-regulatory network of cellular lncRNAs and miRNAs involved in the pathogenesis of KS . We found that the crosstalk between miR-218-5p and lnc-OIP5-AS1 contributed to vIRF1-induced cell motility and proliferation via increasing HMGB2 and CMPK1 expression . In summary , this study constitutes an important discovery related to KS pathogenesis , particularly in the invasiveness and motility of KS tumors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"cell",
"motility",
"luciferase",
"gene",
"regulation",
"natural",
"antisense",
"transcripts",
"enzymes",
"enzymology",
"long",
"non-coding",
"rnas",
"developmental",
"biology",
"micrornas",
"epigenetics",
"dna",
"molecular",
"biology",
"techniques",
"dna",
"methylation",
"chromatin",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"artificial",
"gene",
"amplification",
"and",
"extension",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"oxidoreductases",
"chromatin",
"modification",
"dna",
"modification",
"molecular",
"biology",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"cell",
"migration",
"genetics",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna",
"polymerase",
"chain",
"reaction"
]
| 2019 | Oncogenic KSHV-encoded interferon regulatory factor upregulates HMGB2 and CMPK1 expression to promote cell invasion by disrupting a complex lncRNA-OIP5-AS1/miR-218-5p network |
Most intracellular pathogens that reside in a vacuole prevent transit of their compartment to lysosomal organelles . Effector mechanisms induced by the pro-inflammatory cytokine Interferon-gamma ( IFNγ ) can promote the delivery of pathogen-occupied vacuoles to lysosomes for proteolytic degradation and are therefore important for host defense against intracellular pathogens . The bacterial pathogen Coxiella burnetii is unique in that , transport to the lysosome is essential for replication . The bacterium modulates membrane traffic to create a specialized autophagolysosomal compartment called the Coxiella-containing vacuole ( CCV ) . Importantly , IFNγ signaling inhibits intracellular replication of C . burnetii , raising the question of which IFNγ-activated mechanisms restrict replication of a lysosome-adapted pathogen . To address this question , siRNA was used to silence a panel of IFNγ-induced genes in HeLa cells to identify genes required for restriction of C . burnetii intracellular replication . This screen demonstrated that Indoleamine 2 , 3-dioxygenase 1 ( IDO1 ) contributes to IFNγ-mediated restriction of C . burnetii . IDO1 is an enzyme that catabolizes cellular tryptophan to kynurenine metabolites thereby reducing tryptophan availability in cells . Cells deficient in IDO1 function were more permissive for C . burnetii replication when treated with IFNγ , and supplementing IFNγ-treated cells with tryptophan enhanced intracellular replication . Additionally , ectopic expression of IDO1 in host cells was sufficient to restrict replication of C . burnetii in the absence of IFNγ signaling . Using differentiated THP1 macrophage-like cells it was determined that IFNγ-activation resulted in IDO1 production , and that supplementation of IFNγ-activated THP1 cells with tryptophan enhanced C . burnetii replication . Thus , this study identifies IDO1 production as a key cell-autonomous defense mechanism that limits infection by C . burnetii , which suggests that peptides derived from hydrolysis of proteins in the CCV do not provide an adequate supply of tryptophan for bacterial replication .
Coxiella burnetii is a gram-negative , obligate intracellular pathogen that causes an infectious disease called Q-fever . Humans are occasionally infected through inhalation of aerosols or through close contact with infected livestock , and the symptoms range from mild flu-like illness to vascular complications and fatal endocarditis ( reviewed in [1] ) . Infection of human cells begins with the phagocytosis of C . burnetii . Phagosomes containing C . burnetii undergo endocytic maturation and fuse with lysosomes , which results in the formation of the Coxiella-containing vacuole ( CCV ) . Acidification of the CCV activates the C . burnetii Type IVB secretion system ( T4SS ) called Dot/Icm , which promotes the translocation of roughly 100 different bacterial effector proteins into the host cell cytosol [2 , 3] . Type IV secretion is essential for intracellular replication of C . burnetii and the generation of a spacious CCV that has autophagolysosomal characteristics [3–5] . Individual Type IV effector proteins ( T4E ) facilitate evasion of innate immune surveillance and acquisition of nutrients and membrane for the CCV ( reviewed in [6] ) . The development of an axenic culture medium and genetic manipulation techniques have made C . burnetii an excellent system to study how pathogens adapt to survive and replicate in a lysosome-derived organelle as well as the cell-autonomous immune strategies in place to control their intracellular replication [7 , 8] . Adaptive immune responses lead to the production of IFNγ , which is a critical determinant of host protection against C . burnetii in immunocompetent animals [9 , 10] . IFNγ is a potent pro-inflammatory cytokine secreted by activated lymphocytes during infection . Circulating IFNγ has been reported to be a sensitive and diagnostic biomarker in Q fever patients , which shows that an adaptive cell-mediated immune response has been generated [11 , 12] . IFNγ receptors , ubiquitously expressed on various cell types , bind to IFNγ and stimulate the Janus kinase-Signal transducer and activator of transcription ( JAK-STAT ) signaling cascade that activates expression of hundreds of antimicrobial genes that provide cell-autonomous defense against intracellular pathogens . The functions of IFNγ-induced genes include , but are not restricted to , generation of reactive oxygen and nitrogen radicals , antimicrobial peptides , toxic metabolites , activation of immune signaling , immunoproteasome , antigen presentation , vesicle traffic , autophagy , immune GTPases , small molecule transporters and production of soluble messengers such as cytokines and chemokines ( reviewed in [13] ) . IFNγ-mediated elimination of intravacuolar pathogens ( e . g . Salmonella , Mycobacteria ) involves immune GTPase and autophagic-recognition of the pathogen-containing compartment ( PCV ) and labeling it for lysosomal fusion and degradation [14–16] . In the case of pathogens which rupture their phagosomal vacuole and escape to the cytosol ( e . g . Listeria ) , autophagic response triggers the delivery of the bacteria to the lysosome [17 , 18] . IFNγ-mediated restriction of C . burnetii replication in professional phagocytic cells has been attributed to phenotypes that include CCV alkalinization , TNF-mediated apoptosis , and generation of reactive nitrogen and oxygen species [19–23] . However , restriction mechanisms against pathogens that have evolved to survive and replicate in hostile lysosomal compartments have not been extensively characterized . In an effort to identify and characterize specific host proteins that are induced in IFNγ-activated cells and participate in the restriction of C . burnetii intracellular replication , an siRNA screen using a curated set of IFNγ-induced genes was conducted . Data from the screen shows that Indoleamine 2 , 3-dioxygenase 1 ( IDO1 ) is an IFNγ-induced effector that contributes to the restriction of C . burnetii intracellular replication . IDO1 is an enzyme that catalyzes the conversion of the essential amino acid L-tryptophan to kynurenines , which are then used for the synthesis of the metabolite nicotinamide adenine dinucleotide ( NAD+ ) ( reviewed in [24] ) . Because C . burnetii is a tryptophan auxotroph [25] , these data show that one mechanism by which IFNγ restricts the replication of this intracellular pathogen is through IDO1-mediated depletion of an essential nutrient .
Macrophages treated with IFNγ will restrict C . burnetii replication by a process that is mediated in part by production of inducible nitric oxide synthase and NADPH oxidase [21–23] . Data demonstrating that macrophages deficient in these enzymes still robustly restrict C . burnetii replication indicates that there must be multiple mechanisms by which mammalian cells restrict intracellular replication of C . burnetii upon stimulation by IFNγ [23] . To identify additional pathways by which IFNγ stimulation restricts intracellular replication of C . burnetii , we examined whether treatment of HeLa 229 cells would restrict intracellular replication of this pathogen . The rationale for using HeLa cells was that these cells have many evolutionarily conserved antimicrobial mechanisms that can restrict the replication of intracellular bacterial pathogens ( reviewed in [26] ) , but HeLa cells may not have as many independent pathways to limit the intracellular replication of C . burnetii upon IFNγ treatment as macrophages given that these are human derived epithelial cells and not professional phagocytes . This would increase the likelihood of identifying host factors important for growth restriction by reducing the possibility of redundancy . In addition , HeLa cells are amenable to genetic manipulation . A layout of the experimental setup is presented in Fig 1A . Luminescence generated by a C . burnetii strain expressing the luxCDABE operon constitutively , served as an indicator of bacterial replication ( Fig 1B ) . HeLa cells were infected with C . burnetii and treated with increasing concentrations of IFNγ at 6h post infection ( pi ) . At concentrations as low as 10 ng/ml , there was a significant decrease in the luminescence values observed as early as day 3 ( d3 ) pi . Based on these data , d4 pi was used as the standard time-point for subsequent experiments measuring bacterial luminescence ( Fig 1B ) . In agreement with luminescence readouts , C . burnetii genome equivalents ( GE ) measured by quantitative PCR also showed a significant decline in bacterial numbers from IFNγ-treated cells ( Fig 1C ) . Of note , GE and luminescence data are graphed on log and linear scales respectively , yet reveal the same trend . Reduced bacterial numbers in IFNγ-treated cells also correlated with smaller CCV sizes on d4 pi ( Fig 1D ) . PFA-fixed cells stained for LAMP1 and C . burnetii were visualized by indirect immunofluorescence microscopy . The CCVs in IFNγ-treated cells were small and the bacteria were packed tightly , whereas , untreated cells had larger and more spacious CCVs that contained dispersed bacteria ( Fig 1D ) . Quantification showed that the average CCV size in IFNγ-treated cells was reduced significantly compared to that of untreated cells ( Fig 1D ) . Effector proteins translocated by the Dot/Icm secretion system are required to subvert host vesicle traffic to promote the expansion of the CCV . The observation that the CCVs were smaller in IFNγ-treated cells raised the question of whether IFNγ signaling interferes with the ability of C . burnetii to translocate effector proteins . Examination of Dot/Icm-dependent translocation of the effector proteins was measured using the effector Cbu0077 fused to the translocation reporter BlaM encoding a β-lactamase enzyme that will cleave the fluorescent substrate loaded into host cells . These data demonstrated a significant reduction in BlaM-Cbu0077 translocation in IFNγ-treated cells on d2 pi and on d4 pi ( Fig 1E ) . Because unified CCVs were observed in IFNγ-treated cells , which is a phenotype requiring the effector protein Cig2 [27] , it is likely that C . burnetii are initially capable of translocating early effector proteins in cells that were treated with IFNγ 6h after infection . However , the decrease in translocation of BlaM-Cbu0077 observed at d2 and d4 pi indicates that IFNγ-treatment leads to inhibition of Dot/Icm function at these later times , which is consistent with the decrease observed in bacterial luminescence and replication ( Fig 1E ) . To determine if C . burnetii isolated from IFNγ-treated cells were viable and capable of initiating a secondary infection , a foci-forming unit ( FFU ) assay was performed . Lysates and supernatants of C . burnetii-infected cells that were either untreated or treated with IFNγ were collected on d4 pi . Dilutions of each sample were used to measure the number of bacteria by quantitative PCR ( GE ) and infect untreated HeLa cells . Infected cells were fixed and stained using an anti-C . burnetii antibody on d4 pi and immunofluorescence microscopy was used to determine the FFU value after counting the number of cells containing large CCVs ( Fig 1F , left panel ) . C . burnetii collected from untreated cells ( primary infection ) gave rise to a FFU value that was almost a log higher than that isolated from IFNγ-treated cells infected in parallel . This difference was comparatively larger than that obtained using the GE assay ( Fig 1F , right panel ) . This indicates that IFNγ treatment reduces the viability of C . burnetii that were still detectable using the GE assay . Together , these data demonstrate that IFNγ limits the expansion of the CCV , intracellular replication , metabolic activity and viability of C . burnetii by inducing cell-autonomous defense mechanisms . Thus , in addition to restricting C . burnetii in professional phagocytes [19 , 20 , 22 , 23] , these results indicate that IFNγ induces antimicrobial activities that are capable of restricting C . burnetii replication in cells that are normally non-phagocytic , which suggests there could be a shared IFNγ-activated antimicrobial activity in these cells . Depending on the cell type , IFNγ signaling stimulates the expression of hundreds to thousands of genes . Published data sets profiling genes upregulated by IFNγ in HeLa cells and macrophages were used to curate genes that may be involved in restricting C . burnetii replication [28–31] . Genes encoding proteins that regulate transport and fusion of membranes were given high priority , as were regulators of nutrient transport and cell metabolism . To identify the specific host factors that mediate restriction , a subset of genes upregulated following IFNγ treatment was silenced individually using siRNA . Cell-surface receptors for the cytokine IFNγ ( IFNγR1 and IFNγR2 ) , which signal the kinases of the JAK-STAT signal transduction pathway ( JAK1 , JAK2 ) to activate the transcription factor STAT1 were included as positive controls ( Fig 2 ) . Because intracellular bacterial luminescence provided a robust and sensitive assay for C . burnetii replication ( Fig 1B ) , the C . burnetii lux strain was used to monitor the effect of individual gene knockdowns on bacterial replication . Silencing of genes encoding components of the JAK-STAT signal transduction pathway , IFNγR1 , IFNγR2 , JAK1 , JAK2 and STAT1 , significantly increased C . burnetii replication in IFNγ-treated cells , which indicated that this screen was sensitive enough to identify potential restriction factor candidates ( Fig 2 ) . Among the IFNγ-induced effector genes , silencing of the gene encoding Indoleamine 2 , 3-dioxygenase 1 ( IDO1 ) resulted in the largest increase in C . burnetii replication in IFNγ-treated cells ( Fig 2 ) . These data implicate IDO1 as being a critical effector that mediates restriction of C . burnetii intracellular replication in IFNγ-treated cells . Experiments to validate that IDO1 expression inhibits C . burnetii replication were conducted in HeLa cells ( Fig 3A ) . Immunoblot analysis and quantitative RT-PCR demonstrated robust induction of IDO1 in IFNγ-treated cells , and silencing of IDO1 expression by siRNA treatment ( Fig 3B ) . As suggested in the initial screen , IDO1 silencing significantly increased C . burnetii luminescence in IFNγ-treated cells ( Fig 3C ) . In STAT1-silenced cells it was found that IFNγ treatment resulted in higher levels of C . burnetii luminescence , which indicates that there are additional mechanisms by which IFNγ stimulation of cells restricts intracellular replication of C . burnetii in the absence of IDO1 ( Fig 3C ) . IDO1 is a cytosolic enzyme that catabolizes tryptophan , which will reduce cellular levels of this amino acid . This enzyme can be inhibited by 1-Methyl Trp , which is a competitive analogue of tryptophan ( Fig 3D ) . The addition of 1-Methyl Trp enhanced C . burnetii luminescence in IFNγ-treated cells to a magnitude similar to that observed when IDO1 was silenced ( Fig 3E ) . Thus , the enzymatic activity of IDO1 is important for restriction of C . burnetii in host cells stimulated with IFNγ . Because C . burnetii is a tryptophan auxotroph it must acquire tryptophan from the host cell [25] . IDO1 expression could inhibit C . burnetii replication by depleting tryptophan in the cytosol or by generating kynurenines that have antimicrobial properties . To determine if tryptophan depletion was the primary mechanism by which IDO1 inhibits C . burnetii replication , exogenous tryptophan was added to the tissue culture medium to test whether this was sufficient to suppress IDO1-mediated growth restriction . A schematic showing the timing of tryptophan addition and collection of samples is shown in Fig 4A . Tryptophan supplementation increased C . burnetii luminescence in IFNγ-treated cells to a magnitude similar to that observed following IDO1 knockdown ( Fig 4B ) . Importantly , tryptophan supplementation did not enhance C . burnetii luminescence in IDO1- or STAT1-silenced cells . This indicates that growth restriction mediated by tryptophan depletion is dependent on STAT1 and IDO1 . A decrease in C . burnetii luminescence was observed in the IFNγ-treated cells 3-days post-infection , consistent with a cessation of luciferase production by intracellular bacteria ( Fig 4C ) . By contrast , C . burnetii luminescence values continued to increase over the 5-days of infection when the IFNγ-treated cells were supplemented with tryptophan , which indicates that the intracellular bacteria remained more metabolically active ( Fig 4C ) . Similarly , tryptophan supplementation augmented the function of the type IV secretion system in IFNγ-treated cells as evident from the significant increase in the translocation of the effector protein BlaM-Cbu0077 ( Fig 4D ) . Measurements of CCV area showed that IFNγ treatment resulted in smaller vacuoles containing C . burnetii , and this phenotype was suppressed by the addition of tryptophan ( Fig 4E ) . To gain additional insight into the mechanism of IDO1-mediated growth restriction , a GE assay was used to measure C . burnetii genome expansion . Similar to the luminescence data , a decrease in C . burnetii GE values was observed in the IFNγ-treated cells compared to untreated cells at 4-days post-infection and tryptophan supplementation of cells treated with IFNγ resulted in a significant increase in C . burnetii GE values by d7 pi ( Fig 4F ) . Lastly , the FFU assay showed that the addition of tryptophan increased C . burnetii viability in the IFNγ-treated cells ( Fig 4G ) . Thus , IDO1 production interferes with C . burnetii intracellular replication and survival by depleting free tryptophan in the host cytosol and supplementing the culture medium with excess tryptophan suppresses IDO1-mediated restriction . Transmission electron microscopy ( TEM ) was used to examine whether IFNγ stimulation had any detectible impact on the morphology of the CCV . Several representative images show that CCVs in untreated cells were spacious , which means the individual C . burnetii were dispersed randomly throughout the lumen of the vacuole and were not typically in close contact with each other . Also , numerous vesicles were observed within the lumen of CCVs , which is likely due to robust fusion of autophagosomes with the CCV and formation of internal vesicles by a functional multivesicular body ( MVB ) pathway ( Fig 5 ) . By contrast , CCVs in IFNγ-treated cells were constricted , showed enhanced osmium tetroxide staining , had fewer intraluminal vesicles , and the bacteria inside the vacuole were tightly packed and many showed signs of damage resulting in cellular swelling ( Fig 5 ) . Importantly , the membrane of the CCV appeared to be intact and C . burnetii were not detected in the cytosol . The CCVs in the IFNγ-treated cells supplemented with tryptophan were similar in appearance to the CCVs in the untreated cells . Thus , tryptophan depletion resulting from IDO1 production interferes with the ability of C . burnetii to maintain a spacious CCV that supports intracellular replication . Although IDO1 was important for restriction of C . burnetii replication in cells stimulated with IFNγ , it was unclear whether IDO1 expression would restrict C . burnetii replication in the absence of other IFNγ-induced genes . To address this question , a HeLa cell line that produces IDO1 under the control of a tetracycline-inducible promoter was created . Immunoblot analysis showed that the IDO1 protein was not produced by unstimulated cells but when cells were stimulated with a tetracycline inducer ( doxycycline or anhydrotetracycline ) or treated with IFNγ , there was a concentration-dependent increase in IDO1 protein levels ( Fig 6A ) . The IDO1-inducible cell line was infected with C . burnetii lux and IDO1 expression was stimulated by either anhydrotetracycline ( referred to as Tet ) induction or by stimulation with IFNγ as shown in the schematic ( Fig 6B ) . Anhydrotetracycline induction of IDO1 significantly inhibited C . burnetii replication in the HeLa pTRIPZ-IDO1 cells , but not in the vector control HeLa pTRIPZ-EV cells ( Fig 6C , left panel ) . The addition of tryptophan to the culture medium restored C . burnetii replication in HeLa pTRIPZ-IDO1 cells that had been induced to produce IDO1 ( Fig 6C , right panel ) . The GE assay confirmed data from the luminescence assay and showed that induction of IDO1 production in the HeLa pTRIPZ-IDO1 cells was sufficient to restrict C . burnetii replication by a mechanism that could be suppressed by the addition of tryptophan to the culture medium ( Fig 6D ) . These data indicate that host cell expression of IDO1 will restrict C . burnetii replication in the absence of IFNγ signaling , which validates that depletion of tryptophan in the host cell cytosol is sufficient to disrupt the ability of C . burnetii to replicate in the lysosome-derived CCV . The THP1 cell line was used to determine whether IDO1-mediated restriction of C . burnetii replication is a conserved IFNγ-activated pathway that is operational in human-derived macrophage-like cells . Immunoblot analysis was used to evaluate IDO1 production in undifferentiated THP1 cells and THP1 cells that were differentiated into macrophage-like cells by treatment with PMA . Similar to control HeLa cells , differentiated THP1 cells produced IDO1 protein upon IFNγ stimulation ( Fig 7A ) . Differentiated THP1 cells treated with IFNγ restricted C . burnetii replication as determined by measuring bacterial luminescence ( Fig 7B ) . The culture medium was supplemented with increasing amounts of tryptophan to determine if IDO1-mediated tryptophan depletion has a measurable effect on restriction of C . burnetii replication in differentiated THP1 cells treated with IFNγ . A dose-dependent increase in C . burnetii luminescence was observed upon the addition of tryptophan ( Fig 7B ) . A tryptophan concentration of 0 . 3125 mM resulted in significantly higher levels of C . burnetii luminescence ( Fig 7C ) . Thus , IDO1 is induced and participates in the restriction of C . burnetii replication in macrophages .
Activation of cell-intrinsic defense by the cytokine IFNγ enables mammals to combat a large number of microbial pathogens that are able to survive and replicate in host cells . In this study , C . burnetii was used as a model pathogen to advance our understanding of how IFNγ-induced responses enable host cells to defend themselves against pathogens that have evolved the ability to replicate in lysosomes , which are catabolic and hostile organelles for most microbes . The data revealed that IFNγ , even at relatively low doses , stimulated a response that efficiently restricted C . burnetii replication in HeLa cells , which indicates that non-phagocytic cells are also equipped with cell-intrinsic antimicrobial mechanisms that can limit replication of this pathogen . The IFNγ-induced enzyme IDO1 was found to be important for restriction of C . burnetii replication . IDO1 limits the amount of tryptophan available for C . burnetii , which is a tryptophan auxotroph so must acquire this essential nutrient from the host . Data presented here indicate that IDO1-mediated depletion of tryptophan stalls C . burnetii infection by inhibiting bacterial metabolism and secretion of effector proteins by the Dot/Icm system . This would explain the impact of IDO1 production on the size , morphology and maintenance of the CCV , which is dependent on Dot/Icm function . Importantly , IDO1 is an enzyme that remains localized in the host cell cytosol , whereas , C . burnetii replicates inside a membrane-bound vacuole . Given that the lysosome-derived CCV is an acidified organelle and retains the ability to hydrolyze proteins , these data suggest that the pool of tryptophan generated by lysosomal degradation of proteins inside the CCV is not sufficient to maintain the nutritional requirements for C . burnetii metabolism . Thus , this organism must have the ability to access metabolites such as tryptophan that are in the cytosol . In addition to tryptophan , C . burnetii depends on the host for several other key amino acids , which include arginine , cysteine , histidine , leucine , lysine , phenylalanine , proline , tyrosine , threonine , and valine [25] . An important question for future studies will be to determine which of these essential amino acids can be generated in sufficient quantities by hydrolysis of proteins in the vacuole lumen , and which amino acids must be imported into the vacuole either through the subversion of host transporters or through bacterial transporters that are delivered into the CCV membrane . Phylogenetic analysis of the C . burnetii genome indicates that genome reduction , pseudogenization of genes occurred as Coxiella evolved and adapted from tick-associated lifestyle to infect mammalian hosts [32–36] . C . burnetii tryptophan synthesis genes , in particular , are related to the genes in Simkania negevensis , which is in the phylum Chlamydiae and have been suggested to be acquired from Chlamydial ancestors through horizontal transfer [36 , 37] . A closer analysis of the C . burnetii trp genes provides insight into why this pathogen is unable to synthesize tryptophan [32 , 38] . Multiple frameshift mutations in the trpDG genes ( CBU1153 ) renders the biosynthetic pathway incapable of utilizing chorismate as a precursor for tryptophan synthesis [32] . There is also a fusion of the genes encoding phosphoribosyl anthranilate isomerase ( trpF ) and tryptophan synthase ( trpB ) [32] . Whether this fusion protein ( CBU1155 ) retains enzymatic function is unknown . Regardless , data here indicate that IDO1-mediated depletion of tryptophan restricts C . burnetii replication , which means that C . burnetii is incapable of utilizing secondary metabolites in the tryptophan synthesis pathway or that these substrates are not sufficiently available . Chlamydia , Leishmania and Toxoplasma spp . are also tryptophan auxotrophs and susceptible to IDO1-mediated growth restriction in human cells [39–42] . Similar to what was shown here for C . burnetii , the inclusion containing Chlamydia trachomatis becomes contrasted and Chlamydia exhibit enlarged , atypical morphology during IDO1-mediated tryptophan depletion [43] . Intracellular replication of Mycobacterium tuberculosis ( Mtb ) and Listeria monocytogenes , which are not tryptophan auxotrophs , can also be inhibited by IDO1 activity . For Mtb , IDO1-mediated tryptophan depletion is detrimental to mutants defective in tryptophan biosynthesis [44] . In the case of Listeria , tryptophan catabolites such as kynurenines ( Kyn ) and 3-hydroxy kynurenines ( 3HK ) have been shown have antibacterial effects [45] . It is unlikely that kynurenines have a negative impact on C . burnetii replication because supplementing culture medium with excess tryptophan completely negated the contribution of IDO1 to restricting C . burnetii replication . Additionally , presence of excess tryptophan in the axenic medium used to culture C . burnetii has been shown to maximize log-phase growth and bacterial plating efficiency [8 , 46] . Thus , tryptophan is essential for C . burnetii multiplication in both cell-free and intracellular environments . Several IFNγ-induced effectors did not appear to be necessary for restriction of C . burnetii replication based on this screen . For instance , silencing of the chloride , glucose and organic compound transporter genes ( CLIC2 , SLC2A3 , SLC16A1 and SLC22A2 ) , or the genes involved in phospholipid and cholesterol transport ( PLSCR1 , APOL1 and APOL2 ) , did not increase C . burnetii replication . Likewise for the genes encoding the gp91-phox subunit of the NADPH oxidase ( CYBB ) or nitric oxide synthase ( NOS2 ) . This may be explained by the finding that intracellular bacteria such as Salmonella do not experience significant redox stress in epithelial cells compared to macrophages [47] . In addition , C . burnetii is believed to encode acid phosphatases , superoxide dismutases and peroxiredoxin that neutralize the toxic effect of oxidative radicals [48–52] . Induction of apoptotic cell death pathways is another mechanism by which IFNγ-treated cells can restrict the replication of intracellular pathogens . Bacterial effector proteins delivered into host cell by the C . burnetii T4SS , however , can block intrinsic and extrinsic apoptotic pathways in infected epithelial cells [53 , 54] , which may be why silencing of the TNF receptor ( TNFRSF1A ) did not have an effect on C . burnetii growth restriction . Lastly , it is well appreciated that interferon-inducible guanylate binding proteins and the immunity-related GTPase M ( GBP1 , GBP2 , GBP5 and IRGM ) target pathogen-containing compartments and cytosolic bacteria to promote their destruction by cell-autonomous mechanisms ( reviewed in [55] ) . These GTPases were not essential for inhibiting C . burnetii replication . Because C . burnetii-containing phagosomes are transported along the default endocytic pathway to the lysosomes , these GTPases may not recognize the CCV as a modified phagosome that has evaded maturation and should be targeted for lysosomal fusion and degradation . It remains possible that some of these IFNγ-induced proteins contribute to C . burnetii growth restriction , but because they may have functions that are redundant , there is no enhancement of C . burnetii replication when only one factor is silenced in IFNγ-treated cells . Data presented here demonstrate that ectopic production of IDO1 is sufficient to suppress C . burnetii replication in the absence of IFNγ signaling . IDO1 activity can limit the replication of intracellular pathogens to prevent dissemination , however , is typically not sufficient to kill intracellular microbes and without the participation of other cell autonomous defense pathways , can lead to persistent infection of infected tissues [56 , 57] . It is likely that other IFNγ-dependent mechanisms participate in killing C . burnetii once IDO1 renders the bacteria metabolically incapable of interfering with host cell functions . Consistent with IFNγ-activation leading to nutritional depletion , knockdown of SLC11A1 , the gene encoding the metal transporter NRAMP1 ( Natural resistance-associated macrophage protein 1 ) that restricts iron and manganese levels in phagosomes , led to a small , but significant increase in C . burnetii replication . These data indicate that host modulation of metal concentrations in the CCV could contribute to restriction of intracellular C . burnetii . Thus , future studies will be aimed at identifying additional IFNγ-dependent mechanisms that act synergistically to combat intracellular C . burnetii replication using cells that are deficient in IDO1 .
HeLa 229 cells ( ATCC ) and THP1 ( ATCC ) were used in experiments . HeLa cells were cultured and maintained in Dulbecco’s Modified Eagle Medium ( DMEM , Gibco Cat . 11965–118 ) supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) . THP1 cells were cultured in Roswell Park Memorial Institute 1640 Medium with ATCC modification ( ThermoFisher Scientific , Cat . A1049101 ) supplemented with 10% heat-inactivated FBS . All cells were maintained at 37°C with 5% CO2 . During the length of infection experiments , HeLa cells were maintained with 5% FBS whereas THP1 cells were maintained with 10% FBS . Wild-type Coxiella burnetii Nine Mile phase II ( RSA439 ) was cultured in ACCM-2 media and used for all experiments [58 , 59] . A NMII strain that constitutively expresses luminescence was generated and provided by Shawna C . Reed ( Table 1 ) . C . burnetii strains were grown for 6 days in ACCM-2 at 37°C , 2 . 5% O2 and 5% CO2 with appropriate antibiotic selection ( 375μg/ml kanamycin for Kanr strains , 3μg/ml chloramphenicol for Cmr strains ) as described [58 , 59] . Bacterial cultures were centrifuged at 4000 rpm , 4°C for 15 mins and pellets re-suspended in half the volume with DMEM containing 5% FBS . Bacteria were sonicated for 10’ prior to infection . C . burnetii genome equivalents were measured by quantitative PCR as described previously [27] . Black 96-well plates with clear bottom were purchased from Corning costar ( Cat . 3904 ) . Dharmafect ( T-2001 ) , 5X siRNA buffer ( Cat . B-002000-UB-100 ) and siGENOME SMARTpool siRNAs as listed in Table 2 , were purchased from Dharmacon . Recombinant human IFNγ was obtained from Biolegend ( Cat . 570206 ) . L-tryptophan ( Trp ) and 1-Methyl L-tryptophan ( 1-Methyl Trp ) were obtained from Sigma-Aldrich ( Cat . T0254 , 447439 ) . Trp was re-suspended in Tissue Culture grade water and used at 0 . 3125mM unless otherwise indicated . 1-Methyl Trp was resuspended in 0 . 1N NaOH and used at 0 . 2mM . Phenol-red free DMEM and probenecid were obtained from Thermo Fisher Scientific ( Cat . 21063029 and P36400 ) . LIVEBLAzer™-FRET B/G loading kit which includes the fluorescent substrate CCF4-AM was purchased from Invitrogen ( Cat . K1095 ) . siRNAs were re-suspended in 1X siRNA buffer as 10μM or 2μM stocks and stored at -20°C until use . 48h before infection , HeLa cells ( 104 cells/well ) were reverse-transfected with 25 or 50nM siRNA using Dharmafect ( 0 . 2μl/well ) in 96-well black , clear bottom plates . As controls , cells were transfected with transfection reagent alone ( Mock ) or control non-targeting siRNA . HeLa cells were infected with C . burnetii lux at MOI 100 and 6h later , IFNγ was added . Cells were washed and replenished with fresh media without IFNγ on d1 pi and luminescence values measured on specific days post-infection as indicated . Peak C . burnetii luminescence values measured from untreated cells ( d4 pi or as indicated ) were normalized to 100% . In the experiments performed with Trp and 1-Methyl Trp , C . burnetii- infected cells were supplemented with Trp ( 0 . 3125mM ) or treated with 1-Methyl Trp ( 0 . 2mM ) 1h prior to treatment with IFNγ . Trp or 1-Methyl Trp , but not IFNγ , were added back to the media after washing the cells on d1 pi . Bacterial luminescence was measured using TECAN infinite M1000 plate reader on indicated days post-infection . THP1 cells were plated at 75000/well in 96-well plates and differentiated with 124ng/ml phorbol 12-myristate 13-acetate ( PMA ) overnight or left undifferentiated . Cells were infected with MOI 25 and 6h later , treated with IFNγ 100ng/ml in the presence or absence of additional tryptophan ( concentrations as indicated ) . On d1 pi , cells were washed and replenished with IFNγ-free media and supplemented with tryptophan . Bacterial luminescence was measured using TECAN infinite M1000 plate reader on indicated days post-infection . Effector translocation through the type IV secretion system was assessed by measuring the translocation of β-lactamase ( BlaM ) - effector fusion protein using a FRET-based assay , as previously described [2] . HeLa cells were plated at 2*104 cells/w in 96w black , clear bottom plates 24h prior to infection . Cells were infected with wt or icmL::Tn C . burnetii expressing BlaM alone or BlaM-77 at MOI 500 as listed in Table 3 . Cells were treated with IFNγ 6h later , in the presence or absence of additional tryptophan . On d1 pi , cells were washed and replenished with IFNγ-free media and supplemented with tryptophan where indicated . On d2 or d4 pi , culture medium was replaced with phenol red-free DMEM containing HEPES . Cells were loaded with the fluorescent substrate CCF4/AM using the LIVEBLAzer™-FRET B/G loading kit and probenecid and incubated in dark at RT for 2h . The ratio of signal at 460 and 535nm ( blue:green ) was measured using the TECAN M1000 plate reader and the response ratio was calculated by normalizing the blue: green ratio of infected cells to that of uninfected control . 5*104 cells per well were plated in 24-well plates , one day prior to infection . Cells were infected wt C . burnetii NMII at MOI 100 . Cells were left untreated or treated with IFNγ in the presence or absence of Trp at 6h pi . Infected cells were washed on d1 pi and replenished with IFNγ-free fresh media in the presence or absence of additional Trp . Supernatants and cells lysed with distilled water were combined and collected on d1 , d4 and d7 pi . Genomic DNA was extracted using Illustra bacterial genomicPrep mini spin kit ( Cat no . 28904259 , GE ) and quantified by qPCR using primers for C . burnetii dotA gene ( Table 4 ) . 2 . 5 or 5*104 cells per well were plated in 24-well plates with poly L-lysine coated coverslips , one day prior to infection . Cells were infected at MOI 100 , left untreated or treated with IFNγ and additional Trp as indicated . On d1 pi , cells were washed and replenished with IFNγ-free media in the presence or absence of additional Trp . Cells were fixed on d4 pi with 4% para formaldehyde ( PFA ) for 20 mins at RT . Coverslips were washed at least 6x times with 1x PBS , permeabilized and blocked with 0 . 2% saponin , 0 . 5% BSA and 1% ( v/v ) heat-inactivated FBS in PBS . Coverslips were stained with primary antibodies- rabbit anti-C . burnetii [3] , mouse anti-LAMP1 ( Source: H4A3 Development Studies Hybridoma bank at the University of Iowa ) and DAPI ( 4 , 6-diamidino-2-phenylindole , Cat no . D9542 from Sigma-Aldrich ) at 1:10 , 000 , 1:500 and 1:10 , 000 dilutions respectively , as previously described [63] . Secondary antibodies goat anti-Rabbit IgG , Alexa Fluor 568 ( Life Technologies , A11036 ) and goat anti-Mouse IgG , Alexa Fluor 488 ( Life Technologies , A11029 ) were used at 1:2000 . Stained coverslips were mounted on glass slides using Prolong gold antifade reagent ( Life Technologies ) . Coverslips were visualized with a Photometrics CoolSNAP EZ 20 Mhz digital monochrome camera connected to Nikon Eclipse TE2000-S inverted microscope using Nikon Plan Apo60x objective/ 1 . 4 numerical aperture . Images were acquired using SlideBook software 6 . 0 ( Intelligent Imaging Innovations ) , saved as tiff files and resized and labeled using Adobe Illustrator . To measure the infectivity of C . burnetii in infected cells , Foci Forming Unit ( FFU ) assay was used . HeLa 229 cells were plated in 6-well plates at 200 , 000 cells/w one day prior to infection . Cells were infected with wt C . burnetii at MOI 100 , left untreated or treated with IFNγ and additional Trp as indicated . On d1 pi , cells were washed and replenished with IFNγ-free media in the presence or absence of additional Trp . Similar to GE assay , supernatants and lysates from infected cells were combined and collected on d4 pi and sonicated for 10’ . This culture was serially diluted either in water ( to assess primary infection GE by qPCR ) or in DMEM containing 5% FBS to infect HeLa cells ( plated at 2*104 cells/well in 96-well plate on the previous day ) for a second round of infection . Infected cells in 96-well plates were subsequently washed on d4 pi , fixed with 4% PFA for 20’ at RT , permeabilized with 0 . 5% Triton-X in PBS for 5’ and stained with rabbit anti-C . burnetii at 1:10000 in 0 . 1% Triton-X in PBS . Cells were washed with PBS , stained with DAPI and goat anti-rabbit IgG , Alexa Fluor 488 secondary antibody ( Life Technologies , A11034 ) . Foci were visualized by indirect immunofluorescence , manually counted and represented as FFU per ml of the primary infection sample by accounting for the dilution factor . IDO1 expression in HeLa cells was measured by qRT-PCR using primers listed in Table 4 . Fold increase in IDO1 expression in IFNγ-treated , Mock or IDO1 siRNA-transfected cells was calculated in comparison to untreated , mock transfected cells by assuming 100% efficiency for primers . HeLa 229 cells were infected with wt C . burnetii at MOI 100 and left untreated or treated with IFNγ in the presence or absence of supplemented Trp at 6h pi . Cells were washed on d1 pi and Trp was supplemented in the appropriate samples . The following method was adapted based on published EM studies with C . burnetii [64] and recommendation from the Yale EM facility . On d4 pi , infected cells in petri dishes were fixed in 2 . 5% glutaraldehyde in 0 . 1M sodium cacodylate buffer pH7 . 4 containing 2% sucrose for 1h , then rinsed in buffer and replaced with 0 . 1% tannic acid in buffer for another hour . Buffer-rinsed cells were scraped in 1% gelatin and spun down in 2% agar . Chilled blocks were trimmed and post fixed in 1% osmium tetroxide and 1 . 5% potassium ferrocyanide in buffer for 1 hour . The samples were rinsed in sodium cacodylate and distilled water and en-bloc stained in aqueous 2% uranyl acetate for 1h . This was followed by further rinsing in distilled water and dehydrated through an ethanol series to 100% . The cells were infiltrated with Embed 812 ( Electron Microscopy Sciences ) resin , placed in silicone molds and baked at 60°C for 24h . Hardened blocks were sectioned using a Leica UltraCut UC7 . 60nm sections were collected on formvar coated nickel grids and stained using 2% uranyl acetate and lead citrate . 60nm Grids were viewed FEI Tencai Biotwin TEM at 80Kv . Images were taken using Morada CCD and iTEM ( Olympus ) software . Images were resized and scale bars added using Adobe Illustrator . IDO1 gene was cloned into an empty lentiviral vector pTRIPZ-EV under the tetracycline-inducible promoter at EcorI site by ligase independent cloning using the primers listed in Table 4 . To derive the lentivirus , HEK293T cells were plated in 10cm dish and transfected with pTRIPZ-EV or pTRIPZ-IDO1 with pVSV-G and psPAX2 , as listed in Table 3 , using Lipofectamine 2000 ( Invitrogen ) . Lentiviral particles were obtained by collecting the supernatant at 48h and 72h post-transfection , pooled , filtered using 0 . 45μm low protein binding filter . Lentivirus containing pTRIPZ-EV or pTRIPZ-IDO1 was used to transduce sub-confluent HeLa cells at half the volume of complete media , in the presence of polybrene ( 8μg/ml ) . Transduced cells were maintained in culture using puromycin ( 2 . 25μg/ml ) . To test the efficiency of IDO1 knockdown by siRNA , HeLa cells were plated and treated with IDO1 siRNA and IFNγ 10ng/ml as described for the siRNA screen . 24h later , cells were washed and replenished with IFNγ-free media . Cells were collected two days post IFNγ treatment for immunoblotting . To test the expression of IDO1 in pTRIPZ-IDO1 expressing HeLa cells , 105 cells/w were plated in 24w plates and treated with increasing concentrations of anhydrotetracycline ( referred to as Tet ) or Doxycycline for 3 days or IFNγ 10ng/ml for 4 days . Cells treated with IFNγ were washed and replenished with IFNγ-free media 24h post-treatment . To determine if IDO1 is induced in THP1 cells , 4*105 cells/w were plated in 12w plates and left untreated or differentiated with PMA overnight . Cells were washed and left untreated or treated with IFNγ 100ng/ml . IFNγ-treated cells were washed and replenished with IFNγ-free media after 24h . Cells were harvested 2 days post IFNγ-treatment . In all cases , cell lysates were prepared using Blue Loading Buffer ( Cell signaling , #7722 ) . Anti-IDO1 ( BioLegend , W16073A ) , anti-α-tubulin ( Sigma-Aldrich , T9026 ) prepared at 1:1000 in 5% BSA in 1X TBST and secondary antibodies goat anti-rat or anti-mouse IgG , HRP ( Invitrogen , Cat . 62–9520 and 62–6520 ) were used to detect the proteins by western blotting . Data presented here are representative of at least 2 experiments . The siRNA screen ( Fig 2 ) represents an average of 3 independent trials . Data were graphed and analyzed using Prism 7 software . Statistical significance was determined by t-test , one-way or two-way ANOVA depending on the number of groups and experimental conditions being compared . | Coxiella burnetii is a mammalian pathogen that can cause a predominantly zoonotic disease called Q-fever . In humans , Q-fever manifests as an acute or chronic illness especially in immunocompromised individuals . C . burnetii is uniquely adapted to live in a lysosome-derived vacuole that degrades proteins and provides nutrients that support intracellular replication . From a cell biological perspective , C . burnetii represents an excellent model to study pathogens that survive in harsh cellular environments . The strategies by which infected cells intrinsically combat C . burnetii are not well-established . In this study , we investigate the underlying mechanism by which IFNγ activates cells and prevents C . burnetii from replicating inside cells . The data presented here demonstrate that IFNγ induces the expression of the enzyme Indoleamine 2 , 3-dioxygenase 1 ( IDO1 ) , which degrades the amino acid tryptophan and restricts the intracellular replication of C . burnetii . The production of IDO1 is sufficient to inhibit replication of C . burnetii , indicating that tryptophan depletion is an effective cell-autonomous defense mechanism against this lysosome-adapted pathogen . In addition , these data imply that the degradative vacuole in which this pathogen resides does not generate a supply of tryptophan sufficient to support intracellular replication . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"lysosomes",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"intracellular",
"pathogens",
"hela",
"cells",
"gene",
"regulation",
"pathogens",
"biological",
"cultures",
"microbiology",
"organic",
"compounds",
"electromagnetic",
"radiation",
"cell",
"cultures",
"amino",
"acids",
"cellular",
"structures",
"and",
"organelles",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"coxiella",
"burnetii",
"aromatic",
"amino",
"acids",
"small",
"interfering",
"rnas",
"proteins",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"chemistry",
"luminescence",
"cell",
"lines",
"physics",
"biochemistry",
"rna",
"organic",
"chemistry",
"nucleic",
"acids",
"cell",
"biology",
"tryptophan",
"genetics",
"biology",
"and",
"life",
"sciences",
"interferons",
"physical",
"sciences",
"cultured",
"tumor",
"cells",
"non-coding",
"rna"
]
| 2019 | Host cell depletion of tryptophan by IFNγ-induced Indoleamine 2,3-dioxygenase 1 (IDO1) inhibits lysosomal replication of Coxiella burnetii |
Genome-wide association studies have identified a wealth of genetic variants involved in complex traits and multifactorial diseases . There is now considerable interest in testing variants for association with multiple phenotypes ( pleiotropy ) and for testing multiple variants for association with a single phenotype ( gene-based association tests ) . Such approaches can increase statistical power by combining evidence for association over multiple phenotypes or genetic variants respectively . Canonical Correlation Analysis ( CCA ) measures the correlation between two sets of multidimensional variables , and thus offers the potential to combine these two approaches . To apply CCA , we must restrict the number of attributes relative to the number of samples . Hence we consider modules of genetic variation that can comprise a gene , a pathway or another biologically relevant grouping , and/or a set of phenotypes . In order to do this , we use an attribute selection strategy based on a binary genetic algorithm . Applied to a UK-based prospective cohort study of 4286 women ( the British Women's Heart and Health Study ) , we find improved statistical power in the detection of previously reported genetic associations , and identify a number of novel pleiotropic associations between genetic variants and phenotypes . New discoveries include gene-based association of NSF with triglyceride levels and several genes ( ACSM3 , ERI2 , IL18RAP , IL23RAP and NRG1 ) with left ventricular hypertrophy phenotypes . In multiple-phenotype analyses we find association of NRG1 with left ventricular hypertrophy phenotypes , fibrinogen and urea and pleiotropic relationships of F7 and F10 with Factor VII , Factor IX and cholesterol levels .
Pleiotropy refers to a phenomenon in which a single locus affects two or more apparently unrelated phenotypic traits . It is often identified as a single mutation that affects these two or more wild-type traits [1] . The study of pleiotropic genes usually involves the mapping of phenotypic traits to a single mutant locus . When two or more traits consistently segregate with a particular mutation , this mutation is then classified as pleiotropic . In the case of S . cerevisiae ( yeast ) , it has been argued that the pleiotropic effects of a gene are not usually conferred by multiple molecular functions of the gene , but by multiple consequences ( biological processes ) of a single molecular function [2] . Tyler et al . [3] defined the concept of vertical and horizontal pleiotropy , extending the definition of relational and mosaic pleiotropy proposed by Hadorn and Mittwoch [4] . Vertical or relational pleiotropy appears when a mutation in one gene produces a modification of one particular phenotype , which leads to modification in one or several related phenotypes . By contrast , horizontal or mosaic pleiotropy appears when one mutation in one gene with a causal implication in several biological mechanisms , causes a disruption in these mechanisms . This causes alteration in very different phenotypes , which are observable at the same physiological level . Some papers [5] [6] [7] have established a high level of pleiotropy for certain genes , particularly genes associated with disease [8] . To discover such associations we could use a range of multivariate techniques which highlight the dependence of a single variable on a set of independent variables . Some proposals are based on combining univariate association measures for different phenotypes in order to find pleiotropic effects , such as PRIMe [9] or Yang et al's approach [10] , based in O'Brien's method [11] . An approach taken by Li [5] , uses Fisher's combined p-value approach [12] , adjusting Fisher's combined measure using a Satterwhite approximation method . Other approaches use a Bayesian network approach [13] or multiple regression analysis [14] . However , for the purposes of pleiotropy analysis , we are most interested in finding dependencies between two multivariate sets of variables , rather than a relation of one set with one dependent variable . Various techniques have also been introduced to deal with such multivariate problems [15] . An example of this multiple SNP/multiple phenotype analysis is GUESS [16] , which is an implementation of a Bayesian variable selection algorithm for multiple regression using evolutionary Monte Carlo techniques: the algorithm selects relevant SNPs and identifies the contribution of each SNP to single or multiple traits . In this paper , we will focus on Canonical Correlation Analysis ( CCA ) [17] , which uses linear combinations of variables derived from two sets of data objects and finds those combinations which are maximally correlated with each other . The variables found in the first iteration of the method give the first set of canonical variables . In subsequent iterations we seek variables which maximize the same correlation function , subject to the constraint that they are uncorrelated with previous sets of canonical variables . CCA has been used as an efficient and powerful approach for both univariate and multivariate gene-based association tests . For genomic multivariate data analysis , such an approach would involve finding linear combinations over very large blocks of features , typically involving tens of thousands of features . However , to use CCA , the number of samples should be more than the number of features . To handle this issue , some solutions have been proposed for genomic data integration , such as sparse CCA [18] . With this approach , sparsity is intrinsically achieved by the algorithm so that the number of features used is less than the sample size . This method maximizes the correlations between these selected subsets using a regularization procedure similar to LASSO . Adaptive SCCA [19] selects fewer features which are more correlated and Waaijenborg et al . [20] , propose a method called penalized CCA to find associations between gene expression and copy number variation data . Other variants on CCA which are applicable include non-linear extensions of CCA , such as kernel CCA [21] , [22] , Bayesian approaches to CCA [23] , [24] and sparse CCA models for handling more than two types of data [25] . CCA for association analysis was proposed initially by Ferreira and Purcel [26] and subsequently extended [27] . Both these papers apply CCA to multiple trait/single genotype analysis ( pleiotropy analysis ) , while the latter also considers the case of several markers ( gene centered pleiotropy analysis ) and several traits , or several markers and one trait ( epistasis analysis ) . Since the original publication [26] , CCA has been used for multiple association analysis elsewhere , including a single SNP , multiple phenotype association approach [28] to analyze blood phenotypes related with metabolic syndrome in mice , and use of a sparse version of CCA to discover associations between single locus and multiple neuroimage phenotypes [29] . Further applications of CCA include a study [26] of pleiotropy in white cell related traits using a single locus/multiple trait approach , and use of CCA for single SNP/multiple trait analysis to find different child behavior profiles [30] . In this paper we propose an alternative approach for using CCA in which we select feature sets via biological insight , based on association with a gene , a pathway or another biologically relevant grouping . As detailed below , to maximize the association between genetic data and different phenotypes , we combine the CCA approach proposed by Ferreira and Purcell [26] with an optimization technique , drawn from integer programming . We will refer to any discovered significant associations between subsets of the genetic and phenotype data as putative association rules .
This approach consists of a gene centered association analysis with each single phenotype using simple CCA without any search heuristic . It is exactly the same approach used previously by Tang and Ferreira [27] , consisting of a multiple association of all the SNPs close to a gene ( see Methods for more detail ) with a particular phenotype . In order to correct for multiple testing , we use a Bonferroni correction for 3648 genes and 82 phenotypes , giving a “threshold” p-value of 1 . 67×10−06 corresponding to p = 0 . 05 for a single test . We found 62 genes with significant association ( p<1 . 67×10−06 ) . Most of the time this association reflects the most associated SNP in a gene . The most important associations are presented in the hive plot in Figure 1 . In Table 1 we show some of the associations found and publications that supports these findings . All such associations can be found in Table S1 , where we compare the association values between this approach and conventional single SNP association tests . Although most genes have more than one associated SNP reading , we found a non-reported association ( p = 4 . 93×10−10 ) between the single SNP rs11264341 , located in the intronic region of gene TRIM46 ( ENTREZ GENE # 80128 ) , and the serum magnesium phenotype . This SNP is in LD with SNP rs4072037 ( r2 = 0 . 54 ) in MUC1 , which has been previously related with serum magnesium . Close to this SNP ( >8 kb ) , and not in LD , we found an association in gene MUC1 ( ENTREZ GENE # 4582 ) with serum magnesium ( p = 1 . 37×10−14 ) which has been previously reported by Meyer et al . [31] . We found a previously reported association of gene SURF4 ( ENTREZ GENE # 6836 ) and Von Willebrand Factor ( vWF ) , but also a non-previously-reported association with Factor VIII ( 1 . 57×10−24 ) and alkaline phosphatase ( ALP ) ( 3 . 1×10−9 ) . SNPs in SURF4 are in some LD ( r2 = 0 . 696 ) with SNP in C9orf96 , which has been related with vWF [32] , also with SNPs in ABO ( ENTREZ GENE # 28 ) ( r2 = 0 . 502 ) which has been related with vWF [32] and ALP [33] [34] . We could expect the association between Factor VIII and vWF , because there is a high correlation between its serum concentrations ( 0 . 70 ) , and vWF acts as a carrier protein of Factor VIII . However , the correlation between ALP and vWF are 0 . 14 . This is a clear example of vertical pleiotropy , where variants in SURF4 are causal of vWF , and vWF glycoprotein is the carrier for Factor VIII glycoprotein in blood . Another important association ( p = 2 . 18×10−8 ) , which has not been reported , is between gene NSF ( ENTREZ GENE # 4905 ) and triglycerides . NSF is related with genes KIAA1377 ( ENTREZ GENE # 57562 ) and LUC7L2 ( ENTREZ GENE # 51631 ) , through the PPI network , which are also related with the LPL gene ( ENTREZ GENE # 4023 ) . Finally the MYBPHL gene is associated ( p = 3 . 19×10−8 ) with low density lipoprotein ( LDL ) cholesterol , which has not been previously reported . However , SNPs in this gene are in LD with SNPs in CELSR2 ( ENTREZ GENE # 1952 ) ( r2 = 0 . 546 ) , PSRC1 ( ENTREZ GENE # 84722 ) ( r2 = 1 ) and SORT1 ( ENTREZ GENE # 6272 ) ( r2 1 ) rs12740374 , which is associated [35] , [36] , [37] , [38] , [39] with LDL cholesterol . Left ventricular hypertrophy can be detected through ECG parameters such as Cornell product [40] or QRS product [40] , [41] . Using the CCA gene-centered association approach we have identified a number of genes associated with these two clinical parameters , which are also positively associated with cardiovascular diseases such as stroke [42] . We found association between ACSM3 ( ENTREZ GENE # 6296 ) and Cornell product ( p = 2 . 38×10−8 ) . This gene was previously reported to associate with hypertension in rats [43] and in humans [44] and also with obesity hypertension in humans [45] , but there is some controversy [46] . Other studies relate it with ventricular deformations such as left ventricular mass index and mean wall thickness [47] . The ERI2 gene was also associated with Cornell product ( p = 7 . 87×10−9 ) . This gene overlaps ACSM3 ( ERI2 SNPs is a subset of ACSM3 ) . No association with left ventricular hypertrophy or hypertension has been reported previously . IL18RAP ( ENTREZ GENE # 8807 ) was associated with Cornell product ( p-value 1 . 07×10−8 ) and QRS voltage product ( p-value 1 . 72×10−10 ) . SNPs in this gene have been associated [48] with echocardiography left ventricular obtained measures . In Grisoni et al . [49] , using different SNPs in the same gene , the authors did not find any association between IL18RAP and any cardiovascular diseases ( CVD ) risks . However , Tiret et al . [50] found a significant association between IL18 family gene SNPs and mortality . We found association between IL23R ( ENTREZ GENE # 149233 ) and Cornell product ( 3 . 3×10−12 ) . This gene has been associated with left ventricular hypertrophy [51] and idiopathic dilated cardiomyopathy in Chen et al . [52] . It is interesting to note the importance of autoimmune related genes ( IL18RAP and IL13R ) in left ventricular hypertrophy or idiopathic dilated cardiomyopathy . A relation between autoimmune response and idiopathic dilated cardiomyopathy has been suggested in San Martin et al . [53] and Lappe et al . [54] . Finally , gene NRG1 ( ENTREZ GENE # 3084 ) presents an association with phenotypes Cornell product ( 2 . 71×10−14 ) and QRS voltage product ( 2 . 97×10−11 ) . This gene has been associated to cardiovascular development in mouse [55] , through the NRG1/ErbB signaling pathway [56] , [57] , that is involved in angiogenesis , blood pressure and skeletal muscle response to exercise . In humans , serum NRG-beta has been found elevated in patients with severe systolic heart failure [58] . In McBride et al . [59] , no association was found between SNPs in NRG1 and a group of congenital heart malformations ( left ventricular outflow tract , defects of aortic valve stenosis , coarctation of the aorta and hypoplastic left syndrome ) . One of the drawbacks of CCA analysis , which could affect our understanding of the role of NRG1 , is that this method lacks power when a gene is larger than 100 Kb [27] , and NRG1 has a length of 1 . 1 Mb . In order to analyze the association of all the SNPs in one gene and multiple phenotypes , we use CCA and a genetic algorithm as an optimization method , to select the most important phenotypes , as described in the Methods section . In Table 2 we show some of the most important pleiotropic genotype/multiple phenotype associations , including the p-value of CCA association and the phenotypes with which they are associated . We also show Fisher's combined association value and , in parentheses , the association value of the genes and the single phenotype . In Table S2 we show all the results for associations between one gene/multiple phenotypes . In order to correct for multiple associations , we use a Bonferroni correction for 3648 genes and combinations of 82 phenotypes in subsets of 24 to 2 groups . We chose 24 because it is the maximum number of different phenotypes in one association rule ( an association rule is a combination of a number of phenotypes associated with a number of genes ) selected by the genetic algorithm ( see the multiple test association correction paragraph in Methods ) . This combination gives 5 . 36×1020 different phenotypic rules , giving a threshold p-value of 2 . 55×10−25 equivalent to p = 0 . 05 for a single test . In Figure 2 , we use a heatmap plot to represent the most important ( higher association ) pleiotropic relations between phenotypes and genotypes . Also , we use a hive plot ( interactive plot available online ) in Figure 3 . In this diagram , vertical axis represents the association between the phenotype ( left axis ) and genotype ( right axis ) . Association rules are ordered in the diagram following the association value ( the higher association , the higher in the plot ) . Gene ABO which has an indicated association ( p-value 2 . 47×10−147 ) with coagulation ( tissue plasminogen activation , Factor VIII and Von Willebrand factor levels ) , but also with serum levels of ALP ( previously reported in Yuan et al . [34] and creatinine . Gene SURF4 ( ENTREZ GENE # 6836 ) , which has been previously associated with Von Willebrand Factor , Factor VIII and ALP , is also associated with ECG measures , MMP-9 ( inflammatory marker ) and mean cell volume ( average red blood cell volume ) among others . Gene HRG presents a weak association ( p-value 1 . 88×10−24 , corrected threshold 2 . 55×10−25 ) with some factors related with coagulation , such as activated partial thromboplastin time ( APTT ) , ratio activated protein C ( APC ) /APTT , volume , total protein and Factor IX . Finally gene CETP ( ENTREZ GENE # 1071 ) shows weak association ( p-value 3 . 65×10−20 ) with cholesterol as expected , but also with coagulation factors ( Von Willebrand Factor , Factor VII and sCD40 ligand ) . Gene F10 ( ENTREZ GENE # 2159 ) , presents association with coagulation factors ( Factor VII and Factor IX ) , but also with cholesterol , similar to gene F7 , which also presents association with diastolic blood pressure . For the gene NRG1 , we found association with ECG measures of ventricular hypertrophy , but also with urea and fibrinogen . Gene IL18RAP is weakly associated ( p-value 5 . 07×10−21 ) with white cell counts ( white cells , neutrophils , lymphocytes ) , with alanine transaminase ( ALT ) and glucose , but also with ECG measures of ventricular hypertrophy . Gene IL23R is weakly associated ( p- value 2 . 19×10−18 ) with levels of interleukin 18 but also with adiponectin and ECG measures of ventricular hypertrophy ) . Gene ALOX5AP ( ENTREZ GENE # 241 ) has been related with myocardial infarction and stroke [60] , and also with inflammatory activity and atherosclerosis [61] . In our results it presents some association with some phenotypes related with immune response ( white blood count , neutrophils , CD40 or total protein ) but also with some markers of ECG related with hypertension . And gene GPR98 ( ENTREZ GENE # 84059 ) is related in our analysis with immune response phenotypes and insulin related phenotypes ( insulin , HOMA score ) and in some cases with an association with CVD . No relation between this gene and these phenotypes has been reported , but some association was reported with carotid diseases and body weight [62] , [63] . In this case , instead of selecting the most associated phenotypes for each gene , the GA selects the most associated genes for each phenotype . This operation is more computationally expensive than the previous one , because of the high number of genes ( 3648 ) involved . In order to correct for multiple testing , we use a Bonferroni correction for 82 phenotypes and a combination of 3648 genes in subsets of 29 , 28 , 27…1 groups . We choose 29 because this is the maximum number of different genes in one rule . This combination gives 2 . 03×1072 different genotypic rules , giving a threshold p-value of 2 . 99×10−75 equivalent to p = 0 . 05 for a single test . Ferreira [27] comments there may be a lack of power related with gene size for CCA for the case of multiple gene analysis . However , we consider that this analysis could contribute if the involved genes are small . Some of the most interesting rules are shown in Table 3 . Other significant and non-significant enrichment analyses of the genes in the rules are listed in supplementary Table S3 and S4 . The Von Willebrand factor association ( p-value 1 . 69×10−117–2 . 49×10−119 ) is led by individual association with gene ABO ( p-value 9 . 43×10−112 ) , and two of three significant pathways present more CCA association that Fisher multiple association . The bilirubin association ( p-value 6 . 76×10−115–3 . 79×10−118 ) is most influenced by genes in the UGT1 family , and all pathways present more CCA association than Fisher multiple association . The FVII association is led by genes F7 and EDEM2 ( ENTREZ GENE # 55741 ) or PROCR ( ENTREZ GENE # 10544 ) . Finally FVIII association is led by ABO gene . Regarding the enrichment analysis ( Table S4 ) , some interesting enrichments has been found , such as Factor VII and Human Phenotype Pathway “Abnormality of the coagulation cascade” , KEGG pathway “Complement and coagulation cascades” and Reactome pathway “Formation of Fibrin Clot ( Clotting Cascade ) ” , or Factor VIII and KEGG pathways “ECM-receptor interaction” ( pathway related with hemophilia , directly related with factor VIII ) . From non significant rules , APTT related genes are annotated with GO Terms “negative regulation of blood coagulation” , “blood coagulation fibrin clot formation” , “blood coagulation intrinsic pathway” and Reactome pathway “formation of fibrin clot” . Finally , LDL cholesterol is annotated with LDL gene related annotations Finally , we use a CCA - two population genetic algorithm approach for multiple gene/multiple phenotype rule extraction . As a result , a set of 56 rules that relate the most associated set of genes with phenotypes was obtained . Following our previous multiple association corrections , the maximum number of genes in the obtained rules is 22 and the maximum size of the phenotypes is 9 , so there is a possible population of 1 . 94×1057 gene rules and 3 . 3×1011 phenotypes , that determine a threshold p-value of 7 . 71×10−70 ( equivalent to p = 0 . 05 for a single test ) . Table 4 shows some of these association rules , and a complete list of 56 rules can be found in Table S5 . An enrichment analysis can also be found in Table S6 . The bigger association obtained rule , genes F7 , ABO , MRPS28 , UGT1A3 and SURF4 with phenotypes bilirubin FVII and vWF , presents an association probability under 2 . 22×10−308 , which was below our machine precision and therefore recorded as zero . We have identified some patterns in the multiple genes/multiple phenotype pleiotropic rules . ABO and SURF4 has similar relations with ALP , FVIII and vWF , F7 and F5 with FVII , F5 and HRG with APTT and ratio APC/APTT , F12 with APTT and NRG with Cornell product and QRS voltage product . Most of the rules obtained here are combinations of these . The enrichments analysis of multiple phenotypes reveals interesting results , such as a rule formed by phenotypes bilirubin , alp , APTT , ratio APC/APTT and Von Willebrand Factor which were enriched for HP pathways “Prolonged partial thromboplastin time” and “Prolonged whole-blood clotting time” , KEGG pathway “Complement and coagulation cascades” and Reactome pathway “Formation of Fibrin Clot ( Clotting Cascade ) ” . This rule is not a clear example of pleiotropy , because all genes and phenotypes are related with clotting , but it is clear that the inclusion of all genes and phenotypes in the same rules increases the association . Rules including phenotypes QRS duration , Cornell Index and Cornell Product , are annotated with hypertension GO terms and linked with genes that support these annotations . Also rules including phenotypes related with left ventricular hypertrophy are enriched with the GO term “epithelium development” and linked with genes related with cardiovascular development .
In the case of single gene/single phenotype analysis , we are not looking for pleiotropic effects , but for a combined gene-based association effect , and some interesting results were found . The complete list of gene-based significant and previously reported associations can be found in Table 1 . One of the drawbacks of CCA analysis , which could affect our understanding of the role of NRG1 , is that this method lacks power when a gene is larger than 100 Kb [27] , and NRG1 has a length of 1 . 1 Mb . In the case of single gene and multiple phenotype association , our results show that the p-values ( both CCA and Fisher ) increase when more related phenotypes are included in the phenotype set . As expected , most of these phenotypes are correlated/associated . However , not all phenotypic sets are correlated . An example can be observed in gene F7 ( ENTREZ GENE # 2155 ) , which is associated with phenotypes total cholesterol , Factor VII and Factor IX . Correlation exists between total cholesterol and Factor VII ( 0 . 28 ) , Factor VII and Factor IX ( 0 . 39 ) , but not between cholesterol and Factor IX ( 0 . 09 ) . In some cases , Fisher's combined p-value approach shows equal or bigger association than CCA , which could mean that CCA association shows the cumulative effects of individual associations . In contrast , when Fisher's multiple association p-value is smaller than CCA association , this could suggest that CCA association analysis has found pleiotropic effect between a gene and these phenotypes . Some examples of the first group are association of gene ABO with coagulation phenotypes . In contrast , examples of pleiotropic effects appear in genes F10 ( ENTREZ GENE # 2159 ) or F7 , which presents an association with coagulation factors ( Factor VII and Factor IX ) , but also with cholesterol . In the case of multiple gene/single phenotype , using this robust association threshold , we have identified a set of pathways that are associated significantly with phenotypes Von Willebrand factor , bilirubin , FVII and FVIII . The whole list of pathways is listed in Table S3 . It's interesting to see that there is no significant difference between the CCA and Fisher's association , in contrast with the differences shown in the previous Section , which supports the fact that CCA could detect pleiotropy patterns . In conclusion , in this paper we have applied a canonical correlation analysis approach for association in multivariate datasets , finding correlations between gene-centered genetic variants and phenotypes . This multivariate approach allows us to mine pleiotropic relations between one or a set of genes and a set of phenotypes . In term of single gene/single phenotype association , we have found non-reported associations of gene NSF and triglycerides and genes ACSM3 . ERI2 , IL18RAP , IL23RAP and NRG1 with phenotypes related with left ventricular hypertrophy . We use a genetic algorithm as feature selection algorithm in order to find pleiotropy patterns in phenotypes . Using this approach we found pleiotropy patterns in genes F7 and F10 with phenotypes Factor VII , Factor IX and cholesterol; NRG1 , with left ventricular hypertrophy related phenotypes , but also with fibrinogen and urea or IL18RAP or IL23RAP , related with immune response related phenotypes , but also with ECG measures . Despite the possible drawbacks of CCA , related to power when the length of a gene is greater than 100 Kb , or increases of type I error when features are not normally distributed , we found that CCA can be used as a powerful tool to find gene-centered association , multivariate association and pleiotropic patterns . Also , this tool can be extended to find non-linear canonical correlation relations using kernel based approaches such as KCCA . Future research directions include improving the search method , using other meta-heuristics such as Tabu Search , Simulated Annealing or Particle Swarm Optimization , or sparse regularization methods .
The British Women's Heart and Health Study ( BWHHS ) is a UK-based prospective cohort study of 4286 healthy women aged 60–79 years at baseline ( 1999–2001 ) . Participants were selected at random from general practice registers in 23 UK towns [64] . A range of baseline data sources ( blood samples , anthropometry , health/medical history , echocardiography measures , etc . ) was collected between 1999 and 2001 , and DNA extracted from 3884 participants . Although the cohort has been followed-up in subsequent phases , all data presented here is based on the recruitment ( baseline ) phase . Multi-centre ( London Multi-centre Regional Ethics Committee ) and local research ethics committees provided approval for the BWHHS study and informed consent was obtained from the women to complete the data used in this study . Genotyping was performed using the Illumina HumanCVD BeadArray ( Illumina Inc , San Diego , USA ) , which comprises nearly 48 , 742 SNPs in over 2 , 100 genes selected on the basis of cardiovascular candidacy by an international consortium of experts [65] . Genotypes were called using a Illumina BeadStudio ( v3 ) Genotyping Module . Samples with a genotype call rate <90% , Hardy Weinberg disequilibrium <10−7 and minor allele frequency <1% were excluded from the analysis , following insight from previous work on this array and patient cohort [66] . Non-European samples were also excluded from analysis . Principal components analysis identified no evidence of population stratification ( consistent with self-reported ancestry ) . The different phenotypes used in this study consisted of 11 directed and derived electrocardiogram ( ECG ) measures , obtained as described in Gaunt et al . [67] , 64 blood measures , 2 blood pressure readings , 3 anthropometric measures , HOMA score ( derived from glucose and insulin values ) and an indicator of whether a patient has suffered cardiovascular disease . These data were measured as described in Lawlor et al . [64] . All data were analyzed using R ( The R project for statistical computing , http://www . r-project . org/ ) . Due to the high number of missing values present in the phenotypic data ( 7575 of 312984 values , median of 55 ( 1 . 42% ) missing values per phenotype , max 509 ( 13 . 17% ) and min 19 ( 0 . 49% ) ) , we followed a strategy of phenotypic data imputation based on a k-nearest neighbor approach , implemented in the R package “Imputation” [68] ( http://cran . r-project . org/web/packages/imputation/index . html ) , with a k of 5 . In order to test how these imputed values affected the association profile , we compared the single association values of imputed data versus data with missing values removed . The results show that the associations are the same or lower in the imputed values , so imputation does not create false associations . All phenotypic data was normalized to mean zero and standard deviation one . All the approaches for analysis in this work were based on a “gene-centered” perspective . Genotype data , both intronic and exonic , was assigned to the genomically closest gene using the function “ClosestBED” from the suite “BEDTools” [69] ( http://bedtools . readthedocs . org/en/latest/ ) . In order to avoid multicollinearity in genotype data , we applied two-stage linkage disequilibrium ( LD ) pruning as described in Tang and Ferreira [27] . We removed SNPs with a high LD ( r2>0 . 64 ) with other markers and also a high correlation between linear combinations of SNPs using Variance Inflation Factor ( VIF ) [70] in order to exclude SNPs with a VIF>2 with other markers . In order to select the most appropriate value for r2 , we developed several experiments to test the CCA single gene/single phenotype association using a range of r2 ( 0 . 5–0 . 99 ) , and best results was obtained pruning SNPs with r2>0 . 64 . The value of VIF>2 was selected based in the recommendation of the original CCA paper [27] . As mentioned above , unlike other approaches to pleiotropy analysis , in this study we used a gene-centered approach . This perspective allowed us to capture all the pleiotropic effects in one gene , instead of the pleiotropic effects caused by just one variation . But we are also interested in studying the pleiotropic effects of a set of genes in several phenotypes . In order to do this , we divided the study into four stages . Firstly , we studied the individual association between each gene ( which may consist of one or more SNPs ) and a single phenotype to establish a gene-centered association baseline . This approach did not reveal any pleiotropy , of course , but it is worth pursuing for two reasons . Firstly , it was interesting to find if inclusion of several SNPs increases the association value over a single SNP approach . Secondly , we got a baseline gene association value that we used as a comparator for the CCA association analyses in our subsequent analysis . For our second stage we studied the association between one single gene and a set of phenotypes . The aim of this analysis was to reveal possible gene-based pleiotropic effects . Our next stage was to study association effects between multiple genes and a single phenotype ( gene-based epistasis analysis ) . The aim of this analysis was to discover pathway based baseline association between a set of genes and a single phenotype . Finally , our last stage consisted in studying the association between a set of multiple genes and different phenotypes . Here we expected to find the pleiotropic effects of a set of genes in multiple phenotypes , with increased statistical significance for the indicated association rules . Canonical Correlation Analysis ( CCA ) allows us to find linear combinations of two sets of variables with the highest correlations . The aim of this work was to find correlation between a set of genotype data and a set of phenotype data . The CCA algorithm was based on a method proposed by Tang and Ferreira [27] . In order to test the significance of all canonical correlations , Wilk's Lambda and Rao's F approximation were calculated . Let q be the number of SNPs in the genotype , p the number of phenotypes evaluated , n the number of samples and cj the number of canonical components calculated . Wilk's Lambda is calculated as follow:And Rao's F approximation:WhereThe methods for CCA analysis analyzed in the previous Section could be computationally expensive with a large number of features and samples . In order to use standard CCA we can also divide the feature set into small subsets using biological insight ( eg the set of SNPs in the region of a specific gene ) . In this paper we will simply use feature sets in which the SNPs are linked to a single gene , defining the link to a gene by genomic proximity . In this case , feature selection is not necessary because the number of samples is larger than the number of features . To find those sets which have a high correlation , according to CCA , we need to use an optimization method , with the association value as the fitness function for this optimization procedure . We have formulated this optimization step as an integer programming problem which can therefore be addressed using a metaheuristic procedure to find an approximately good solution in a computationally tractable time . We have decided not to use methods such as hill climbing or similar local methods , because they are prone to capture by local minima . In this particular problem , any big single association could be assumed to be a local minimum and a hill climbing approach could not exit easily . Instead of this , we have decided to use global methods , such as Tabu Search [71] , Particle Swarm Optimization [72] or Genetic Algorithm ( GA ) used here , as a well known and well used approach to this type of problem and with an effective means for evading local minima . A GA is a metaheuristic , initially proposed by Holland [73] and Goldberg [74] . This procedure is based on the principles of evolution and natural selection , with steps analogous to inheritance , mutation and crossover . It is initialized with a set of solutions , each representing one possible solution to the problem . The performance of each proposed solution is estimated using the fitness function , which measures how well an individual solution is adapted to the proposed problem . The method then iteratively evolves a high-fitness solution . The “genalg” R package ( http://cran . r-project . org/web/packages/genalg/index . html ) was used as a binary implementation of a GA . However , because of the requisites of the multiple gene/multiple phenotype analysis , this code was modified in order to include two population searches ( Modified source of genealg package is available in http://github . com/jseoane/gaCCA ) . One of those populations represents different solutions for gene selection , and the other represents different solutions for phenotype selection . The search strategy is applied in parallel over the two populations and the fitness function is evaluated simultaneously over the selected set of genes and selected set of phenotypes , when calculating their CCA association value . The encoding of the genetic algorithm is a binary encoding , widely used in feature selection approaches , where if the feature is set to 1 , it is included in the analysis and is not included otherwise . The fitness function in the three versions is defined by the CCA association value . Regarding the parameterization , the population size is 100 for single gene/multiple phenotype , 600 for multiple gene/single phenotype and 1000 for multiple gene/multiple phenotype . The mutation depends on the size of the GA chromosome ( 1/82 in the case of multiple phenotype , 1/3248 for multiple gene ) . The elitism ( how many samples of the population are conserved between generations ) is 20/35/100 , respectively for each of the versions . Finally , the “zero to one ratio” , which controls the number of features in the chromosome is set to 50 in the case of multiple phenotype and 700 in the case of multiple gene . In order to avoid multiple testing associations which arise by chance , we applied a Bonferroni correction . In this case , the Bonferroni correction should be applied to both sides of the association . In this case the association is calculated over p phenotypes and g genes , so a 0 . 05 of confidence should need 0 . 05/ ( p*g ) . But when a GA is used , millions of associations are considered , so we approximate the Bonferroni correction over the search space of the algorithm ( i . e . if we expect rules of p′ phenotypes and g′ genes , the search space over genes are combinations of g different genes over g′ , g′-1 , g′-2 , ‥ , 2 , and the search space over phenotypes are combinations of p different phenotypes over p′ , p′-1 , p′-2 , ‥ , 2 ) . The highly conservative final association threshold proposed is 0 . 05/ ( length of search space in genes * length of search space in phenotypes ) , though we ranked and considered all results by p-value in our analysis . In order to compare the CCA combined association measure with other measures , we have chosen a statistical measure based on Fisher's combined p-value approach proposed in Li et al . [5] . During phase two and phase four of the analysis , a set of genes related with one or several phenotypes is obtained . In order to functionally annotate these sets of genes , we perform an enrichment analysis , detecting GO ontology terms , KEGG , Reactome or Phenotype annotations that are significantly present in our pathways . We use the enrichment analysis tool g:Profiler [75] ( http://biit . cs . ut . ee/gprofiler/ ) , through R package “gProfiler” ( http://cran . r-project . org/web/packages/gProfileR/index . html ) . In order to calculate the p-values for each enrichment , the method first simulate 10 millions of queries ( sets of genes ) randomly to see how was the p-values distribution according the query size . Then analytically derived the p-value threshold for each query size ( for more details consult g:SCS threshold section in the Reimand paper ) . Because some association values were close to zero , note that all calculations were performed in a 64-bit Linux R environment where the lowest positive value is 2 . 22×10−308 , which means that values below this threshold were treated as zero . | Pleiotropy appears when a variation in one gene affects to several non-related phenotypes . The study of this phenomenon can be useful in gene function discovery , but also in the study of the evolution of a gene . In this paper , we present a methodology , based on Canonical Correlation Analysis , which studies gene-centered multiple association of the variation of SNPs in one or a set of genes with one or a set of phenotypes . The resulting methodology can be applied in gene-centered association analysis , multiple association analysis or pleiotropic pattern discovery . We apply this methodology with a genotype dataset and a set of cardiovascular related phenotypes , and discover new gene association between gene NRG1 and phenotypes related with left ventricular hypertrophy , and pleiotropic effects of this gene with other phenotypes as coagulation factors and urea or pleiotropic effects between coagulation related genes F7 and F10 with coagulation factors and cholesterol levels . This methodology could be also used to find multiple associations in other omics datasets . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"genome-wide",
"association",
"studies",
"medicine",
"and",
"health",
"sciences",
"applied",
"mathematics",
"algorithms",
"mathematics",
"statistics",
"(mathematics)",
"genome",
"analysis",
"trait",
"locus",
"analysis",
"population",
"biology",
"biostatistics",
"genetic",
"epidemiology",
"cardiology",
"epidemiology",
"statistical",
"methods",
"genetic",
"association",
"studies",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"computational",
"biology",
"human",
"genetics"
]
| 2014 | Canonical Correlation Analysis for Gene-Based Pleiotropy Discovery |
Planar cell polarity ( PCP ) instructs tissue patterning in a wide range of organisms from fruit flies to humans . PCP signaling coordinates cell behavior across tissues and is integrated by cells to couple cell fate identity with position in a developing tissue . In the fly eye , PCP signaling is required for the specification of R3 and R4 photoreceptors based upon their positioning relative to the dorso-ventral axis . The ‘core’ PCP pathway involves the asymmetric localization of two distinct membrane-bound complexes , one containing Frizzled ( Fz , required in R3 ) and the other Van Gogh ( Vang , required in R4 ) . Inhibitory interactions between the cytosolic components of each complex reinforce asymmetric localization . Prickle ( Pk ) and Spiny-legs ( Pk-Sple ) are two antagonistic isoforms of the prickle ( pk ) gene and are cytoplasmic components of the Vang complex . The balance between their levels is critical for tissue patterning , with Pk-Sple being the major functional isoform in the eye . Here we uncover a post-translational role for Nemo kinase in limiting the amount of the minor isoform Pk . We identified Pk as a Nemo substrate in a genome-wide in vitro band-shift screen . In vivo , nemo genetically interacts with pkpk but not pksple and enhances PCP defects in the eye and leg . Nemo phosphorylation limits Pk levels and is required specifically in the R4 photoreceptor like the major isoform , Pk-Sple . Genetic interaction and biochemical data suggest that Nemo phosphorylation of Pk leads to its proteasomal degradation via the Cullin1/SkpA/Slmb complex . dTAK and Homeodomain interacting protein kinase ( Hipk ) may also act together with Nemo to target Pk for degradation , consistent with similar observations in mammalian studies . Our results therefore demonstrate a mechanism to maintain low levels of the minor Pk isoform , allowing PCP complexes to form correctly and specify cell fate .
Planar cell polarity ( PCP ) instructs tissue patterning in a wide range of organisms from Drosophila to humans , through input into cellular orientation across tissues , individual cell fate decisions , and the coordinated movement of groups of cells [1–8] . In the Drosophila eye , Frizzled core PCP signaling coordinates the cell fate decisions of individual photoreceptors , and their subsequent collective movements during ommatidial rotation , via asymmetric localization of two distinct membrane-bound complexes on opposite sides of a cell [1–7 , 9] . The two core Fz/PCP pathway complexes comprise of Frizzled/Dishevelled/Diego ( Fz/Dsh/Dgo ) in one complex , and Van Gogh/Prickle ( Vang/Pk ) ( Vang , also known as Strabismus/Stbm ) in the other . These complexes are localized to opposite sides of the cell and stabilized intercellularly via the atypical cadherin Flamingo ( Fmi ) associating with both complexes [1–7] . Each component is highly conserved between Drosophila and vertebrates , and mutation of PCP genes in humans is linked to a range of diseases from spina bifida to polycystic kidney disease and epilepsy [10] . Feedback between the two complexes is essential to reinforce Wnt-induced cellular orientation bias [11–13] into coordinated tissue-wide polarity . Positive intercellular interactions between transmembrane factors , Fz , Vang and Fmi , relay positional information and negative intracellular interactions between cytosolic factors Pk and Dsh/Dgo enhance asymmetry on a cellular level [1 , 3 , 5 , 7 , 14] . In mammals there are four prickle genes and although there is only one prickle gene in Drosophila the range of Prickle functions are performed by distinct isoforms [15] , Prickle ( Pk ) , Spiny-legs ( Pk-Sple ) and PrickleM . Pk and Pk-Sple are the two functionally relevant isoforms during establishment of PCP . The balance between the two isoforms is tissue specific: Pk-Sple is the ‘major’ isoform in eyes and legs , and Pk the ‘major’ isoform in wings [15–17] . The precise balance has functional significance since the two isoforms can antagonize each other and/or the other’s function , although the underlying mechanism is not well understood [15–17] . Recent work has shown that this isoform balance is regulated transcriptionally at the tissue level in the wing , where Pk mRNA is present at 10-15-fold higher levels than Pk-Sple mRNA [17] . However , it is unclear how the balance is maintained in the eye , because it cannot be explained by transcriptional regulation: Pk mRNA is actually expressed at slightly higher levels than Pk-Sple mRNA [17] , even though Pk-Sple is the major functional isoform here . When the Pk/Pk-Sple balance is perturbed it causes PCP defects , for example by overexpression of one isoform or isoform-specific alleles such as pkpk1 , in which expression of only the Pk isoform is lost and Pk-Sple expression is unchanged [15–17] . Recent reports demonstrate that imbalance between the specific Pk and Pk-Sple isoforms causes seizures in Drosophila and , moreover , they suggest that disrupting PRICKLE genes underlies cases of epilepsy in humans [18 , 19] . The Drosophila eye is a compound eye with ~800 individual ommatidia , each containing eight photoreceptor neurons ( R1-8 ) arranged as a trapezoid [20–22] ( also Suppl S2 Fig ) . Chirality of the trapezoid is determined by positioning of the R3/R4 pair , whose fate is specified by Fz/PCP signaling . Fz activity is higher at the dorso-ventral midline ( equator ) of developing eye discs [20] . Consequently , for each R3/R4 pair the cell closer to the midline exhibits increased Fz activity , adopts the R3 fate , and signals to its neighbor to induce it as R4 . Ommatidial preclusters then undergo a 90° rotation that is coordinated across the field by PCP activity , resulting in a line of symmetry around the equator . Disruption of PCP signaling causes chirality defects , whereby the R3/R4 fate decision becomes uncoupled from positional information or fails to be resolved [20] . PCP defects also involve misregulation of ommatidial rotation ( OR ) such that OR is no longer coordinated across the tissue [20–22] . Here , we report a new function for Nemo ( Nmo ) kinase , a classic ‘OR’ gene [23 , 24] , and demonstrate its role in regulating levels of the Pk isoform of pk via direct phosphorylation of Pk and its targeting for proteasomal degradation . Nmo is required in specific cells , the R4 cells , where the Pk isoform needs to be suppressed in the eye , and also in PCP mediated leg patterning , where Pk is also the minor functional isoform . Our results establish a new regulatory mechanism of PCP factors with tissue- and cell-specific regulation of core PCP protein degradation being coupled to PCP-mediated cell fate induction and function .
In order to better understand how Nmo acts as a PCP effector in the eye , we performed a genome-wide , gel-shift based screen for novel Nmo kinase substrates , similarly to our recent studies identifying novel dROK and Hpo substrates [25 , 26] . In brief , pooled cDNA clones were in vitro translated and labeled with [35]S-Methionine , and then incubated with purified Nmo in the presence of unlabeled ATP . Reduced mobility on Anderson gels was used as the criterion to select candidate substrates . Surprisingly , one of the candidates identified in this screen was Prickle , an ‘upstream’ core component of PCP complexes . Nmo , but not dRok or Hpo kinases , was able to induce a band shift of Prickle ( Fig 1A ) , as well as the positive control Pan/dTCF [27] , but not the negative control , Mbs ( Fig 1A and Suppl . S1 Fig ) . Band shift assays of cell culture extracts confirmed that Nmo kinase activity reduces the mobility of Prickle ( Suppl . S1 Fig ) . Using purified Nmo kinase we performed in vitro kinase assays and determined that Nmo directly phosphorylated the common region of Pk . The two main Prickle isoforms required for PCP , Pk and Pk-Sple , are largely identical except for an extended Pk-Sple N-terminal region ( outlined in Fig 1B ) . To identify the Nmo phosphorylation site ( s ) within the common Pk sequence , we designed a series of deletion constructs ( outlined in Fig 1B ) with the Vang C-term included as a negative control [24] . Through comparison of the phosphorylation of these constructs in in vitro kinase assays , the phosphorylation sites were mapped to a middle region of the common Pk sequences ( fragment M , Fig 1B and 1C ) . Nmo did not phosphorylate the Sple N-terminus , PET and LIM domains , or the region within the Pk C-terminus required for binding to Vang ( contained within fragment C1 [28] ) . Our previous studies had suggested a model whereby Vang recruited Nmo to the membrane in ‘mature’ ommatidial clusters , where it acted as an effector and phosphorylated β-catenin to promote cluster rotation [24] . Prickle being a Nmo substrate raised the possibility that Nmo had an earlier role in regulating the PCP complexes themselves . Although the zebrafish nmo homolog , nemo-like kinase had been known to genetically interact with the Wnt/PCP pathway during the coordinated movements of convergent extension in the embryo [29] , our result is the first indication that nmo directly affects the core PCP factors during the establishment of PCP itself . Given that Nmo phosphorylated a region that is shared between the two prickle isoforms , Pk and Pk-Sple , we wanted to determine whether one or both isoforms were functionally affected in vivo . We performed genetic interaction assays with nmo mutants and isoform specific alleles , pkpk1 and pksple1 ( loss of function , LOF ) or over-expressed individual Pk and Pk-Sple isoforms ( gain of function , GOF ) . We first examined the interaction between nmo and pksple1 in the eye . Wild-type and nmoP hypomorphs show wild-type ommatidial chirality , in addition to the previously-described underrotation in nmo mutant clusters ( Fig 2A , 2B and 2M ) [23 , 24 , 30] . pksple1 clusters adopt almost random chirality , whereby the R3/R4 fate decision is resolved , but the dorsal vs ventral chiral arrangements are intermixed , termed ‘chirality flips’ , and R3/R4 fate is uncoupled from dorso-ventral positioning ( Fig 2C and 2M ) . However , in the pksple1; nmoP double mutant , a significant proportion of achiral , symmetrical clusters were observed; a phenotype rarely observed in either single mutant ( Fig 2D and 2M , note over 20-fold increase in symmetrical clusters—see Suppl . S2A Fig for schematic of photoreceptor arrangement ) . We confirmed the achiral nature of the clusters by immunostaining larval eye discs using the mδ-LacZ construct , which in wild-type specifically labels the R4 precursor [31 , 32] ( Fig 2E and 2F and Suppl . S2B Fig for overview ) . Clusters with two negative or two positive cells can often be seen in the double mutant , but not wild-type , where there is a regularly-spaced array of one β-gal positive cell per cluster ( Fig 2E and 2F ) . Such symmetrical clusters are considered the strongest PCP defect as they completely fail to resolve the R3/R4 fates , indicating an inability to establish PCP-mediated cell fate differences at all [20 , 22 , 33] . Interestingly , symmetrical clusters were also observed in nmo , fz double mutant clones [34] , but this was neither commented upon nor followed up on in that study . To further define the genetic interaction between pksple1 and nmoP , we analyzed the PCP phenotype in the tarsal region of the leg . pksple1 mutants display supernumerary tarsal joints with altered bristle polarity ( Fig 2G , 2I and 2N ) [15] . This results in spiny-looking legs , giving the allele its name . Compared to pksple1 single mutants , joint number significantly increased in pksple1; nmoP double mutants ( Fig 2I , 2J and 2N ) . As the only isoform expressed in pksple1 adults is Pk , we confirmed the genetic interaction in a Pk GOF assay . Consistent with the notion that eye patterning is very sensitive to Pk isoform levels , even the low level overexpression of Pk ( via direct act-Pk , without Gal4-associated amplification ) is sufficient to unsettle the balance between Pk and Pk-Sple and cause defects: act-EGFP-Pk animals [35] displayed predominantly ‘flips’ in the eye ( Fig 2M and Suppl . S2G Fig ) . The chirality defects in act-EGFP-Pk animals were dominantly enhanced by nmoDB/+ ( Fig 2M and Suppl . S2G and S2H Fig ) . As act-EGFP-Pk is expressed at low levels throughout the animal , we also analyzed the leg phenotype . Consistent with the eye results , loss of nmo function enhanced the ectopic joint phenotype associated with act-EGFP-Pk ( Fig 2K , 2L and 2N; note that like the eye the PCP leg patterning is also sensitive to the Pk/Pk-Sple balance ) . In contrast , we did not observe PCP defects in wings of pksple1; nmoP mutants or act-EGFP-Pk and act-EGFP-Pk/nmoDB animals; all wings displayed a wild-type appearance ( Suppl . S3 Fig , note Pk is the major isoform in the wing ) . Examining whether there was an interaction between pkpk1 ( where Pk-Sple is the isoform expressed ) and nmo in the eye , we detected no chirality defects ( Suppl . S2C , S2D and S2I Fig , note that pkpk1; nmo double mutant ommatidia displayed only the expected nmo rotation phenotypes ) . Moreover , in GOF scenarios with overexpressed Pk-Sple , no interaction between nmo and Pk-Sple was observed in either tissue studied ( Suppl . S4 Fig ) , although act-EGFP-Sple wings did have a strong PCP phenotype [17 , 35] ( Suppl . S4C–S4E Fig ) . Importantly , in a pk null background ( pkpk-sple13 , with neither Pk or Pk-Sple isoforms expressed ) , we did not detect an increase in symmetrical cluster formation in pk-; nmo double mutant ommatidia ( Suppl . S2E , S2F and S2I Fig ) . The double mutant phenotype resembles that of pk null eyes , further reinforcing the notion that Nmo acts specifically on the Pk isoform , and thus if both isoforms are absent ( as with the null allele ) Nmo has no PCP substrate upon which to act . Collectively , the above data suggest that Nmo phosphorylates the Pk isoform of the pk gene , which has functional consequences in eyes and legs; tissues where both isoforms are expressed and Pk-Sple is the major protein isoform . Given the importance of the Pk/Pk-Sple balance , Nmo might be required to repress the Pk isoform in these tissues . The genetic interaction between pksple1 and nmoP being similar to that between Pk over-expression and nmo LOF ( Fig 2M and 2N ) suggests that the phenotype of the pksple1; nmoP double mutants results from increased Pk activity and/or amount . We therefore proceeded to examine the effect of Nmo phosphorylation on Pk activity . The sequence of the Nmo target fragment of Pk , region M ( Fig 1B and 1C ) , contains two clusters of 4 potential MAPK phosphorylation sites ( Fig 3A; Nmo is a member of the MAPK family ) . All 4 serine/threonine residues in each cluster were mutated to alanine to create phospho-mutant cluster 1 , cluster 2 , or a construct with all 8 sites mutated ( mut1+2; Fig 3A ) . These fragments were tested in in vitro kinase assays , which revealed that mutations of either cluster alone had little effect; but when both clusters were mutated phosphorylation was markedly reduced ( Fig 3A ) . These data identify two clusters of MAPK consensus sites within the common M fragment of Pk as direct phosphorylation targets of Nmo . To investigate the effect of these Nmo phosphorylation sites in vivo , we generated transgenic flies of either wild-type myc-Pk , or “phospho-mutant” myc-Pk ( PkWT and PkMut1&2 , where both clusters of Nmo phosphorylation sites were mutated to alanine ) ( Fig 3A and Methods ) . Pk overexpression in both cells of the R3/R4 precursor pair during PCP signaling ( under sevenless-Gal4 control: sev>Pk; [28 , 36] ) produced a phenotype with rotation and chirality defects ( Fig 3B and 3D; chirality defects were mainly ‘flips’ , although some symmetrical clusters were observed ) . Comparing Gal4-driven expression of wild-type Pk and phospho-mutant Pk ( Pkmut1+2 ) the phospho-mutant displayed more severe phenotypes with an increase in chirality defects and particularly symmetrical clusters ( Fig 3B and 3D; both transgenic constructs were inserted in the same genomic att-site and are thus transcriptionally expressed at equal levels; Methods ) . The above data mimic the effect of nmo loss-of-function on wt-Pk ( Fig 2M ) and below , corroborating the notion that the phosphorylation event causes a post-transcriptional reduction in Pk activity/levels . We modulated levels of Nmo and assessed the effects in the sev-driven Pk GOF assay in the eye . We used either nmo loss-of-function alleles ( nmoP and nmoDB alleles: Fig 3F and 3G , hypomorphic and null , respectively ) or Nmo co-overexpression ( Fig 3H ) [24] . There was a dose-dependent effect of loss of nmo function on the sev>Pk phenotype . Chirality defects increased in nmo heterozygotes with an increased number of symmetrical clusters in particular ( Fig 3E–3G and 3I ) . Conversely , Nmo co-overexpression with Pk suppressed the sev>Pk phenotype , markedly reducing the number of symmetrical clusters ( Fig 3E , 3H and 3I ) . Consistently with this , increasing the levels of Pk in a pksple1 background causes a synergistic increase in symmetrical clusters , similarly to Nmo LOF ( Suppl . S5B and S5C Fig ) . These data are consistent with a hypothesis that Nmo function is required to limit Pk activity or levels . Consistent with our earlier results , we did not detect an effect of nmo LOF on the Pk-Sple isoform overexpression phenotype ( Suppl . S4C–S4I Fig ) or differences in the activity of wild-type Pk-Sple compared to to Pk-SpleMut1&2 ( sev-SpleWT vs sev-SpleMut1&2 , the equivalent mutations to PkMut1&2; Suppl . S5E and S5F Fig ) . Moreover , mutation of the Nmo sites in Pk-Sple , does not affect the ability of the Pk-Sple isoform to rescue the chirality defects present in the pk null mutant ( Suppl . S5G Fig ) . We did not observe a difference in the activity levels of wild-type and phospho-mutant Pk in the wing either ( nub>PkWT and >PkMut1&2; Suppl . S5D Fig ) . Collectively , these results support a model in which Nmo acts primarily on Pk to maintain a tissue-specific balance of Pk/Pk-Sple activity in the eye and leg , where both are expressed with Pk-Sple being the major isoform . Therefore , post-translational regulation of Pk acts in the eye , in addition to the transcriptional control of isoform expression previously described for the wing [17] . Although the Nmo phosphorylation sites are shared between all pk isoforms , we have no evidence to suggest that the Pk-Sple isoform is affected . The Pk and Pk-Sple isoforms are known to form different protein complexes [37] , potentially explaining this disparity . It remains possible that Nmo does also phosphorylate Pk-Sple , but that Pk-Sple phenotypes are not affected by Nmo , potentially because the Pk-Sple specific N-terminus interferes with phosphorylation and/or masks the biological read-out . Based on these results we would predict Nmo to be required in the polar R4 precursor to limit Pk activity , because the pk gene ( the Pk-Sple isoform ) is required in the polar cell to establish PCP complexes and direct proper cell fate [28] . Genetic mosaic analysis is highly useful to determine which cell of the R3/R4 precursor pair requires an individual core PCP gene [14 , 28 , 34 , 38 , 39] . As nmo mutants show frequent chirality defects only in a pksple1 background rather than wild-type , we performed mosaic analysis in the pksple1 genetic background . Specifically we induced clones of nmo- cells in pksple1 eyes and analyzed R3/R4 pairs that were bisected by the clonal boundary; one cell of the R3/R4 pair was nmo+ and the other nmo- ( Methods ) . Clones of nmo- mutant cells were marked by absence of pigment ( Fig 4A and 4B shows schematic and example image ) . If nmo were specifically required only in one cell of the pair , for instance the R4 precursor , then we would expect to see more clusters developing with wild-type chirality when nmo function is removed from the other cell , R3 , than from R4 . Strikingly , when nmo function was only removed from R4 , ommatidia displaying wild-type chirality were markedly reduced ( to 42% ) , very similar to the fully double mutant pksple1; nmoP eyes ( Fig 4C and 4D’ ) . In contrast , when only R3 was nmo- , the proportion of wild-type clusters was 68% , very similar to pksple1 single mutant eyes ( Fig 4C and 4D” ) . These data indicate that this pk-associated nmo function is specifically required in the R4 cell . This was confirmed when the chirality defect ratios were compared with the pksple1 single and pksple1; nmoP double mutants ( Fig 4C , 4D”’ and 4D‘”’ ) . Thus the mosaic analyses suggested that for chirality establishment and core PCP function , it is the presence of Nmo in R4 that is required . The combined mosaic analyses of nmo ( this work ) and our previous study of pk [28] demonstrate that Nmo is required in the same cell as the pk gene to reduce the activity/levels of Pk in order for functional PCP complexes to be established . Compare this result to the genetic requirement of nmo in ommatidial rotation , when it functions as an effector of PCP , and is required in all R-cells and even cone cells [24 , 40] . Together these results demonstrate a spatially- and temporally-distinct role for nmo in regulating Pk in core PCP complex establishment in R4 , and then subsequently acting as an effector downstream of the core PCP complexes throughout the ommatidial cluster to regulate ommatidial rotation . One possibility of how Nmo could limit Pk activity is to regulate the levels of the Pk isoform , thereby preventing excess Pk disrupting the Pk/Pk-Sple isoform balance . To examine nmo loss-of-function effects on Pk levels , we performed Western blots with larval eye disc lysates . We first compared the levels of wild-type Pk and PkMut1&2 that were expressed under actin-Gal4 control . Compared to wild-type myc-Pk , mutation of the Nmo phosphorylation sites resulted in an increased protein level , as quantified as signal ratio of myc:gamma tubulin antibodies ( Fig 5A and Suppl . S6D and S6H Fig ) . We then compared levels of act-EGFP-Pk in either wild-type or nmoDB/+ backgrounds ( Fig 5B and Suppl . S6B and S6I Fig ) . nmo loss-of-function resulted in an increase in EGFP-Pk levels ( Fig 5B ) , similarly to mutation of Nmo phosphorylation sites ( Fig 5A ) . As nmo is only expressed in a stripe posteriorly to the furrow [41] , and its effect on Pk is specifically required during R3/R4 specification , there is a sizeable amount of act-EGFP-Pk that is unaffected , explaining the subtle increase , which is nevertheless significant . Conversely and as a specificity control , we did not detect a similar change in EGFP-Sple levels in a nmoDB/+ background , confirming that nmo acts specifically on the Pk isoform ( Fig 5B and Suppl . S6A–S6C Fig ) . These data are consistent with Nmo functioning to maintain lower levels of the Pk isoform , thus limiting the Pk/Pk-Sple ratio . Given the increase in PkMut1&2 , which is expressed from a transgene lacking endogenous 3’ UTR , under control of Gal4/UAS system , this also indicated that Nmo regulates Pk at the post-translational level . Degradation by the proteasome could be a means to regulate Pk levels . We examined this hypothesis by first co-expressing a dominant negative ( DN ) form of the proteasome 20S β2 subunit , Prosbeta2 [42–44] along with Pk under sev-Gal4 control and analyzing the adult phenotype . DNProsbeta2 expression synergized with the Pk GOF phenotype to enhance chirality defects , particularly symmetrical clusters ( Fig 5C , 5D and 5G ) . In a complementary approach , we used the milder act-EGFP-Pk phenotype and examined the effect of another DN proteasome component—this time β6 ( Prosbeta6 ) [42–44] , under the control of GMR-Gal4 , which is expressed in all post-mitotic , differentiating cells in the eye . We again saw an increase in chirality defects , and of symmetrical clusters in particular ( Fig 5E–5G ) . Moreover in this scenario , we also saw an increase in act-EGFP-Pk protein levels ( Fig 5H and Suppl . S6E Fig ) . In both cases , the control animals , with mildly reduced proteasome function alone , did not induce chirality defects ( Suppl . S6F and S6G Fig ) . It has been suggested that Pk levels are constitutively regulated by the Cullin1/ SkpA/Supernumary limbs ( Slmb ) SCF E3-ubiquitin ligase complex in Drosophila wings [45 , 46] . Consequently , we reasoned that the SCF complex might be operating in the eye to regulate Pk levels through ubiquitination and promoting subsequent proteasomal degradation of the Pk isoform . To test this hypothesis , we reduced the activity of the Cul1/SkpA/Slmb complex in the eye in the Pk GOF assay . We used RNAi to knockdown components of the complex temporally during establishment of PCP in the eye . Co-expressing the respective RNAi constructs , enhanced the Pk gain-of-function effects ( Fig 6A–6E and Suppl . S7D Fig; in control animals , the SCF LOF alone did not cause chirality defects; Suppl . S7A–S7C Fig ) . As with nmo LOF alleles , knockdown of each SCF component caused an increase in chirality defects , and symmetrical clusters in particular , in the Pk GOF scenario . Knockdown of slmb also caused a severe loss of photoreceptors , as well as many symmetrical clusters ( Suppl . S7D Fig ) and therefore the effect was confirmed by analyzing sev>Pk in a slmb00295/+ background ( Fig 6D ) . For SkpAIR , although there was a reproducible , mild enhancement of the phenotype , there was also loss of photoreceptors and of tissue integrity . Repeating the experiment at higher temperature to increase the knockdown , resulted in lethality ( sev-Gal4 includes the sev enhancer coupled to a heat shock promoter ) , so we were limited in terms of the temperature range in which we could work . We next examined the effect of reduction in SCF complex function on Pk protein levels . We examined act-EGFP-Pk levels in a slmb00295/+ background . Similarly to nmoDB/+ and GMR> DNProsbeta6 , we saw an increase in EGFP-Pk protein levels in the slmb LOF background ( Fig 6F and Suppl . S7E and S7F Fig ) . Taken all together , our data suggest that the increase in chirality defects and symmetrical clusters in the Nmo-Pk phosphorylation context is a result of an altered Pk/Pk-Sple isoform balance , which is caused by reduced Cul1-SkpA-Slmb-mediated proteasomal targeting of Pk . Previous studies have linked Nmo phosphorylation of a substrate to ubiquitination by the SCF complex and the proteasome; in mammalian cells , phosphorylation by Nemo-like kinase ( NLK ) acting downstream of TGF-β activated kinase ( TAK ) induces ubiquitination and proteasomal degradation of c-myb [47 , 48] . To investigate whether a dTAK-Nmo-protein degradation link is conserved and acts in the Nmo-Pk and PCP-signaling context , we tested for potential effects of the dTAK179 allele [49] on the sev>Pk phenotype . The chirality defects induced by sev>Pk were indeed enhanced in dTAK179/+ heterozygous females ( Fig 7A and 7B ) , suggesting that dTAK functions upstream of Nmo to limit Pk activity during establishment of PCP . In the TAK1-NLK-SCF complex axis , phosphorylation by Homeodomain Interacting Protein Kinase 2 ( HIPK2 ) also occurs and promotes substrate degradation [47 , 48] . We investigated whether the Drosophila Hipk homologue was also involved in regulating Pk . Scanning the Pk sequence for potential Hipk consensus sites [50] we detected a putative site in the C-terminus of the protein ( Fig 7G ) . Furthermore , in the Pk GOF assay we noted that knockdown of Hipk enhanced the PCP phenotype , similarly to knockdown of Nmo ( Fig 7C–7F ) . Together these data suggest that Nmo acts with dTAK and Hipk to phosphorylate Pk and recruit the SCF complex , promoting proteasomal degradation of Pk to maintain the Pk/Pk-Sple balance . In our systematic , genome-wide screen to identify Nmo substrates during its role in ommatidial rotation , we identified Pk as an unexpected bona fide target . Our functional studies then established a role for Nmo kinase during PCP establishment in addition to its known role during the subsequent rotation process . The phosphorylation of the Pk isoform serves as a way to limit the activity of the minor isoform in tissues where Pk-Sple is the major functional isoform . It was previously documented that the isoform balance is regulated at the transcriptional level within the wing , but it was unclear how Pk-Sple was able to act as the ‘major’ isoform in eyes , given that Pk mRNA was even expressed at higher levels [17] . Here we define a novel cell-specific requirement for post-translational regulation of Pk through phosphorylation and associated proteasomal targeting within the R4 cell . Our results point to a new paradigm of PCP modulation in which spatially-dependent regulation of core PCP protein degradation is required for robust PCP-cell fate coupling . Although it has been well documented that the balance of Pk isoforms is important [15 , 17] , it has remained unclear as to how such a balance would be maintained and/or reinforced post-translationally . Importantly , the Pk and Pk-Sple isoforms can form different protein complexes and localize to different sites within wing cells [37] . In particular , the correct presence of either Pk or Pk-Sple appears critical in the context of coupling the orientations of the core PCP complex alignments with the Fat/Ds-system polarity orientation [17 , 37 , 51] . While in the eye Fz-core PCP and Fat/Ds orientation is anti-parallel , the two systems are aligned in a parallel manner in the wing for example [51 , 52] . These opposing alignments correlate with differential requirements of either the Pk ( wing ) or Pk-Sple ( eye ) isoforms , and hence it is critical to maintain the correct levels of the individual isoforms . Our data suggest that Nmo-mediated phosphorylation of the Pk isoform participates in this context . Nmo phosphorylation of Pk is required in a tissue and cell-specific manner to maintain low levels of the Pk isoform and favor Pk-Sple in contexts where this is the ‘major’ isoform . Our mosaic analysis demonstrates that Nmo is required in the same photoreceptor cell as Pk-Sple and inhibits Pk function . It has been shown previously that the Cul1/SkpA/Slmb complex regulates overall Pk levels throughout the wing [45 , 46] , but in this case Pk is the major isoform and optimal Pk levels are required to prevent interference with core PCP complex function , particularly the internalization of transmembrane PCP components . The Cul1/SkpA/Slmb complex appears to play a maintenance role in the wing in preventing Pk hyperactivity throughout the tissue . Interestingly in this case , Pk-Sple accumulated in Cul1 LOF clones , similarly to Pk [45] . In contrast , our results show that the same machinery operates in a spatially restricted and isoform-specific manner in the eye . How might this specificity be achieved ? One possibility emerges from comparison with mammalian studies [47 , 48] . In this case , Wnt1 ligand acts upstream of TAK1-NLK-HIPK and subsequent SCF/proteasomal degradation of c-myb . In the case of the developing ommatidia , we observed a requirement for nmo in the R4 cell . This is the polar cell of the R3/R4 pair and the one that is closer to the Wingless/dWnt4 ligand sources at the dorsal and ventral poles of the imaginal disc [13 , 53] . This raises the possibility that Wingless/dWnt4 are acting upstream of dTAK and Nmo in regulating Pk levels . The involvement of Hipk may be complicated by its pleiotropic roles in regulating Wingless and Hedgehog signaling during eye development , in part through phosphorylation of Slmb itself [54] . This molecular circuitry governing Pk isoform degradation adds another layer of regulation to the intricate feedback mechanisms within a cell to increase robustness of position-based cell fate decisions and helps to coordinate tissue-wide patterning events . Our previous work demonstrated that Vang recruits Nmo to PCP complexes [24] . Based on our current study , this raises the possibility that when Pk , the minor isoform , is recruited into PCP complexes instead of Pk-Sple , Nmo serves a ‘gatekeeping’ role and phosphorylates Pk , targeting it for degradation . The importance of appropriate levels of Pk ubiqutination and degradation is underlined by a recent study that identified USP9X , a de-ubiqutinase , as a regulator of PRICKLE-mediated seizures in mammals [55] . In zebrafish , nemo-like kinase ( nlk ) genetically interacts with non-canonical wnt11 during convergent extension [29] , suggesting that Nemo/Nlk regulation of Prickle in PCP patterning processes is conserved . Together , these studies and our work highlight the importance of regulating the balance of Prickle family proteins during embryonic development and adult homeostasis to prevent disease .
The genetic tools and alleles used in this study are listed here along with their Flybase ID ( www . flybase . org ) : w1118 ( FBal0018186 ) ; pkpk1 ( FBal0013838 ) ; pksple1 ( FBal0016024 ) ; pkpk-sple13 ( FBal0060943 ) ; nmoP ( FBti0003251 ) ; eyFLP ( FBti0015982 ) ; pw+ , FRT80B ( FBst0001940 ) ; dTAK179 ( FBst0026275 ) ; UbxFLP ( FBti0150346 ) ; actGal4 ( FBst0003954 ) ; white RNAi ( FBst0031088 ) Cul1 RNAi ( FBst0029520 ) ; SkpA RNAi ( FBst0028974 ) ; slmb RNAi ( FBst0031056 ) ; slmb00295 ( FBst0011493 ) ; nmo RNAi ( FBst0025793 ) ; Hipk RNAi ( FBst0035363 ) . act-EGFP-Pk and act-FRT-STOP-FRT-EGFP-Sple flies , nmoDB , mδ-lacZ , and GMR>DNProsbeta6 ( also termed GMR>Dts5 ) flies were generous gifts from David Strutt ( University of Sheffield , UK ) , Esther Verheyen ( Simon Fraser University , Canada ) , Sarah Bray ( University of Cambridge , UK ) , and Hermann Steller ( The Rockefeller University , USA ) , respectively . sevGal4 , UAS-Pk; sevGal4 , UAS-Sple [14 , 28]; sev-Sple-WT [14 , 28]; UAS-Nmo , [24] , nubGal4 [25] . UAS-DNProsβ2 and UAS-DNProsβ6 ( also called Dts5 ) [42 , 56] flies were as described ( see refs above ) . pUAS attB_myc-Pk-Wt and pUAS attB_myc-Pk-mut1&2 were used to generate transgenic flies in the y1 w1118; PBac{y+-attP-9A}VK00027 ( FBst0009744 ) background . Using the specific AttP integration site , we ensured that each transgene was expressed from the same genomic locus and thus to a comparable level . Several independent transgenic insertions were tested to confirm comparable phenotypes . Genotypes for all figures are as follows: Tangential eye sections were prepared as described [57] . Three to ten independent eyes were analyzed per genotype , with over 300 ommatidia scored in each case , except for the mosaic analysis ( see below ) , or as noted in figure legends . Only ommatidia with a full complement of photoreceptors were scored for chirality defects . Rotation defects were not analyzed . χ2 or Fisher’s exact test were performed on adult eye data based on the number in each category , depending on which criteria were met regarding sample size and values of each category . Data in graphs are shown as a percentage for clarity . Animals of the genotype eyFLP; pksple1; nmoDBFRT80B/pw+FRT80B were analyzed . The white pigment marks all heterozygous cells and those homozygous for pw+FRT80B following eyFLP-mediated recombination . Cells lacking white pigment were therefore homozygous for the nmoDB allele . Tangential eye sections were carefully studied to identify R3/R4 pairs in which one cell lacked pigment ( nmo- mutant ) and the other produced pigment and therefore had at least one functioning copy of nmo ( nmo+/+ or nmo+/- ) . Pairs were scored for the number of times the R3/R4 fate decision was correctly resolved ( WT ) or where PCP defects occurred ( flipped or symmetrical clusters ) , when either R3 had Nmo function or when Nmo function was only present in R4 . Over 140 R3/R4 pairs were analyzed from 12 individual adults . Mosaic eye data were analyzed by Fisher’s exact test . For comparison to the pksple1 and pksple1; nmoP mutants , the data from Fig 1 were shown with flips and symmetrical clusters combined into one PCP defect category . Adult legs were dissected and incubated overnight in 70% ethanol . Legs were then washed three times in PBS with 0 . 1% Triton-X100 and once in PBS . Legs were mounted in 80% glycerol in PBS . The Mann-Whitney U test was used to compare number of tarsal joints . Over 54 legs were analyzed per genotype . Wings were dissected and incubated in PBS with 0 . 1% Triton-X100 for at least an hour before being mounted in 80% glycerol in PBS . Immunofluorescence was performed on third instar larval eye discs and imaged as described [58] , using the rabbit anti-beta Galactosidase antibody ( Immunology Consultants Laboratory , Inc . 1:200 ) and mouse anti-Elav antibody ( Developmental Studies Hybridoma Bank ( DSHB ) 1:50 ) . S2 cells were transfected using Qiagen effectene reagent according to manufacturer’s instructions . Cells expressing Myc-Nmo and HA-Pk were lysed and treated in phosphatase assay as described [58] . Eye discs from third instar larvae were dissected and collected in PBS on ice ( 10 μl per pair of eye discs ) . An appropriate volume of 5X laemmli sample buffer was added and samples were boiled for 10 minutes . 5–10 disc pairs were loaded per lane of a standard SDS-PAGE gel . Membranes were incubated with anti-myc ( Mouse , 1:1000 Santa Cruz Biotechnology ) , anti-γ-tubulin ( Mouse , 1:1000 , Sigma ) , anti-GFP antibody ( Mouse , 1:1000 Roche ) , and anti-Arm ( Mouse , 1:20 DSHB ) antibody ( DSHB ) . Secondary antibodies and signal visualization were performed using a ChemiDoc MP imager ( BioRad , Hercules , CA ) as described [58] . The intensity of the EGFP-Pk band was quantified as a signal ratio of myc:γ-tubulin , GFP:γ-tubulin , or GFP:Armadillo antibodies ( Arm , Drosophila β-catenin , whose levels are unaffected by Nmo [24] ) . Intensities were measured in ImageJ . Background intensity was subtracted and then the ratio of signal from GFP:Arm calculated . A paired t-test was used to compare the ratio intensities from four independent experiments . All PCR products were verified by sequencing or by replacing internal fragments with versions from cDNA clones . Gst-Stbm/Vang-Cterm , Gst-Pk-Cterm , Gst-PkΔSacI ( N-Δ1 in Fig 1B ) , Gst-PkΔNruI ( N-Δ2 in Fig 1B ) , pAct_mod_Cterm were described in [28] . pGexKG-EcoCterm ( C1 in Fig 1B ) was made by cloning the EcoRI fragment of pAct_mod_Cterm [28] into the EcoRI site of pGExKG . pGexKG_Suf ( C2 in Fig 1B ) was made by re-circularizing SacI digested pGexKG-EcoCterm , thus deleting the C-terminal SacI fragment . pGex4T1_common was cloned by transferring the BamHI fragment of pTopo4 . 0_Common [28] into the BamHI site of pGex4T1 . Topo4 . 0_Sple_C encoding the PkSple specific N-terminal extension contains a PCR product amplified with Sple_C_upper/lower ( Sple_C_upper TATGGATCCATGAGCAGCCTGTCAACC , Sple_C_lower ATAGAATTCTCACTCATTTGACTCCTGCTGG ) . Its insert was then cloned as BamHI/ EcoRI fragment into the corresponding sites of pGex4T1 to give pGex4T1_Sple_C . pBSIISKP motifs ( Dom in Fig 1B ) contains a PCR product amplified with motifs_upper/lower ( motifs_upper TATGGATCCGGCGGACCGCACATGG motifs_lower ATACTCGAGTCACCCCTTGCTGCAGGCG ) and encodes the PET and LIM domains of Pk . Its insert was transferred as BamHI/ XhoI fragment into pGex4T1 to give pGex4T1_motifs . pCR-Topo2 . 1_PET corresponds to the PET domain of Pk that was amplified by PCR using motifs_upper and Pet_lower_XhoI ( motifs_upper TATGGATCCGGCGGACCGCACATGG , Pet_lower_XhoI ATACTCGAGTCATCGGGCGCTCATCAGCTG ) . Its insert was transferred as BamHI/ XhoI fragment into pGex4T1 to give pGex4T1_PET . pQE31_PkBE ( M in Fig 1B ) was cloned by inserting a BssHII ( blunt ) / EcoRI ( blunt ) fragment of Pk into the SmaI site of pQE31 . pGex4T1_PkBE was cloned as BamHI/ EcoRI fragment that was amplified with PkBssRI_WT_for_Bam and T3XL ( PkBssRI_WT_for_Bam TATGGATCCCTGCCGGCGCGCATTCCC , T3XL CGAAATTAACCCTCACTAAAGGGA ) into the corresponding sites of pGex4T1 . The eight candidate Nmo sites ( Ser/Thr followed by Pro ) were mutated to Ala as follows: first , using pBS2SKP_Sple_cloneable [28] as template , ‘mega primers’ were amplified with primers Pk_B/Emut_for_3_BamHI and PK_BE_Mut rev 1 , and Pk_BE_Mut for2 and PK_BE_Mut rev 2 , respectively . These were used as primers for amplification with PK_BE_Mut rev 2 , and Pk_B/Emut_for_3_BamHI , respectively ( Pk_B/Emut_for_3_BamHI TATGGATCCCTGCCGGCGCGCATTCCCAGCAGCCACGCCTCCAGCGCACCGCCCATGGCACCGCAACAGCAGCAGCAG , PK_BE_Mut rev 1 CTGGAAGTCGCCGGGTGCGTTCAGAGGCGCTAGGTTCTGCGAGGT , Pk_BE_Mut for2 GCCCGCTCCCAACTTGAGCGTGGCTTCCACCGCCTTGCCGCCAGAGCTTATGGGCGCCCCCACCCACTCGGCGGGCGACAGGTCGCTGAACGCGCCCATG , PK_BE_Mut rev 2 GTCCGGAATTCCCTCGAAGCGCACGCCCTTCTTCTTGGCCGGCTCCCCGCTCATCGGCGCGGAGGAGGA ) . The final fragment used further corresponds to a BssHII/ EcoRI fragment of Pk with an added upstream BamHI site in pBS2SKP ( after correcting PCR mistakes ) . Note that the 8 mutations roughly cluster in two regions that are partitioned in two groups of 4 by a SphI site ( the first and second cluster are marked by a NcoI or a NarI site , respectively ) . The BamHI/ EcoRI insert of pBS2SKP_PkBE_mut was inserted into the corresponding site of pGex4T1 to give Gex4T1_PkBE_mut . Plasmids encoding Gst-Pk versions with one mutant cluster only each were generated by replacement of the other cluster with wild-type SphI/ EcoRI and BamHI/ SphI fragments from pGex4T1_PkBE , respectively , to give Gex4T1_PkBE_mut-first and Gex4T1_PkBE_mut-second . pFastBacHisC_Nmo was cloned by inserting a BamHI/ PstI fragment of pEGFPN3-Nmo into the appropriate sites of pFastBacHisC . Baulovirus recombination and protein expression and purification were done as described [25] . Construction , expression , and purification of RokCat was described in [25] and recombinant Mst1/Mst2 were from Invitrogen [26] . Gst and His tag fusion proteins were prepared as in [14 , 25 , 28] . [32]P kinase assays were done using 1μg of the indicated Gst protein in 20μl reactions using 0 . 5μl Nmo ( about 350 ng ) as described [25] . Gel shift kinase assays were done as in [25] using 2μl of [35]S labeled in vitro translated proteins ( translated from pOT-108-pan ( a kind gift of K . Basler , University of Zuerich ) , pβTH_common [28] and RE63915 ( mbs ) . pUAS attB_myc-Pk-Wt and pUAS attB_myc-Pk-mut were cloned as follows: Pk EcoRINtermMycATGF and PKRA 1232R XhoI primers were used to amplify the 5′ region and to add a myc tag ( Pk EcoRINterm Myc ATG F AACGCACCATGGAACAAAAACTTATTTCTGAAGAAGATCTGATGGATACCCCAAATCAAATGCC , PKRA 1232 R XhoI CCGCTCGAGAAAGCCGGCGATAGCTGGTG ) . The 3′ region was digested using SalI and NotI from either pBS2SKP_PkBE ( WT ) or pBS2SKP_PkBE_mut to generate myc-Pk-WT or myc-Pk-mut , respectively . These fragments were ligated into pUAS attB . The generation of sev-PkSple flies was as described in [28] . The construct corresponds to the Sple cDNAs [15] cloned as blunted DraI/AseI fragment into the blunted EcoRI site of pKB267PL ( modified after [59] ) . For sev-SpleMut , the Sple-BssHII-EcoRI fragment of pBS2SKP_PkBE_mut ( see above ) was cloned into the corresponding BssHII ( partial digest ) / EcoRI sites of pBS2SKP_Sple_cloneable [28] . From that construct , a AgeI/EcoRI fragment was cloned into the corresponding sites of sev-PkSple . | For functional tissues to form , individual cells must correctly orient themselves and function appropriately for their particular location in the body . The Planar Cell Polarity ( PCP ) complexes transmit one set of spatial cues by acting as signposts to mark direction across an epithelial layer . PCP signals can direct and coordinate cell differentiation , the behavior of groups of cells , or the orientation of individual cellular protrusions , depending on the tissue . PCP signals act as a polarization relay with two different complexes being positioned on opposite sides of each cell . This pattern of polarity is transmitted to neighboring cells and so extends across the tissue . In the fly eye , PCP signals control the differentiation of a pair of photoreceptors , R3 and R4 , where the cell that is positioned closer to the dorso-ventral midline becomes R3 . An excess of the PCP protein Prickle prevents the proper assembly of PCP complexes in the eye and so alters R3/R4 fate . Here we show that Nemo kinase is required in the R4 cell to phosphorylate Prickle and promote its degradation by the proteasome . Maintenance of low Prickle levels allows proper formation of PCP complexes , cell fate specification , and eye development . | [
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
]
| [
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"social",
"sciences",
"cloning",
"neuroscience",
"pigments",
"materials",
"science",
"molecular",
"biology",
"techniques",
"epigenetics",
"eyes",
"research",
"and",
"analysis",
"methods",
"animal",
"cells",
"proteins",
"genetic",
"interference",
"gene",
"expression",
"materials",
"by",
"attribute",
"sensory",
"receptors",
"head",
"molecular",
"biology",
"proteasomes",
"biochemistry",
"signal",
"transduction",
"rna",
"cellular",
"neuroscience",
"psychology",
"protein",
"complexes",
"anatomy",
"post-translational",
"modification",
"cell",
"biology",
"phenotypes",
"nucleic",
"acids",
"neurons",
"genetics",
"photoreceptors",
"biology",
"and",
"life",
"sciences",
"ocular",
"system",
"cellular",
"types",
"afferent",
"neurons",
"sensory",
"perception",
"physical",
"sciences"
]
| 2018 | Prickle is phosphorylated by Nemo and targeted for degradation to maintain Prickle/Spiny-legs isoform balance during planar cell polarity establishment |
Copy number variants ( CNVs ) are a pervasive source of genetic variation and evolutionary potential , but the dynamics and diversity of CNVs within evolving populations remain unclear . Long-term evolution experiments in chemostats provide an ideal system for studying the molecular processes underlying CNV formation and the temporal dynamics with which they are generated , selected , and maintained . Here , we developed a fluorescent CNV reporter to detect de novo gene amplifications and deletions in individual cells . We used the CNV reporter in Saccharomyces cerevisiae to study CNV formation at the GAP1 locus , which encodes the general amino acid permease , in different nutrient-limited chemostat conditions . We find that under strong selection , GAP1 CNVs are repeatedly generated and selected during the early stages of adaptive evolution , resulting in predictable dynamics . Molecular characterization of CNV-containing lineages shows that the CNV reporter detects different classes of CNVs , including aneuploidies , nonreciprocal translocations , tandem duplications , and complex CNVs . Despite GAP1’s proximity to repeat sequences that facilitate intrachromosomal recombination , breakpoint analysis revealed that short inverted repeat sequences mediate formation of at least 50% of GAP1 CNVs . Inverted repeat sequences are also found at breakpoints at the DUR3 locus , where CNVs are selected in urea-limited chemostats . Analysis of 28 CNV breakpoints indicates that inverted repeats are typically 8 nucleotides in length and separated by 40 bases . The features of these CNVs are consistent with origin-dependent inverted-repeat amplification ( ODIRA ) , suggesting that replication-based mechanisms of CNV formation may be a common source of gene amplification . We combined the CNV reporter with barcode lineage tracking and found that 102–104 independent CNV-containing lineages initially compete within populations , resulting in extreme clonal interference . However , only a small number ( 18–21 ) of CNV lineages ever constitute more than 1% of the CNV subpopulation , and as selection progresses , the diversity of CNV lineages declines . Our study introduces a novel means of studying CNVs in heterogeneous cell populations and provides insight into their dynamics , diversity , and formation mechanisms in the context of adaptive evolution .
Copy number variants ( CNVs ) drive rapid adaptive evolution in diverse scenarios ranging from niche specialization to speciation and tumor evolution [1–4] . CNVs , which include duplications and deletions of genomic segments , underlie phenotypic diversity in natural populations [5–10] and provide a substrate for evolutionary novelty through modification of existing heritable material [11–14] . Beneficial CNVs are associated with defense against disease in plants , increased nutrient transport in microbes , and drug-resistant phenotypes in parasites and viruses [9 , 15–18] . Despite the importance of CNVs for phenotypic variation , evolution , and disease , the dynamics with which these alleles are generated and selected in evolving populations are not well understood . Long-term experimental evolution provides an efficient means of gaining insights into evolutionary processes using controlled and replicated selective conditions [19 , 20] . Chemostats are devices that maintain cells in a constant nutrient-poor growth state using continuous culturing [21] . Nutrient limitation in chemostats provides a defined and strong selective pressure in which CNVs have been repeatedly identified as major drivers of adaptation . CNVs containing the gene responsible for transporting the limiting nutrient are repeatedly selected in a variety of organisms and conditions including Escherichia coli limited for lactose [22] , Salmonella typhimurium in different carbon source limitations [23] , and Saccharomyces cerevisiae in glucose- , phosphate- , sulfur- , and nitrogen-limited chemostats [24–30] . CNVs confer large selective advantages , and multiple , independent CNV alleles have been identified within experimental evolution populations [25–27 , 31] . These findings suggest that CNVs are generated at a high rate , but estimates differ greatly , ranging from 1 × 10−10 to 3 . 4 × 10−6 duplications per cell per division , with variation in CNV formation rates potentially differing between loci and/or condition [32 , 33] . A high rate of CNV formation suggests that multiple , independent CNV-containing lineages may compete during adaptive evolution , resulting in clonal interference , which is characteristic of large , evolving populations [29 , 34–36] . However , the extent to which clonal interference among CNV-containing lineages influences the dynamics of adaptation is unknown . The general amino acid permease gene , GAP1 , is well suited to studying the role of CNVs in adaptive evolution . GAP1 encodes a high-affinity transporter for all naturally occurring amino acids , and it is highly expressed in nitrogen-poor conditions [37 , 38] . We have previously shown that two classes of CNVs are selected at the GAP1 locus in S . cerevisiae when a sole nitrogen source is provided: GAP1 amplification alleles are selected in glutamine and glutamate-limited chemostats , and GAP1 deletion alleles are selected in urea- and allantoin-limited chemostats [24 , 25] . GAP1 CNVs are also found in natural populations . In the nectar yeast Metschnikowia reukaufii , multiple tandem copies of GAP1 result in a competitive advantage over other microbes when amino acids are scarce [39] . As a target of selection in adverse environments in both experimental and natural populations , GAP1 is a model locus for studying the dynamics and mechanisms underlying both gene amplification and deletion in evolving populations . CNVs are generated by two primary classes of mechanisms: homologous recombination and DNA replication [40–42] . DNA double-strand breaks ( DSBs ) are typically repaired by homologous recombination and do not result in CNV formation . However , nonallelic homologous recombination ( NAHR ) can generate CNVs when the incorrect repair template is used , which occurs more often with repetitive DNA sequences such as transposable elements and long terminal repeats ( LTRs ) [43] . During DNA replication , stalled and broken replication forks can reinitiate DNA replication through processes including break-induced replication ( BIR ) , microhomology-mediated break-induced replication ( MMBIR ) , and fork stalling and template switching ( FoSTes ) [44–46] . BIR is driven by homologous sequences , whereas MMBIR relies on shorter stretches of sequence homology . Recently , origin-dependent inverted-repeat amplification ( ODIRA ) has been identified as a novel mechanism underlying amplification of the SUL1 locus in yeast [47 , 48] . ODIRA is mediated by short inverted repeat sequences that facilitate ligation of the leading and lagging strands following regression of the replication fork during DNA synthesis . ODIRA is hypothesized to involve the formation of an extrachromosomal circular intermediate that replicates independently and therefore requires an origin of replication within the amplified region . Subsequent integration of the circle into the original locus via homologous recombination results in an inverted triplication . Extrachromosomal circular DNA is common in yeast [49] , can drive tumorigenesis [50] , and may represent a rapid and reversible mechanism of generating adaptive CNVs [51 , 52] . Previously , we found that some GAP1 amplifications are extrachromosomal circular elements . We hypothesized that GAP1circle alleles are generated as a result of NAHR between flanking LTRs , resulting in their excision from the chromosome [25] . Identifying the mechanisms underlying CNV formation is required for understanding the roles of CNVs in evolutionary processes and human disease . A key limitation to the study of CNVs in evolving populations is the challenge of identifying them at low frequencies in heterogeneous populations . CNVs are typically detected using molecular methods including quantitative PCR ( qPCR ) , Southern blotting , DNA microarrays , and sequencing [24–26] . However , using any of these methods , de novo CNVs are undetectable in a heterogeneous population until present at high frequency ( e . g . , >50% ) . This precludes analysis of the early dynamics with which CNVs emerge and compete in evolving populations . As CNVs usually comprise genomic regions that include multiple neighboring genes [24] , we hypothesized that CNVs could be identified on the basis of increased expression of a constitutively expressed fluorescent reporter gene inserted adjacent to a target gene of interest . A major benefit of this approach is that it detects CNVs independently of whole-genome sequencing , enabling a high-resolution and efficient assay of CNV dynamics with single-cell resolution in evolving populations . In this study , we constructed strains containing a fluorescent CNV reporter adjacent to GAP1 in S . cerevisiae and performed evolution experiments in different selective environments using chemostats . The CNV reporter allowed us to visualize selection of CNVs at the GAP1 locus in real time with unprecedented temporal resolution . We find that CNV dynamics occur in two distinct phases: CNVs are selected early during adaptive evolution and quickly rise to high frequencies , but the subsequent dynamics are complex . We find that GAP1 CNVs are diverse in size and copy number and can be generated by a range of processes including aneuploidy , nonreciprocal translocations , and tandem duplication by NAHR . Nucleotide resolution analysis of GAP1 CNV breakpoints revealed that CNV formation is mediated by short , interrupted inverted repeats for half of the resolvable cases , suggesting that replication-based mechanisms also underlie gene amplification at the GAP1 locus . The presence of inverted repeats , in combination with a replication origin and inverted triplication , is consistent with GAP1 CNV formation through ODIRA . ODIRA may be a major source of de novo CNVs in yeast , as these breakpoint features also characterize CNVs at an additional locus identified in our study , DUR3 . To determine the underlying structure of the CNV subpopulation , we generated a lineage-tracking library using random DNA barcodes . Fluorescence-activated cell sorting ( FACS ) -based fractionation of CNV lineages and barcode sequencing identified hundreds to thousands of individual CNV lineages within populations , consistent with a high CNV supply rate and extreme clonal interference . Together , our results show that CNVs are generated repeatedly by diverse processes , resulting in predictable dynamics , but that the long-term fate of CNV-containing lineages in evolving populations is shaped by clonal interference and additional variation .
We sought to construct a reporter for CNVs that occur at a given locus of interest . Based on previous studies [53–56] , we hypothesized that CNVs that alter the number of copies of a constitutively expressed fluorescent protein gene would facilitate single-cell detection of de novo copy number variation . To test the feasibility of this approach , we constructed haploid S . cerevisiae strains isogenic to the reference strain ( S288c ) with one or two copies of a constitutively expressed green fluorescent protein ( GFP ) variant mCitrine [57] and diploid strains with 1–4 copies of mCitrine integrated into the genome ( S1 Table ) . Flow cytometry analysis confirmed that additional copies of mCitrine produce quantitatively distinct distributions of protein fluorescence ( Fig 1A ) . Haploid cells with two copies of mCitrine have higher fluorescence than those with a single copy , and there is minimal overlap between the distributions of fluorescent signal in the two strains . Normalization of the fluorescent signal by forward scatter , which is correlated with cell size , shows that the concentration of fluorescent protein is proportional to the ploidy normalized copy number of the mCitrine gene ( i . e . , one copy in a haploid results in a signal equivalent to two copies in a diploid , and two copies in a haploid results in a signal similar to four copies in a diploid ) . Thus , the cell size–normalized fluorescent signal , or concentration , accurately reports on the number of copies of the fluorescent gene in single cells . Therefore , integrating a constitutively expressed fluorescent protein gene proximate to an anticipated target of selection functions as a CNV reporter for tracking gene amplifications and deletions in evolving populations ( Fig 1B ) . Previous work has shown that spontaneous GAP1 amplifications are positively selected when glutamine is the sole limiting nitrogen source during evolution experiments in chemostats [25] . GAP1 copy number amplifications result in increased amino acid transporters on the plasma membrane , providing cells with a selective advantage when nitrogen is scarce [24 , 25] . Conversely , GAP1 deletions provide a fitness benefit and are selected in urea-limited conditions [25] , which may be due to two nonexclusive reasons: either ( 1 ) because GAP1 is highly expressed regardless of the type of limiting nitrogen source [58] but unable to transport urea , it confers a gene expression burden; or ( 2 ) when the extracellular concentration of amino acids is low compared to the intracellular concentration , the electrochemical gradient drives their export through the GAP1 permease . Thus , the use of different nitrogen sources in nitrogen-limited chemostats enables the study of both GAP1 amplification and deletion , making it an ideal system for studying the dynamics of CNV selection in evolving populations . We constructed a haploid strain containing a mCitrine CNV reporter located 1 , 118 bases upstream of the GAP1 start codon to ensure that the native regulation of GAP1 was unaffected [59] . We inoculated the GAP1 CNV reporter strain into 9 glutamine- , 9 urea- , and 8 glucose-limited chemostats for a total of 26 populations ( S2 Table ) . For each of the three selection conditions , we included two control populations: one containing a single copy of the mCitrine CNV reporter at a neutral locus ( one copy control ) and one containing two copies of the mCitrine CNV reporter at two neutral loci ( two copy control ) . All populations were maintained in continuous mode ( dilution rate = 0 . 12 culture volumes/hour; population doubling time = 5 . 8 hours ) for 267 generations over 65 days . We sampled each of the 32 populations every 8 generations and used flow cytometry to measure fluorescence of 100 , 000 cells per sample . Experimental evolution in a glutamine-limited chemostat resulted in clear increases in fluorescence in individual cells containing the GAP1 CNV reporter by generation 79 ( Fig 2A ) . By contrast , populations containing one or two copies of mCitrine at neutral loci exhibited stable fluorescence for the duration of the experiment ( Fig 2A ) . Maintenance of protein fluorescence in one- and two-copy control populations is consistent with the absence of a detectable fitness cost associated with one or two copies of the CNV reporter in glutamine-limited chemostats , which we confirmed using competition assays ( S1 Fig ) . Analysis of eight additional independent populations evolving in glutamine-limited chemostats showed qualitatively similar dynamics of single-cell fluorescence over time ( S2 Fig ) . To summarize the dynamics of CNVs in evolving populations , we determined the median normalized fluorescence in each population at each time point . The fluorescent signal of the GAP1 CNV reporter increases during selection in all populations evolving in glutamine-limited chemostats ( Fig 2B ) , consistent with the de novo generation and selection of CNVs at the GAP1 locus in all 9 populations . Populations evolving in urea-limited and glucose-limited chemostats do not show substantial changes in fluorescence , with one exception ( Fig 2B ) . In a single urea-limited population ( ure_05 ) , we detected a complete loss of fluorescent signal by generation 125 , indicating the occurrence of a GAP1 deletion that subsequently swept to fixation . Thus , the GAP1 CNV reporter detects both amplification and deletion alleles at the GAP1 locus in evolving populations . The absence of increases or decreases in fluorescence in all glucose-limited populations is consistent with the absence of selection for GAP1 CNVs in conditions that are irrelevant for GAP1 function . To quantify the proportion of cells containing a GAP1 duplication , we used one- and two-copy control strains to define flow cytometry gates . We found that the fluorescence of control strains varied slightly ( S3A Fig ) , which may be indicative of either instrument variation or changes in cell physiology and morphology during the experiment , as suggested by systematic changes in forward scatter with time ( S3B Fig ) . Using a conservative method to classify individual cells containing GAP1 amplifications ( Methods ) , we find that GAP1 amplification alleles are selected with remarkably reproducible dynamics in the nine glutamine-limited populations ( Fig 2C ) . CNVs are predominantly duplications ( two copies ) , but quantification of fluorescence suggests that many cells contain three or more copies of the GAP1 locus ( S4 Fig ) . We quantified the dynamics of CNVs in each population evolved in glutamine-limited chemostats using metrics defined by Lang and colleagues [60] . CNVs are detected by generation 70–75 ( average = 72 . 8 ) in all 9 populations ( Tup ) ( Table 1 ) . To estimate the fitness of all CNV lineages relative to the mean population fitness , we calculated Sup , the rate of increase in the abundance of the CNV subpopulation ( see Methods and S1 Text ) . The average relative fitness of the CNV subpopulation is 1 . 077 ( Sup ) , and CNV alleles are at frequencies greater than 75% in all populations by 250 generations ( Table 1 ) . Thus , in all replicated glutamine-limited selection experiments , GAP1 amplifications emerge early , increase in frequency rapidly , and are maintained in each population throughout the selection . GAP1 CNVs undergo two distinct phases of population dynamics . The initial dynamics with which CNV subpopulations emerge and increase in frequency are highly reproducible in independent evolving populations . However , after 125 generations , the trajectories of the CNV subpopulation in the different replicate populations diverge . Many populations maintain a high frequency of GAP1 amplification alleles , but in some populations , they decrease in frequency . In one population , GAP1 CNV alleles are nearly lost from the population before subsequently increasing to an appreciable frequency ( gln_07 ) . Based on prior studies [24 , 26] , we hypothesized that multiple CNV alleles exist within each population . To characterize the diversity of GAP1 CNVs , we isolated a total of 29 clones containing increased fluorescence from glutamine-limited chemostats at 150 and 250 generations for whole-genome sequencing ( S3 Table ) . We used read depth to calculate GAP1 copy number and to estimate CNV boundaries ( Fig 3A , S4 Table , and Methods ) . We find that GAP1 copy number estimated by sequencing read depth correlates with the fluorescent signal for individual clones ( Fig 3B ) , indicating that fluorescent signal is predictive of copy number . In 3 clones , we find increased read depth across the entirety of Chromosome XI consistent with aneuploidy . Thus , the CNV reporter is able to detect aneuploid chromosomes as well as subchromosomal CNVs . We identified diverse GAP1 CNVs between and within populations ( Fig 3C ) . In the majority of populations ( 6/9 ) , different clones had different CNVs . For example , in population gln_01 at generation 150 , we identified a large GAP1 CNV that includes the entire right arm of Chromosome XI and another clone that was aneuploid for Chromosome XI . At generation 250 , clones isolated from population gln_01 have CNV alleles that are distinct from each other and from those observed at generation 150 . Clones from the 8 additional glutamine-limited populations show evidence for CNV diversity within and between the two time points analyzed ( Fig 3C ) , suggesting the presence of multiple CNV lineages within evolving populations . Furthermore , the diversity of GAP1 CNVs indicates that they are not predominantly formed through a recurrent mechanism as might be anticipated by the presence of proximate repetitive elements . We used pulsed-field gel electrophoresis and Southern blotting to confirm CNV structures ( S5 Fig ) . Using GAP1 and CEN11 probes for Southern blotting , we identified size shifts in some samples consistent with the large CNVs ( >140 kilobases ) we identified in several clones . In some cases , we identified two discrete bands in our GAP1 Southern blot , indicating that the additional copies of GAP1 were not contained on Chromosome XI . The GAP1 Southern also provided further evidence for the GAP1 deletion in a clone isolated from urea limitation . Whereas control populations evolving in glutamine-limited chemostats did not show evidence for GAP1 CNVs on the basis of fluorescence , sequence and Southern blotting analysis identified GAP1 amplifications in lineages isolated from these populations ( S2 Text and S5 Fig ) . As one- and two-copy control strains do not have the GAP1 CNV reporter , this suggests that GAP1 CNV formation and selection are not affected by the reporter . Moreover , we find no evidence that the molecular features of GAP1 CNVs are affected by the presence of the CNV reporter . We determined the fitness of GAP1 CNV-containing clones using pairwise competitive fitness assays in glutamine-limited chemostats ( S6 Fig and Fig 3C ) . Four independent competition assays with the ancestral strain containing the GAP1 CNV reporter showed no significant differences in fitness compared to the isogenic nonfluorescent reference strain . The majority of evolved clones ( 18/28 ) have higher relative fitness than the ancestor , indicating that GAP1 CNVs typically confer large fitness benefits . Several clones have neutral ( 8/28 ) or lower ( 2/28 ) relative fitness , which indicates that either ( 1 ) the fitness effect of GAP1 CNVs may be context specific or ( 2 ) not all GAP1 CNVs confer a fitness benefit . We analyzed the genome sequences of 21 clones that were randomly isolated from urea-limited populations at generation 150 and generation 250 and identified multiple CNVs at the DUR3 locus ( S7A Fig and S2 Text ) . DUR3 encodes a high-affinity urea transporter , and we have previously reported DUR3 amplifications during experimental evolution in a urea-limited chemostat [24] . We compared properties of GAP1 and DUR3 amplifications and found that the average copy number for clones with GAP1 CNVs is 3 ( S7B Fig ) , whereas clones with DUR3 CNVs contain significantly more copies , with an average copy number of 5 ( S7C Fig , t test , p-value < 0 . 01 ) . Copy number within clones does not significantly increase between 150 and 250 generations at either locus . DUR3 CNV alleles ( average of 26 kilobases ) are also significantly smaller than GAP1 CNVs ( average of 105 kilobases ) ( S7D–S7E Fig , t test , p-value < 0 . 01 ) . Thus , comparison of GAP1 and DUR3 CNVs suggests differences in the properties of selected CNVs as a function of locus and selective condition . To resolve CNV breakpoint sequences , we generated a pipeline integrating CNV calls from multiple existing CNV detection methods ( CNVnator , Pindel , LUMPY , and SvABA [61–64] ) and optimized their performance on synthetic yeast genome data ( S3 Text ) simulating both clonal ( S8 Fig ) and heterogeneous populations ( S9 Fig ) . Although these algorithms perform well using simulated data , we found that they had a high false positive and false negative rate when applied to real data ( S5 Table and S6 Table ) and , in general , were not informative about the novel sequence formed at CNV boundaries . Therefore , we developed a breakpoint detection pipeline that integrates information from read depth , discordant reads , and split reads . To define the breakpoint sequence , we performed de novo assembly using split reads and aligned the resulting contig against the reference genome ( Methods ) . In addition to GAP1 and DUR3 CNVs , we identified 3 structural variants in our clonal sequencing data using this method ( S7 Table ) . A read depth–based approach was also used to characterize CNVs genome-wide ( S8 Table ) and calculate ribosomal DNA ( rDNA ) and CUP1 copy number , which exhibit variation among lineages ( S4 Table ) . We analyzed 29 lineages containing GAP1 CNVs and inferred the underlying mechanisms for 19 ( 66% ) of them on the basis of copy number and breakpoint sequences ( Methods ) . Of the 19 GAP1 CNVs that can be reliably resolved , 3 are the result of aneuploidies and 2 are the result of nonreciprocal interchromosomal translocations ( S5 Table ) . Translocations were confirmed using pulsed-field gel electrophoresis and Southern blot analysis ( S5 Fig ) , which clearly shows that the second copy of GAP1 is located on a different chromosome . Southern blotting also indicates that an additional 3 GAP1 CNVs are the result of partial ( i . e . , segmental ) aneuploidies , which include the Chromosome XI centromere ( CEN11 ) but are smaller than the ancestral Chromosome XI ( S5 Fig ) . At least 4 GAP1 CNVs appear to be the result of a tandem duplication mediated by NAHR . For two of these CNVs , novel junction sequences were obtained that included a hybrid sequence composed of half of each flanking LTR ( YKRCdelta11/YKRCdelta12 ) , similar to our previous report [25] . This mechanism is also likely to underlie the GAP1 deletion that we identified in one urea-limited population . For 12 out of 29 ( 41% ) GAP1 CNVs and 8 out of 9 ( 89% ) DUR3 CNVs , we identified a pair of short , interrupted , inverted repeats proximate to at least one breakpoint ( Fig 4 and S2 Text ) . We were able to resolve breakpoints at both ends of the CNV for 12 of the 20 CNVs . Analysis of these breakpoints indicates that inverted repeat sequences range in length from 4 to 24 base pairs ( Fig 4D ) and are typically separated by 40 base pairs ( Fig 4E ) . Microhomology at breakpoint junctions is characteristic of replication-based CNV formation , including MMBIR and ODIRA . ODIRA has several other requirements , including the presence of at least one replication origin within the CNV , an internal inversion , and an odd copy number . The identification of inverted sequence relative to the reference at all identified breakpoint junctions is consistent with an inverted structure . We find that 6/29 GAP1 CNVs and 8/9 DUR3 CNVs meet these criteria and thus are likely the result of ODIRA . In cases when the CNV lacks an odd copy number ( see Methods ) we cannot reliably infer the mechanism ( S5 Table ) . In one case ( ure_07_c1 ) , the CNV meets all the requirements of ODIRA but does not contain a DNA replication origin ( see Discussion ) . To comprehensively characterize genomic variation in populations , we performed whole-population , whole-genome sequencing of glutamine- , urea- , and glucose-limited populations at generations 150 and 250 ( S3 Table ) . Analysis of relative sequence read depth is consistent with high-frequency GAP1 CNVs in glutamine-limited populations ( S2 Text ) . Population sequencing also confirmed the fixation of a GAP1 deletion ( ure_05 ) in a urea-limited population . Relative sequence read depth at the GAP1 locus correlates well with the normalized fluorescence of the GAP1 CNV reporter in populations ( S10 Fig ) , providing additional evidence for the utility of the CNV reporter . In glutamine-limited chemostats , GAP1 copy number estimated within populations ( which is a function of copy number within clones and allele frequencies ) ranges from 2 to 4 copies , with a trend toward increased copy number over time ( S10 Fig ) . We performed single-nucleotide variant ( SNV ) analysis using genome sequencing data from populations ( S9 Table ) and clones ( S10 Table ) at generations 150 and 250 . More nonsynonymous SNVs were identified in glucose-limited populations than the glutamine- and urea-limited populations ( Table 2 ) , which contained GAP1 and DUR3 amplifications at high frequencies at 150 and 250 generations . In contrast to previous studies [28 , 29] , we did not identify CNVs at the HXT6/7 locus in glucose-limited populations . Increased nucleotide variation within these populations may reflect alternative adaptive strategies in glucose-limited populations . We find several genes with multiple independent , nonsynonymous variation in glutamine-limited populations ( Table 3 ) , including MCK1 , a protein kinase with potential roles in nonhomologous end joining ( NHEJ ) ; SOG2 , a member of the regulation of Ace2p activity and cellular morphogenesis ( RAM ) signaling pathway and regulator of bud separation after mitosis; and TAO3 , another member of the RAM network . We previously reported mutations in MCK1 from selection in glutamine- and arginine-limited chemostats [24] , suggesting that it is a recurrent target of selection in these conditions . Changes in cell morphology are potentially adaptive in nutrient-poor conditions , which may result from defects in cell cycle progression and bud separation associated with mutations in the RAM pathway [65] . However , the effect of these mutations on bud separation is likely to be minor , as we did not observe increases in forward scatter ( which varies with cell size ) in flow cytometry data , except in one glucose-limited population ( S3 Fig ) . In the nine urea-limited populations , we identified 14 independent nonsynonymous variants in DUR1 , 2 ( Table 3 ) . DUR1 , 2 encodes urea amidolyase , which metabolizes urea to ammonium . At two different nucleotide positions , we find that the same nucleotide was mutated multiple times independently . In a third location , we identified an SNV at the exact nucleotide position as we previously reported [24] . Thus , a subset of variants in DUR1 , 2 appear to be uniquely beneficial and recurrently selected in urea-limited environments . In glucose-limited populations , we identified multiple , independent mutations in four genes ( Table 3 ) : TRK1 , a component of the potassium transport system; SVF1 , which is important for the diauxic growth shift and is implicated in cell survival during aneuploidy [66]; CDC48 , an ATPase associated with diverse cellular activities ( AAA ) ; and WHI2 , which is a mediator of the cellular stress response . Previous studies have identified loss-of-function mutations in WHI2 , suggesting it is a general target of selection across different conditions [24 , 27 , 67] . Analysis of clonal samples ( S10 Table ) was largely consistent with population sequencing . We identified two cases in which SNVs occurred within GAP1 CNVs . These SNVs are present at frequencies of 53% in a lineage containing a GAP1 duplication and 30% in a lineage containing a GAP1 triplication , indicating that they are present on only one of the copies within the CNV . We also identified polymorphisms within DUR3 amplifications ( S10 Table ) . This suggests that individual copies of a gene within a CNV can accumulate additional nucleotide variation even in relatively short-term evolutionary scenarios . Eight of the 9 clones with DUR3 amplifications also acquired a variant in DUR1 , 2 , which may be indicative of a synergistic relationship between CNVs and SNVs . The reproducible dynamics of CNV lineages observed during glutamine-limited experimental evolution may be due to two nonexclusive reasons: either ( 1 ) a high supply rate of de novo CNVs or ( 2 ) preexisting CNVs in the ancestral population ( S11 Fig ) . In both scenarios , a single CNV or multiple , competing CNVs may underlie the reproducible dynamics . Sequence analysis of clonal lineages suggests at least two , and as many as four , CNV lineages may coexist in populations ( Fig 3 ) ; however , genome sequencing is uninformative about the total number of lineages for two key reasons . First , the recurrent formation of CNVs confounds distinguishing CNVs that are identical by state from those that are identical by descent . Second , CNVs that arise de novo may subsequently diversify over time , resulting in distinct alleles that are derived from a common event . To quantify the number , relationship , and dynamics of individual CNV lineages , we constructed a lineage-tracking library using random DNA barcodes [68] . We constructed a library of approximately 80 , 000 unique barcodes ( S12 Fig ) in the background of the GAP1 CNV reporter and performed six independent replicate experiments in glutamine-limited chemostats . Real-time monitoring of CNV dynamics using the GAP1 CNV reporter recapitulated the dynamics of our original experiment ( Fig 5A , S13A Fig , and S11 Table ) , although CNV lineages appeared significantly earlier in these populations ( Tup; t test p-value < 0 . 01 ) . As the lineage-tracking strain was independently derived from the strain used in our original experiment , these results indicate that selection of GAP1 CNVs in glutamine-limited chemostats is reproducible and independent of genetic background . To quantify individual lineages , we isolated the subpopulation containing CNVs from two populations ( bc01 and bc02 ) at multiple time points ( generations 70 , 90 , 150 , and 270 ) . Isolation of the CNV subpopulation was performed by FACS using gates based on one- and two-copy control populations ( Fig 5A , S14 Fig ) . We sequenced barcodes from the CNV subpopulation at each time point and determined the number of unique lineages ( [69] and Methods ) . To account for variation in the purity of the FACS-isolated CNV subpopulation , we analyzed individual clones using a flow cytometer . Using these data , we estimated a false positive rate , which we find varies between time points ( S13B Fig and Methods ) , and applied this correction to barcode counts ( Table 4 ) . We detect thousands of independent GAP1 CNV lineages at generation 70 , indicating that a large number of independent GAP1 CNVs are generated and selected in the early stages of the evolution experiments ( Fig 5B ) . Applying a conservative false positive correction , we identified 7 , 067 GAP1 CNV lineages in bc01 and 5 , 305 GAP1 CNV lineages in bc02 at generation 70 ( Table 4 ) . If we only consider lineages detected in the CNV subpopulation at multiple time points , we identify 891 CNV lineages in bc01 and 2 , 676 CNV lineages at generation 70 ( Table 4 ) . Thus , between 102 and 104 independent CNV lineages in each population of 108 cells initially compete with each other . The overall diversity of CNV lineages decreases with time , consistent with decreases in lineage diversity observed in other evolution experiments [68 , 70] . By generation 270 , we detect only 76 CNV lineages in bc01 and 28 CNV lineages in bc02 . To determine the dominant lineages in each population , we identified barcodes that reached greater than 1% frequency in the CNV subpopulation in at least one time point: 21 independent lineages are found at greater than 1% frequency in bc01 , and 18 independent lineages are found at greater than 1% frequency in bc02 ( Fig 5B ) . These results indicate the presence and persistence of multiple GAP1 CNVs across hundreds of generation of selection , during which there is a continuous reduction in the overall diversity of CNV lineages . Although CNVs rise to high frequencies in both populations ( Fig 5A ) , the composition of competing CNV lineages is dramatically different: in bc02 , a single lineage dominates the population by generation 150 ( Fig 5B ) , whereas in bc01 , there is much greater diversity at later time points . In both populations , several CNV lineages that comprise a large fraction of the CNV subpopulation at early generations ( generations 70 , 90 , or 150 ) are extinct by generation 270 . Thus , within populations , individual CNV lineages do not increase in frequency with uniform dynamics , despite the consistent and reproducible dynamics of the entire CNV subpopulations ( Fig 5A and Fig 2 ) . Differences in fitness between individual CNV lineages , possibly as a result of variation in copy number , CNV size , and secondary adaptive mutations , are likely to contribute to these dynamics . To distinguish the contribution of preexisting genetic variation ( i . e . , CNVs introduced to the population before chemostat inoculation; S11 Fig ) and de novo variation ( i . e . , CNVs introduced to the population following chemostat inoculation ) to CNV lineage dynamics , we assessed whether barcodes were shared between CNV lineages in independent populations . We identified four barcodes at greater than 1% frequency that are common to both populations ( Fig 5B ) . At generation 70 , one of these barcodes ( indicated in light purple ) was present at 14% and 19% in bc01 and bc02 , respectively . We find that the barcode for this lineage was overrepresented in the ancestral unselected population ( an initial frequency of 0 . 014% , which is one order of magnitude greater than the average starting frequency of 0 . 0011%; S12 Fig ) . Although there is a possibility that de novo CNVs formed independently in this barcode lineage , it is more likely that this lineage contained a preexisting CNV in the ancestral population . Although this lineage represented a sizable fraction of the CNV subpopulation in both replicate populations , it was only maintained at high frequency in one of them ( bc01 ) . Only one of the four preexisting CNV lineages persists throughout the experiment in both populations . By contrast , in each population , we identified 17 and 14 unique high-frequency CNV lineages that are most likely new CNVs . These results indicate that both preexisting CNVs and de novo CNVs that arise during glutamine limitation contribute to adaptive evolution .
To determine the dynamics with which CNVs are selected at the GAP1 locus , we inserted a constitutively expressed fluorescent gene adjacent to GAP1 and tracked changes in single-cell fluorescence over time . Whereas one- and two-copy control strains with mCitrine at neutral loci maintain a steady fluorescent signal over 250 generations of selection , all glutamine-limited populations with the GAP1 CNV reporter show increased fluorescence by generation 75 . The structure and breakpoints of CNVs within and between populations are different , indicating independent formation of CNVs . Control strains were inoculated independently and have different genetic backgrounds but also form CNVs at the GAP1 locus , as determined by whole-genome sequencing and Southern blot analysis . These data indicate that GAP1 CNVs are positively selected early and repeatedly in glutamine-limited environments . Although the majority of evolved clones with GAP1 CNVs ( 18/28 ) have higher relative fitness in glutamine-limited chemostats compared to the ancestor , several clones have neutral ( 8/28 ) or lower ( 2/28 ) relative fitness . CNV-containing clones were selected on the basis of increased fluorescence , which does not necessarily mean the clone had higher fitness than the ancestor . The fitness effect of a CNV within the chemostat environment is context specific and may depend on factors such as frequency-dependent selection . In addition , if GAP1 CNVs are generated at a high rate , as we have hypothesized , neutral or deleterious CNVs could be present for several generations before these lineages are purged from the population or acquire additional adaptive mutations . Whole-genome sequencing of GAP1 CNV lineages isolated on the basis of increased fluorescence uncovered a wide range of CNV structures within and between populations . We found cases in which distinct alleles were identified within populations at different time points and cases in which we identified the same CNV allele 100 generations later . GAP1 CNV alleles are 105 kilobases on average but can include the entire right arm of Chromosome XI ( 260 kilobases ) . A previous study in bacteria showed that there is a cost to gene duplication , with a fitness reduction of 0 . 15% per kilobase [71] . Therefore , we hypothesized that CNVs would decrease in size over evolutionary time through a refinement process in order to reduce the fitness burden . However , we failed to detect a significant reduction in CNV allele size over time . This may be because increased CNV size does not confer a fitness cost in yeast , the fitness benefit of the GAP1 CNV outweighs this cost , or there are other genes within the CNV whose amplification confers a fitness benefit . Our reporter detects increases in gene copy number that result from a variety of processes such as aneuploidy , nonreciprocal translocation , tandem duplication , and complex CNVs , including inverted triplications . The ability to track and isolate these diverse gene amplifications allows us to enumerate the frequency of each type and characterize the mechanisms underlying their formation . Combining our approach with molecular techniques allowed us to further understand the nature of these GAP1 CNVs . Three particularly interesting GAP1 CNV-containing clones appear to have partial ( i . e . , segmental ) aneuploidies that encompass centromere 11 ( S5 Fig ) . As the presence of two centromeres in one chromosome is extremely unlikely , it is plausible that these exist as independent , supernumerary chromosomes [72] . Similar adaptive rearrangements occur in other yeast species: isochromosome formation , potentially mediated by the presence of inverted repeats , has been observed during treatment of Candida albicans with antifungal drugs [73] . The use of a CNV reporter should facilitate determination of the frequency with which these and other complex mechanisms give rise to CNVs at a given locus . Breakpoint analysis provided further insight into the mechanisms underlying CNV formation . We identified breakpoints within LTRs and other repetitive elements for 4 unique glutamine-limited clones that have 2 copies of GAP1 . These findings suggest that these CNVs were formed by a tandem duplication mediated through NAHR . Of these , 3 GAP1 gene amplifications ( 3/28 ) are formed after NAHR between flanking LTRs YKRCdelta11 and YKRCdelta12 . The GAP1 deletion , which occurred in one population undergoing urea limitation , also had breakpoints in these flanking elements consistent with NAHR-mediated gene deletion . NAHR may drive the nonreciprocal translocations we identified and additional unresolved events with breakpoints adjacent to LTRs . We did not find evidence for the selection of GAP1circle CNVs in any population . Thus , it may be that circular elements containing beneficial genes only exist transiently in cells and may rapidly resolve to chromosomal amplifications via homologous recombination–mediated reintegration . We identified 9 GAP1 CNVs and 8 DUR3 CNVs that contain breakpoints characterized by closely spaced inverted repeat sequences . Of these , the majority ( 14/17 ) also had an odd copy number and contained an origin of replication consistent with the ODIRA mechanism [47 , 48] . However , we also identified one DUR3 CNV that does not include a replication origin ( ure_07_c1 ) , although the origin is nearby ( <1 kilobase ) . This could result from a distinct replication-based mechanism of CNV generation . For example , MMBIR is a RAD51-independent process that relies on short stretches of homology ( “microhomology” ) to restart a stalled replication fork [45] . Though we cannot explicitly distinguish between these models , the short stretches of homology in the inverted repeats is inconsistent with formation of this CNV by NAHR . Thus , while NAHR plays an important role in CNV formation , our results suggest that replication-based mechanisms may be a major source of gene amplification in yeast . This is consistent with increasing evidence for replication-based CNV formation in diverse organisms including yeast , mice , and humans [74–77] . Comparison between DUR3 and GAP1 CNVs identified quantitative differences in CNV formation at the two loci . We primarily identified CNVs with 2 or 3 copies of GAP1 in glutamine-limited clones , but urea-limited clones always contained 5 copies of DUR3 . The size ( average of 26 kilobases ) of DUR3 CNVs was also significantly smaller than GAP1 CNVs . Molecular characterization revealed a diverse range of processes underlying GAP1 CNV formation , whereas DUR3 CNVs are all characterized by inversions mediated by short , interrupted , inverted repeats . These data suggest that generation and selection of CNVs vary as a function of locus and selective condition . The CNV reporter can readily be integrated throughout the genome to further test whether there are fundamental differences in CNV formation mechanisms at different loci and how these differences change the temporal dynamics of CNV selection . By combining a CNV reporter with lineage tracking , we identified a surprisingly large number of independent CNV lineages . Whereas clonal isolation and sequencing suggested at least four independent lineages within populations , lineage tracking indicates that hundreds to thousands of individual CNV lineages emerge within fewer than 100 generations . Most of these lineages do not achieve high frequency , as we identified only 18–21 lineages present at >1% frequency in the CNV subpopulation . The number of independent CNV lineages we identified is remarkable . Although we have attempted to account for technical factors that may inflate this number , unanticipated aspects of barcode transformation and library construction , cell sorting , and barcode sequencing and identification may impact this estimation . Conversely , the exact number of CNV lineages may be underestimated , as the unselected barcode library was not maximally diverse and each unique barcode was shared by multiple founding cells . Although we found lineages that were common to both populations ( at least one of which is likely to contain a preexisting CNV ) , ancestral CNV lineages do not drive the evolutionary dynamics . Preexisting CNV lineages have different dynamics in each population and do not prevent the emergence of unique de novo CNV lineages . This demonstrates that the ultimate fate of a CNV lineage depends on multiple factors , and a high frequency at an early generation does not guarantee that a lineage will persist in the population . Thus , CNV dynamics result from preexisting and de novo variation and are characterized by extensive clonal interference and replacement among competing CNV lineages . The large number of CNV lineages identified in our study indicates that they occur at a high rate . Recent studies have suggested that adaptive mutations may be stimulated by the environment . Stress can lead to increases in genome-wide mutation rates in both bacteria and yeast [78–80] , and replicative stress can lead directly to increased formation of CNVs [81 , 82] . Other groups have proposed an interplay between transcription and CNV generation and that active transcription units might even be “hotspots” of CNV formation [83–85] . These hotspots , often designated as common fragile sites , may occur in long , late-replicating genes , with large interorigin distances [82] . Local transcription at the rDNA locus leads to rDNA amplification and is thought to be regulated in response to the environment [86 , 87] . Transcription of the CUP1 locus in response to environmental copper leads to promoter activity that further destabilizes stalled replication forks and generates CNVs [88] . Given the high level of GAP1 transcription in nitrogen-limited chemostats [58] , it is tempting to speculate that this condition may promote the formation of GAP1 CNVs . Further studies are required to understand the full extent of processes that underlie CNV formation at the GAP1 locus and how these different mechanisms may contribute to the fitness and overall success of CNV lineages . The frequency of GAP1 CNVs can be attributed to a combination of factors , including a high mutation supply rate due in part to the large chemostat population size ( approximately 108 cells ) , the strength of selection , and the fitness benefit typically conferred by GAP1 amplification . Together , these factors contribute to an early , deterministic phase , during which CNVs are formed at a high rate and thousands of lineages with CNVs rapidly increase in frequency . During a second phase , the dynamics are more variable , as competition from different types of adaptive lineages and additional acquired variation influence evolutionary trajectories of individual CNV lineages . This phenomenon has recently been observed in other evolution experiments , in which early events are driven by multiple competing single-mutant lineages [70] , but later dynamics are influenced by stochastic factors and secondary mutations [68] . The high degree of clonal interference observed among a single class of adaptive mutations may have important implications for adaptive evolution . CNVs are alleles of large effect that can simultaneously change the dosage of multiple protein-coding genes and subsequently lead to changes in cell physiology . Epistatic relationships between CNVs and other adaptive mutations could therefore dramatically alter the fitness landscape [31] . Additionally , CNVs can confer a fitness benefit per se but also serve to increase the amount of DNA in the genome that can accumulate mutations . Therefore , CNVs can potentially increase the rate of adaptive evolution by increasing the target size for adaptive mutations . In this study , we found evidence for polymorphisms within individual CNVs and potential epistasis between SNVs and CNV alleles , two phenomena that require further exploration as we continue to define the role of CNVs in driving rapid adaptive evolution . The combined use of a fluorescent CNV reporter and barcode lineage tracking provides unprecedented insight into this important class of mutation . Previous studies have tracked specific mutations and their fitness effects [60] , but ours is the first single cell–based approach to identify an entire class of mutations and follow evolutionary trajectories with high resolution . Whereas barcode tracking alone provides information about the number of adaptive lineages and their fitness effects , the CNV reporter enables us to specifically determine the number of unique CNV events . In addition , the reporter provides an estimate of the total proportion of CNVs in the population , which we can use to inform our understanding of lineage dynamics . Using these tools , we have shown that CNVs are generated at a high rate through diverse mechanisms including homologous recombination and replication-based errors . These processes lead to the formation of many distinct CNV alleles segregating within populations . One limitation of our approach is that a complex CNV could be the product of multiple , independent events ( e . g . , a duplication followed by a subsequent triplication ) . Evolution experiments that start with a preexisting CNV would be informative for studying how CNVs diversify when maintained under selection . Our results demonstrate an important role for CNVs in driving rapid adaptive evolution in microbial populations but could be broadly applicable to plants , animals , and humans . Our system provides a facile means for studying the molecular processes underlying CNV generation as well as evolutionary aspects of CNVs , including whether there are fundamental differences in CNV formation and selection at different loci , the impact of a high rate of CNV formation on the evolutionary dynamics of other adaptive lineages , how CNVs are maintained or refined over longer evolutionary timescales , how CNVs interact with other adaptive mutations to influence fitness landscapes , whether there are consequences and tradeoffs in alternative environments , and how the formation of CNVs impacts gene expression and genome architecture . Extension of this method is likely to be useful for addressing additional fundamental questions regarding the evolutionary and pathogenic role of CNVs in diverse systems .
We used FY4 and FY4/5 , haploid and diploid derivatives of the reference strain S288c , for all experiments . S1 Table is a comprehensive list of strains constructed and used in this study . To generate fluorescent strains , we performed high-efficiency yeast transformation [89] with an mCitrine gene under control of the constitutively expressed ACT1 promoter ( ACT1pr::mCitrine::ADH1term ) and marked by the KanMX G418-resistance cassette ( TEFpr::KanMX::TEFterm ) . The entire construct , which we refer to as the mCitrine CNV reporter , is 3 , 375 base pairs . For control strains , the mCitrine reporter was integrated at two neutral loci: HO ( YDL227C ) on Chromosome IV and the dubious ORF , YLR123C , on Chromosome XII . Diploid control strains containing 3 and 4 copies of the mCitrine CNV reporter were generated using a combination of backcrossing and mating . We constructed the GAP1 CNV reporter by integrating the mCitrine construct at an intergenic region 1 , 118 base pairs upstream of GAP1 ( integration coordinates , Chromosome XI: 513945–517320 ) . PCR and Sanger sequencing were used to confirm integration of the GAP1 CNV reporter at each location ( all PCR primer sequences are provided in S12 Table ) . Transformants were subsequently backcrossed and sporulated , and the resulting segregants were genotyped . For the purpose of lineage tracking , we constructed a strain containing a landing pad and the GAP1 CNV reporter by segregation analysis after mating the original GAP1 CNV reporter strain to a landing pad strain ( derived from BY4709 ) [68] . As the kanMX cassette is present at two loci in this cross , we performed tetrad dissection and identified four spore tetrads that exhibited 2:2 G418 resistance . A segregant with the correct genotype ( G418 resistant , ura- ) was identified and confirmed using a combination of PCR ( S12 Table ) and fluorescence analysis . We introduced a library of random barcodes by transformation and selection on SC-ura plates [68] . We plated an average of 500 transformants on 200 petri plates and estimated 78 , 000 independent transformants . Nitrogen-limiting media ( glutamine and urea limitations ) contained 800 μM nitrogen regardless of molecular form and 1 g/L CaCl2-2H2O , 1 g/L of NaCl , 5 g/L of MgSO4-7H2O , 10 g/L KH2PO4 , 2% glucose and trace metals and vitamins as previously described [24] . Glucose-limiting media contained 0 . 08% glucose , 1 g/L CaCl2-2H2O , 1 g/L of NaCl , 5 g/L of MgSO4-7H2O , 10 g/L KH2PO4 , 50 g/L ( NH4 ) 2SO4 and trace metals and vitamins [90] . We inoculated the GAP1 CNV reporter strain into 20-mL ministat vessels [91] containing either glutamine- , urea- , or glucose-limited media . Control populations containing either one or two copies of the CNV reporter at neutral loci ( HO and YLR123C ) were also inoculated in ministat vessels for each media condition . Ministats were maintained at 30°C in aerobic conditions and diluted at a rate of 0 . 12 hour−1 ( corresponding to a population doubling time of 5 . 8 hours ) . Steady-state populations of 3 × 108 cells were maintained in continuous mode for 270 generations ( 65 days ) . Every 30 generations , we archived 2-mL population samples at −80°C in 15% glycerol . To monitor the dynamics of CNVs , we sampled 1 mL from each population about every 8 generations . We performed sonication to disrupt any cellular aggregates and immediately analyzed the samples on an Accuri flow cytometer , measuring 100 , 000 cells per population for mCitrine fluorescence signal ( excitation = 516 nm , emission = 529 nm , filter = 514/20 nm ) , cell size ( forward scatter ) , and cell complexity ( side scatter ) . We generated a modified version of our laboratory flow cytometry pipeline for this analysis ( https://github . com/GreshamLab/flow ) , which uses the R package flowCore [92] . We used forward scatter height ( FSC-H ) and forward scatter area ( FSC-A ) to filter out doublets and FSC-A and side scatter area ( SSC-A ) to filter debris . We quantified fluorescence for each cell and divided this value by the forward scatter measurement for the cell to account for differences in cell size . To determine population frequencies of cells with zero , one , two , and three or more copies of GAP1 , we used one- and two-copy control strains grown in glutamine-limited chemostats to define gates and perform manual gating . We used a conservative gating approach to reduce the number of false positive CNV calls by manually drawing first a liberal gate for the one-copy control strain and then a nonoverlapping gate for the two-copy control strain . Flow cytometry data and code used to generate all figures and tables can be accessed in OSF: https://osf . io/fxhze/ . To quantify the dynamics of CNVs in evolving populations , we defined summary statistics as in [60] . Tup is the generation at which CNVs are initially detected , and Sup is the slope of the linear fit during initial population expansion of CNVs . We first determined the proportion of cells with a CNV and the proportion of cells without CNVs at each time point , using the manually defined gates . To calculate Tup , we defined a false positive rate for CNV detection in evolving one-copy control strains from generations 1–153 ( defined as the average plus one standard deviation = 7 . 1% ) . We designate Tup once an experimental population surpasses this threshold . To calculate Sup , we plotted the natural log of the ratio of the proportion of cells with and without a CNV against time and calculated the linear fit during initial population expansion of CNVs . We defined the linear phase on the basis of R2 values ( S1 Text ) . Sup can also be defined as the percent increase in CNVs per generation , which is an approximation for the relative average fitness of all CNV alleles in the population . Clonal isolates were obtained from each glutamine- and urea-limited population at generation 150 and generation 250 . We isolated clones by plating cells onto rich media ( YPD ) and randomly selecting individual colonies . We inoculated each clone into 96-well plates containing the limited media used for evolution experiments and analyzed them on an Accuri flow cytometer following 24 hours of growth . We compared fluorescence to unevolved ancestral strains , evolved 1- and 2-copy controls grown under the same conditions , and chose a subset of clones for whole-genome sequencing ( S4 Table ) . To measure the fitness coefficient of evolved clones , we performed pairwise competitive fitness assays in glutamine-limited chemostats using the same glutamine-limited conditions as our evolution experiments [24] . We cocultured our fluorescent evolved strains with a nonfluorescent , unevolved reference strain ( FY4 ) . We determined the relative abundance of each strain every 2–3 generations for approximately 15 generations using flow cytometry . We performed linear analysis of the natural log of the ratio of the two genotypes against time and estimated the fitness and associated error relative to the ancestral strain . Evolved clones were grown overnight in glutamine-limited media and embedded in agarose using Bio-rad plug molds . Plugs were incubated in zymolyase T100 ( 200 μg/mL ) overnight at 37°C , proteinase K ( 4 mg/mL ) overnight at 50°C , and PMSF ( 1 mM ) for 1 hour at 4°C . PMSF was removed by washing plugs with 1 mL of CHEF TE 3 times for 30 minutes . Plugs were subsequently run in a 1X TAE , 1% agarose gel using a Bio-rad CHEF-DR II . Southern blotting was performed by alkaline transfer using Hybond-XL membranes . Blots were subsequently probed with 32P-labeled DNA complementary to GAP1 or CEN11 . Probes were created using nested PCR with primers listed in S12 Table . Signal from blots was detected using FujiFilm imaging plates and imaged using Typhoon FLA9000 . For both population and clonal samples , we performed genomic DNA extraction using a modified Hoffman-Winston protocol [93] . We used SYBR Green I to measure gDNA concentration , standardized each sample to 2 . 5 ng/μL , and constructed libraries using tagmentation following a modified Illumina Nextera library preparation protocol [94] . To perform PCR clean-up and size selection , we used an Agilent Bravo liquid-handling robot . We measured the concentration of purified libraries using SYBR Green I and pooled libraries by balancing their concentrations . We measured fragment size with an Agilent TapeStation 2200 and performed qPCR to determine the final library concentration . DNA libraries were sequenced using a paired-end ( 2 × 75 ) protocol on an Illumina NextSeq 500 . Standard metrics were used to assess data quality ( Q30 and %PF ) . To remove reads from a potentially contaminating organism that was introduced after recovery from the chemostats , we filtered any reads that aligned to Pichia kudriavzevii . Given the evolutionary divergence between these species , the majority of filtered reads belonged to rDNA and similar , deeply conserved sequences . The median percent contamination was 1 . 165% . We modified the S . cerevisiae reference genome from NCBI ( assembly R64 ) to include the entire GAP1 CNV reporter and aligned all reads to this reference . We aligned reads using bwa mem ( [95] , version 0 . 7 . 15 ) and generated BAM files using samtools ( [96] , version 1 . 3 . 1 ) . Summary statistics for all sequenced samples are provided in S3 Table . FASTQ files for all sequencing are available from the SRA ( accession SRP142330 ) . Sequencing data and code used to generate all figures and tables can be accessed in OSF: https://osf . io/fxhze/ . To assess the performance of CNV detection algorithms , we simulated CNVs ranging in size from 50 to 100 , 000 base pairs in 100 synthetic yeast genomes . We used SURVIVOR [97] to simulate CNVs in the reference yeast genome and wgsim [96] to generate corresponding paired-end FASTQ files . We used bwa mem [95] to map reads back to the reference and called CNVs with Pindel , CNVnator , LUMPY , and SvABA [61–64] . We assessed the effect of read depth on algorithm performance by downsampling a 100× coverage BAM file to 80× , 50× , 20× , 10× , and 5× coverage . We defined a CNV as being correctly predicted if the simulated and detected CNVs were ( 1 ) of the same type ( e . g . , duplication ) , ( 2 ) predicted to be on the same chromosome , and ( 3 ) contained in the same interval ( defined by the start and stop position ) , which were considered overlapping if there was no gap between them ( maxgap = 0 ) and had minimum overlap of 1 base pair ( minoverlap = 1 ) . For intervals [a , b] and [c , d] , for which a ≤ b and c ≤ d , when c ≤ b and d ≥ a the two intervals overlap , and when c > b or d < a the two intervals do not overlap . If the gap between these two intervals is ≤maxgap and the length of overlap between these two intervals is ≥minoverlap , the two intervals are considered to be overlapping . To assess the performance of these tools on heterogeneous population samples , we also simulated mixed samples by combining reads from a simulated CNV-containing genome and an unmodified reference yeast genome at varying proportions . The ratio of the reads from the CNV-containing genome varied between 20% and 90% , and the total coverage was 50× . Performance comparisons for all benchmarking were based on false discovery rate ( FDR ) and F-score . The F-score ( also known as F1 measure ) combines sensitivity/recall ( r ) and precision ( p ) with an equal weight using the formula F = ( 2pr ) / ( p + r ) [98] . An F-score reaches its best value at 1 and worst at 0 and was multiplied by 100 to convert to a percentage value . We called CNVs for each clone and population sample using an in-house pipeline that collates results from Pindel , SvABA , and LUMPY ( S5 Table and S6 Table ) . Data and code used to generate these figures can be accessed in OSF: https://osf . io/fxhze/ . To manually estimate CNVs boundaries , we used a read depth–based approach . For each sample sequenced , we used samtools [96] to determine the read depth for each nucleotide in the genome . We liberally defined CNVs by identifying ≥300 base pairs of contiguous sequence when read depth was ≥3 times the standard deviation across Chromosome XI for GAP1 or Chromosome VIII for DUR3 . These boundaries were further refined by visual inspection of contiguous sequence ≥100 base pairs with read depth ≥3 times the standard deviation . These analyses were only performed on sequenced clones because population samples are likely to have multiple CNVs and breakpoints , thereby confounding read depth–based approaches . We compared manually estimated breakpoints to those identified by the algorithms ( S5 Table ) and defined a set of “high-confidence breakpoints . ” To determine CNV breakpoints at nucleotide resolution , we extracted split and discordant reads from bam files using samblaster [99] . Both split reads and discordant reads were used to identify breakpoints using a weighted scoring method wherein a split read was worth 1 and discordant reads were worth 3 . Positively identified breakpoints required at least 4 split reads and a combined score of at least 9 . Breakpoint sequences were generated by making local assemblies of breakpoint-associated split reads using MAFFT , EMBOSS , and velvet [100–102] . The relationship between breakpoint sequences and the reference genome was determined using BLAST+ [103] , with blastn and blastn-short using default settings . To infer the underlying mechanism by which CNVs were formed , we applied the following criteria . If at least one of the two CNV boundaries contained inverted repeat sequences , and we estimated an odd number of copies in the CNV , we classified the mechanism as ODIRA [26 , 47 , 48] . If both of the CNV boundaries occurred within repetitive sequence elements ( LTRs or telomeres ) and had two copies , we inferred tandem duplication by NAHR [40] . Aneuploids were defined on the basis of increased read depth throughout the entire chromosome but no detected novel sequence junctions . Translocations were identified by LUMPY and Southern blot analysis . All breakpoints that failed to meet these criteria were defined as unresolved . In addition to CNVs at GAP1 and DUR3 , we also identified additional structural variants ( S7 Table ) and CNVs ( S8 Table ) . Structural variants were identified using the split and discordant read approach described above . Additional CNVs were identified using a two-pass genome-wide read-depth approach . In the initial pass , each sample was scanned for regions ( 400 nucleotide minimum size ) with read depth higher than 3 standard deviations relative to the genome . During the second pass , the read depth of each candidate is normalized by the median read depth of that region , as calculated using a subset of clones that lack a candidate in that region . This normalization allows for the correction of sequencing artifacts , batch effects , and the removal of CNV regions that are not substantially different between the evolved and ancestral clones ( i . e . , rDNA , Ty elements , etc . ) SNVs and indel variants were first identified using GATK4’s Mutect2 [104] , which allows for the identification of variants in evolved samples ( “Tumor” ) after filtering using matched unevolved samples ( “Normal” ) and pool of normals ( PON ) . The PON was constructed using 6 sequenced ancestral clones , whereas the paired normal was a single , deeply sequenced ancestor . Variants were further filtered using GATK’s FilterMutectCalls to remove low-quality predictions; only variants flagged as “passed” or “germline risk” were retained . Given the haploid nature of the evolved population and further downstream filtering of “too-recurrent” mutations , we allowed germline risk variants to be retained . Variants were further filtered if they occurred in low-complexity sequence; i . e . , variants were filtered if the SNV or indel occurred in or generated a homogenous nucleotide stretch of five or more of the same nucleotide . Variants from within populations that were detected at less than 5% frequency were considered low confidence and excluded . Finally , variants were filtered if they were found to be “too recurrent”; i . e . , if the exact nucleotide variant was identified in more than three independently evolved lineages , we deemed it more parsimonious to assume that the variant was present in the ancestor at low frequency . We inoculated the lineage-tracking library into 20-mL ministat vessels [91] containing glutamine-limited media . Control populations containing either zero , one or two copies of the GAP1 CNV reporter at neutral loci ( HO and YLR123C ) were also inoculated in ministat vessels for each media condition . Control populations did not contain lineage-tracking barcodes . Ministat vessels were maintained and archived as above . Samples were taken for flow cytometry about every 8 generations and analyzed as previously described . We used FACS to isolate the subpopulation of cells containing two or more copies of the mCitrine CNV reporter using a FACSAria . We defined our gates using zero- , one- , and two-copy mCitrine control strains sampled from ministat vessels at the corresponding time points: 70 , 90 , 150 , and 265 generations . Depending on the sample , we isolated 500 , 000–1 , 000 , 000 cells with increased fluorescence , corresponding to 2 or more copies of the reporter . We grew the isolated subpopulation containing CNVs for 48 hours in glutamine-limited media and performed genomic DNA extraction using a modified Hoffman-Winston protocol [93] . We verified FACS isolation of true CNVs by isolating clones from subpopulations sorted at generation 70 , 90 , and 150 ( sorted from all lineage-tracking populations , bc01–bc06 ) and performing independent flow cytometry analysis using an Accuri . We estimated the average false positive rate of CNV isolation at each time point as the percent of clones from a population with FL1 less than one standard deviation above the median FL1 in the one copy control strain . Only subpopulations with fluorescence measurements for at least 25 clones were included in calculations of false positive rate . We performed a sequential PCR protocol to amplify DNA barcodes and purified the products using a Nucleospin PCR clean-up kit [68] . We quantified DNA concentrations by qPCR before balancing and pooling libraries . DNA libraries were sequenced using a paired-end ( 2 × 150 ) protocol on an Illumina MiSeq 300 Cycle v2 . Standard metrics were used to assess data quality ( Q30 and %PF , S3 Table ) . However , the reverse read failed because of overclustering , so all analyses were performed only using the forward read . We used the Bartender algorithm with UMI handling to account for PCR duplicates and to cluster sequences with merging decisions based solely on distance except in cases of low coverage ( <500 reads/barcode ) , for which the default cluster merging threshold was used [69] . Clusters with a size less than 4 or with high entropy ( >0 . 75 quality score ) were discarded . We estimated relative abundance of barcodes using the number of unique reads supporting a cluster compared to total library size . Data and code used to generate these figures and tables can be accessed in OSF: https://osf . io/fxhze/ . | Duplications and deletions of genomic sequence , known as copy number variants , are a common source of genetic diversity across all domains of life . Copy number variants play a crucial role in driving evolutionary processes but can also cause genetic disease and cancer . Although copy number variants are important drivers of diversity , adaptation , and disease , the underlying dynamics of their formation and selection are poorly understood . Copy number variants are difficult to detect , especially when present at low frequencies in heterogenous evolving populations . To overcome this challenge , we developed a novel fluorescent reporter that allows us to visualize copy number variants as they emerge and to track them throughout hundreds of generations of laboratory evolution . We show that copy number variants arise early and repeatedly , that they are diverse in size and copy number , and that they are generated at a high rate , leading to competition among cells containing different copy number variants . Molecular characterization of copy number variants indicates that many of them are likely generated by errors during DNA replication . This method is broadly applicable to studying the molecular mechanisms underlying formation of copy number variants , as well as their role in driving evolutionary processes and cancer . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"ecology",
"and",
"environmental",
"sciences",
"computational",
"biology",
"cloning",
"fluorescence-activated",
"cell",
"sorting",
"fungal",
"evolution",
"copy",
"number",
"variation",
"molecular",
"biology",
"techniques",
"evolutionary",
"adaptation",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"ecological",
"metrics",
"genome",
"complexity",
"mycology",
"bioinformatics",
"species",
"diversity",
"molecular",
"biology",
"genetic",
"loci",
"spectrophotometry",
"cytophotometry",
"ecology",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"evolutionary",
"processes",
"spectrum",
"analysis",
"techniques"
]
| 2018 | Single-cell copy number variant detection reveals the dynamics and diversity of adaptation |
The Dot/Icm type IVB secretion system ( T4BSS ) is a pivotal determinant of Legionella pneumophila pathogenesis . L . pneumophila translocate more than 100 effector proteins into host cytoplasm using Dot/Icm T4BSS , modulating host cellular functions to establish a replicative niche within host cells . The T4BSS core complex spanning the inner and outer membranes is thought to be made up of at least five proteins: DotC , DotD , DotF , DotG and DotH . DotH is the outer membrane protein; its targeting depends on lipoproteins DotC and DotD . However , the core complex structure and assembly mechanism are still unknown . Here , we report the crystal structure of DotD at 2 . 0 Å resolution . The structure of DotD is distinct from that of VirB7 , the outer membrane lipoprotein of the type IVA secretion system . In contrast , the C-terminal domain of DotD is remarkably similar to the N-terminal subdomain of secretins , the integral outer membrane proteins that form substrate conduits for the type II and the type III secretion systems ( T2SS and T3SS ) . A short β-segment in the otherwise disordered N-terminal region , located on the hydrophobic cleft of the C-terminal domain , is essential for outer membrane targeting of DotH and Dot/Icm T4BSS core complex formation . These findings uncover an intriguing link between T4BSS and T2SS/T3SS .
Pathogenic bacteria transport functional proteins , such as effector proteins and exotoxins , across bacterial membranes . These bacterial proteins interact with host proteins to manipulate host cellular functions . Therefore , the secretion process plays a central role in bacterial pathogenesis . To accomplish this , bacteria have evolved various secretion systems . The type II secretion system ( T2SS ) is specialized to export periplasmic protein substrates , such as cholera toxin of Vibrio cholerae and heat-labile enterotoxin of enterotoxigenic Escherichia coli ( ETEC ) , across outer bacterial membranes to the extracellular milieu [1] , [2] , [3] , [4] , [5] . The type III secretion system ( T3SS ) is a protein-transport mechanism that translocates cytoplasmic substrates directly into the host cytoplasm . It plays a critical role in pathogenesis for a number of important bacterial pathogens , including enteropathogenic E . coli ( EPEC ) [6] . T3SS is ancestrally related to the bacterial flagellar system , and its core apparatus has a characteristic structure , often referred as “needle complex” [7] , [8] . The type IV secretion system ( T4SS ) is related to the conjugation system , and is apparently a very versatile secretion system for biological macromolecules [9] , [10] . For example , the Agrobacterium tumefaciens VirB/VirD system , one of the best studied T4SSs , is able to transport DNA-protein complex ( T-DNA ) into host cells . Bordetella pertussis secretes periplasmic pertussis holotoxin across outer membrane via the Ptl T4SS [11] , [12] . Many intracellular pathogens , including Legionella pneumophila , translocate a large array of effector proteins to the host cytosol using T4SSs [13] , [14] . T4SSs are further divided into two subgroups , type IVA ( T4ASS ) and type IVB ( T4BSS ) [15] , [16] . These two subgroups of T4SS are not related to each other at sequence level—with some exceptions , including secretion ATPases VirB11/DotB [17] , [18] , [19] . T4ASS is related to the conjugation systems of plasmids RP4 , R388 and pKM101 , and found in a number of bacterial pathogens , including plant pathogen A . tumefaciens . T4BSS is related to the conjugation systems of plasmids ColIb-P9 and R64 , and was originally found in human pathogen L . pneumophila [19] , [20] , [21] . L . pneumophila are gram-negative bacteria ubiquitous in fresh water and soil environments [22] , [23] . L . pneumophila infect and replicate within a wide variety of phagocytic eukaryotic cells , ranging from unicellular amoeba to human macrophages . In cases of human infection , L . pneumophila infection can result in a severe form of pneumonia known as Legionnaires' disease . Intracellular replication of L . pneumophila requires functional Dot/Icm T4BSS , irrespectively of host species [18] , [19] . It has been well established that L . pneumophila translocate more than 100 effector proteins into host cytoplasm using the Dot/Icm T4BSS [13] . The zoonotic pathogen Coxiella burnetii and the anthropod pathogen Rickettsiella grylli are some of known closest relatives to Legionella , and both carry T4BSSs closely related to that of Legionella [21] , [24] , [25] , [26] . A growing body of bacterial genomic information now suggests that over 20 pathogenic and environmental bacteria carry T4BSSs ( Figure S1 ) . Recent studies demonstrated that the pKM101 conjugation system , a T4ASS , has a lantern-shaped core complex composed of three proteins TraN/VirB7 , TraO/VirB9 and TraF/VirB10 [27] , [28] . This core complex spans both inner and outer membranes , but its structure is different from other double membrane-spanning secretion systems in architecture and composition . No other double membrane-spanning complex has been isolated and characterized from both T4ASS and T4BSS . However , a putative core complex of the Dot/Icm T4BSS has been suggested through a biochemical study of component proteins [29] . The complex is supposed to contain two outer membrane lipoproteins , DotC and DotD [30] , two inner membrane-spanning proteins , DotF and DotG , and one outer membrane-associated protein , DotH . In the absence of other components of the Dot/Icm T4BSS DotH remains unassociated with outer membrane , while lipoproteins DotC and DotD are targeted to outer membrane ( [29] and Figure S2 ) . The outer membrane targeting of DotH depends on lipoproteins DotC and DotD , presumably in a manner analogous with the function of pilotin in targeting secretin inT2SS [29]: the outer membrane lipoprotein pilotin is required for the stabilization and outer membrane targeting of the secretin [31] , [32] . The outer membrane lipoprotein VirB7 of Agrobacterium T4ASS forms a heterodimer with a core component , VirB9 , and stabilizes several VirB proteins , including VirB9 [33]; these are thought to be initial steps in assembling the T4ASS complex [9] . Lately , it has been demonstrated that the lipidation site cysteine of pKM101 TraN/VirB7 is essential to the outer membrane association of pKM101 T4ASS core complex , suggesting that TraN/VirB7 has pilotin-like function in terms of outer membrane targeting of the core complex [28] . Thus , the outer membrane lipoproteins DotC/DotD of T4BSS , T2SS pilotins and T4ASS VirB7 seem to play some overlapping role in the secretion apparatus assembly , although they are dissimilar in size and amino acid sequences . To clarify the structure and the molecular mechanism of Dot/Icm T4BSS , we crystallized DotD without the first 20 amino acids which contain the signal sequence for secretion across inner membranes and the lipidation site , Cys20 ( DotDΔN , Figure 1A ) , and determined the structure at 2 . 0 Å resolution .
DotDΔN is composed of the compact C-terminal domain ( DotD domain ) and the N-terminal disordered region ( Figure 1 ) . The N-terminal third of DotDΔN was invisible in the electron density map , except for a short β-strand ( Ala-37 to Ala-42 ) that we call the “lid . ” The N-terminal disordering was confirmed with limited digestion experiments using trypsin or V8 protease ( Figure S3 ) . Sub-stable products of digestion were identified by mass spectroscopic analysis . All preferential cleavage sites ( residues 25 , 55 , 58 and 61 ) are found in the N-terminal region , suggesting the N-terminal region is readily accessible to proteases used in this analysis . In addition , all three methionine residues of DotDΔN are in the disordered N-terminal region ( shown red letters in Figure 1C ) . These explain why we were not able to solve the structure by single-wavelength anomalous dispersion ( SAD ) phasing with the SeMet labeled DotDΔN crystals . Instead , we solved the structure using SAD phasing with Os derivative of DotDΔN crystals . The DotD domain forms a βαβ sandwich fold composed of two α-helices flanked by an antiparallel three-stranded β-sheet on one side and a mixed three-stranded β-sheet on the other side ( Figure 1B ) . The DotD domain does not show structural similarity to T4ASS VirB7 [27] , [34] . DaliLite database search [35] showed that the DotD domain has striking structural similarity to the N-terminal subdomain of secretins , which are outer membrane pore-forming proteins of T2SS and T3SS ( Figure 2 ) . The N0 domain of the ETEC GspD ( the protein database ( PDB ) ID 3ezj , mol-A ) , a T2SS secretin , is superimposable onto the DotD domain with a root-mean-square deviation ( rmsd ) of 2 . 1 Å ( Figure 2A ) . The T3S ( type III-specific ) domain of the EPEC EscC ( PDB ID 3gr5 , mol-A ) , a T3SS secretin , is superimposable onto the DotD domain with a rmsd of 2 . 5 Å ( Figure 2B ) . Sequence identity between the DotD domain and N0/T3S domains is only 6 . 4% and 10 . 6% for 78 and 75 amino acids , respectively . In fact , these domains have never been implicated as a conserved domain at the amino acid sequence level . Thus the DotD domain and the N0/T3S domains of the type II/III secretins share remarkable structural homology , although they are poorly related in amino acid sequences . The most remarkable difference between DotD and the N0/T3S domains of secretins is that DotD has a lid that makes β-strand addition [36] to β-strands β1 and β3 . The lid is located on β1 and covers the hydrophobic cleft formed by α1 , α2 and β1 ( Figures 3 and S4 ) . The side chains of two hydrophobic residues in the lid ( Ile-39 and Leu-41 ) stick into the cleft . The hydrophobic nature of most secretin residues—corresponding to the DotD residues that form the hydrophobic cleft—is conserved ( Figure 2C , yellow shaded ) . However , some bulky side-chains ( Phe-5 , Phe-9 , Asn-23 and Tyr-51 of GspD , Tyr-32 , Ile-34 , Ile-44 and Asn-51 of EscC; shown in italics in Figure 2C and in dark blue in Figure S5 ) protrude inwards and fill the secretin subdomain clefts . As the chain connecting the lid to the DotD domain is invisible in the crystal , it is unclear whether the lid and the DotD domain are in the same molecule . The 16 amino-acid segment ( Ala-37 to Met-52 ) containing the lid is well conserved among DotD orthologs from closely related bacteria ( 50% identity , Figure 1C ) , compared with other segments in the N-terminal region of DotDΔN . To evaluate the biological function of the lid , we constructed a lid mutant ( DotDAA ) , carrying alanine substitutions of both Ile-39 and Leu-41 residues , the side chains which stick into the cleft . The mutant protein was expressed at a level similar to that of wild-type protein in culture-grown L . pneumophila , and was equally targeted to bacterial outer membrane ( Figure 4A ) . Similarly , the expression and the localization of another lipoprotein , DotC , were not affected by the DotD mutation ( Figure 4A ) . In contrast , outer membrane targeting of DotH was abrogated in L . pneumophila producing the DotDAA mutant as much as in the DotD deletion strain ( Figure 4A ) . L . pneumophila strains producing single mutants DotDI39A or DotDL41A behaved like isogenic wild-type strain ( Figure S6 ) , suggesting that alanine substitutions of both residues are required for the defect . Furthermore , immunoprecipitation analyses indicate that all interactions between putative core components DotC , DotD , DotF , DotG and DotH were severely impaired in the L . pneumophila producing the mutant DotDAA ( Figure 4B ) . It should be noted that the apparent difference in mobility in SDS gel between the wild-type and the mutant DotD proteins is due to the intrinsic property of these proteins , because the purified mutant protein from overexpressing E . coli showed the same anomaly in the gel motility . These data indicate that the lid plays a significant role in the assembly process of the core complex of the Dot/Icm T4BSS . However , it remains unclear whether the lid functions directly through interaction with the DotD domain , or with other partners such as DotC and DotH , or indirectly through opening the cleft in the DotD domain . Secretins form a protein family that participates in several macromolecule translocation processes across bacterial outer membranes [37] , [38] , including type II and type III secretion , type IV pilus biogenesis and filamentous phage extrusion . Secretins extracted from membranes are multimeric and have a stacked-ring structure of cylindrical shape . Cryo-electron microscopic analyses of various secretins suggest that secretin rings have 12- or 14-fold rotational symmetry [39] , [40] , [41] . Their protease-resistant C-terminal domains contain single well-conserved secretin domains ( red boxes in Figure 5A ) , which embed in bacterial outer membranes . The N-terminal region of secretins extends into the periplasm and may interact with inner membrane partner proteins as well as substrates [42] , [43] . The N-terminal region is less conserved and always contains one N-terminal domain , which is related to the TonB-dependent outer membrane receptor domain , followed by one or more repeats of domains with the so-called “KH-fold” ( green and blue boxes in Figure 5A ) . The N0/T3S domains are the secretin subdomain closest to the N-terminal that follow the signal sequences for secretion across inner membranes via the Sec machinery . Crystal structures of the periplasmic regions of GspD [44] and EscC [45] secretins , spanning the N0/T3S and the Secretin_N domains , were not captured as multimers of cylindrical shape . However , taken together with available electron micrographic data , it has been suggested that these domains of secretins form periplasmic rings underneath the outer membrane rings . These findings imply that the DotD domain may form a periplasmic ring that is a part of a higher order complex spanning the outer membrane , plausibly composed of DotC , DotD and DotH ( Figure 5B ) . Our attempts to biochemically isolate a complex containing DotD has not succeeded yet , whereas it is possible to explore the propensity of DotD to form ring using DotD atomic coordinates obtained by this study and an in silico approach . To this end we used SymmDock , an algorithm for prediction of complexes with rotation symmetry by geometry based docking [46] , [47] Similar approach using SymmDock has been used for constructing the ring model for GspD periplasmic domain[44] . SymmDock predicted reasonable ring structures of the DotD domain ( without the lid ) , irrespective of the assumption of rotation symmetry , C12 or C14 ( Figure 5C and 5D , respectively ) . These ring models look alike in terms of monomer arrangement in the complexes . The analysis of amino acid conservation pattern using the ConSurf server [48] showed that many conserved residues of both ring models are found at the inner surface and the monomer interface of the DotD domain , while most variable residues are found at the outer surface of the ring ( Figure S7 ) . The cleft , the lid-associating site of the DotD domain , comes near the periplasmic surface in the ring models ( Figure S8 ) . Therefore , it may affect the interactions between DotD and inner membrane components of the Dot/Icm T4BSS or its substrates . Alternatively , the lid may not be in place and extend towards the outer membrane when the ring is formed . This would be a reasonable possibility given that the N-terminal residue of the DotD domain ( Ser-66 ) is situated at the outer membrane surface of the ring models and the N-terminus of DotD must be anchored in the outer membrane . The crystal structure of the outer membrane complex of the pKM101 T4ASS , containing full-length TraN/VirB7 , was recently reported [27] . The TraN/VirB7 in the complex takes an extended conformation and wraps around the outer membrane complex . Mature TraN/VirB7 is a small peptide , 33 residues long , comparable in size to the N-terminal disordered region of DotD ( 46 amino acids ) . Along these lines , the N-terminal disordered region of DotD , which contains the lid and the conserved segment ( Ala-37 to Met-52 ) , may interact with outer membrane components such as DotC and DotH . In summary the DotD domain might have the propensity to form a ring like the EscC T3S and the GspD N0 domains; however , the ring models must be validated by future experimental confirmation . There are several intriguing parallels between secretins and DotC , DotD and DotH , aside from being outer membrane components essential for bacterial secretion systems . The protease-resistant C-terminal domain , representing about two thirds of DotH , is predicted to be rich in β strands using PHDsec [49] ( NN and HN , unpublished ) , which is commonly true of integral outer membrane proteins such as secretins . Lipoproteins DotC and DotD are required for the outer membrane targeting of DotH; likewise pilotins are required for the outer membrane targeting of cognate secretins . Together with the remarkable structural similarity between DotD and a periplasmic subdomain of secretins , it is possible that the putative complex of DotC , DotD and DotH is a secretin counterpart of T4BSS ( Figure 5B ) . Moreover , there is another link between the secretion ATPases of T4BSS and T2SS . Unlike other T4BSS components , secretion ATPase DotB shares sequence-level similarity with ATPases of T2SS , T4ASS , and the type IV pilus biogenesis system ( T4PBS ) which is closely related to T2SS . Phylogenetic analysis of these ATPases showed DotB to be closely related to the T4PBS ATPase , PilT [17] . In fact , DotB was found in the major group consisting of T2SS and T4PBS ATPases in the phylogenetic tree inferred for secretion ATPases . Importantly , this group is distinct from the major group in which T4ASS ATPases ( VirB11 ) are found . Collectively , the data from structural and phylogenetic analyses raise the rather unexpected possibility that the architecture of T4BSS machinery shares similarity with T2SS to a certain extent , and can be significantly different from T4ASS . In conclusion , the present study revealed that the structurally conserved DotD/N0/T3S domain is widely spread throughout outer membrane complexes of even distantly related secretion systems , including T2SS , T3SS and T4BSS . Although the DotD ring models must be validated by future experimental studies , the finding raised the possibility that transport machinery of T4BSS may adopt mosaic architectures of T4ASS and T2SS/T3SS . Future elucidation of structures and functions of bacterial secretion apparatus will give new insights into the molecular mechanism of protein transport across membranes—a central process essential for bacterial pathogenesis .
Bacterial strains and plasmids used in this study are provided in Table S1 . L . pneumophila strains defective in dotD as well as L . pneumophila strains carrying the M45 epitope tagged dotF or the dotDAA ( I39AL41A ) mutation were constructed by allelic exchange [50] . Rabbit anti-sera against DotC , DotG , DotH and M45 epitope were raised by immunization of KHL-conjugated synthesized peptides CMDYVKPEAPNVTLLPKTKA ( DotC ) , CWKQVETQVYTEGTEETK ( DotG ) , CYGPNAKSMPTEEGIPPS ( DotH ) , and CDRSRDRLPPFETETRIL ( M45 ) , respectively . Polyclonal antibodies were purified from the anti-sera by affinity chromatography using peptide-conjugated SulfoLink resins ( Pierce ) . Antibodies against DotA ( mAb2 . 29 ) and DotD were described previously [51] , [52] . E . coli cells overproducing DotDΔN with a hexa-histidine tag were collected by centrifugation and resuspended with 50 mM Tris-HCl pH 7 . 5 , 50 mM NaCl , 1 mM EDTA containing Complete Protease Inhibitor Cocktail ( Roche Diagnostics ) . Cells were disrupted , centrifuged ( 30 , 000×g , 20 min ) , and the soluble fraction was loaded on a SP sepharose column ( GE Healthcare ) . His-tagged DotDΔN was eluted by a step gradient of NaCl in 20 mM Tris-HCl pH 7 . 5 and was loaded on a HisSelect column ( Sigma-Aldrich ) . His-tagged DotDΔN was eluted by a step gradient of imidazole in 20 mM Tris-HCl pH 7 . 5 , 150 mM NaCl . Peak fractions were pooled and dialyzed against 20 mM Tris-HCl pH 7 . 5 , 200 mM NaCl . After removal of the His-tag by thrombin digestion , DotDΔN was loaded onto a HiLoad Superdex 75 gel filtration column ( GE Healthcare ) . Purified protein was eluted in 20 mM Tris-HCl pH 7 . 5 and concentrated using a Vivaspin 20 concentrator ( Sartrius ) . Se-Met DotDΔN was purified using the procedure described above . Crystals suitable for X-ray analysis were obtained at 4°C using the sitting-drop vapor-diffusion method . I23 crystals of DotDΔN with unit cell dimensions a = b = c = 103 . 9 Å were grown from drops prepared by mixing 1 µl protein solution ( 11 . 2 mg/ml ) with 1 µl reservoir solution containing 8% ( v/v ) PEG 8000 and 0 . 1 M CHES-NaOH pH 10 . 0 . Crystals of Se–Met-labeled protein grown under the same conditions as the native crystals were also obtained . Crystals were soaked in a solution containing 90% ( v/v ) of the reservoir solution and 10% ( v/v ) MPD for a few seconds , and then immediately transferred into liquid nitrogen for freezing . X-ray diffraction data were collected at a synchrotron beamline BL41XU of SPring-8 ( Harima , Japan ) with the approval of the Japan Synchrotron Radiation Research Institute ( JASRI ) . The data were recorded under nitrogen gas flow at 90 K , and under He gas flow at 40 K , for native and derivative crystals , respectively . The data were processed with MOSFLM [53] and scaled with SCALA [54] . Initially , we tried to solve the structure using the anomalous data of Se-Met derivative crystals , but the anomalous signal was too weak to determine the phase . Therefore , we prepared Os derivative crystals by soaking the crystals into a reservoir solution containing K2OsCl6 at 50% saturation for four hours . Initial SAD phase was calculated from the anomalous diffraction data of the Os derivative crystal using SOLVE [55] . The phase was improved and extended to 2 . 0 Å with DM using a native data set . The model was constructed with COOT [56] , and was refined to 2 . 0 Å using program CNS [57] . A 5% fraction of the data was excluded from the data for the R-free calculation . During the refinement process , iterative manual modifications were performed using an “omit map . ” The refinement converged to an R factor of 22 . 5% and a free R factor of 24 . 7% . The Ramachandran plot indicated that 91 . 1 % and 8 . 9 % residues were located in the most favorable and allowed regions , respectively . Data collection and refinement statistics are summarized in Tables 1 and 2 . Fractionation of L . pneumophila membranes was carried out essentially as described [58] . Briefly , 20 ml of bacterial culture grown in AYE for 18 h with starting OD600 = 0 . 1 was centrifuged and bacterial pellets were suspended with 5 ml of 10 mM HEPES pH 7 . 4 , 20% ( w/w ) sucrose . After the addition of RNaseA ( final 10 µg/ml ) , DNaseI ( 2 µg/ml ) and phenylmethylsulfonyl fluoride ( PMSF , 1 mM ) , cells were lysed by two passages of a chilled French Pressure Cell ( Thermo Scientific ) ; unlysed cells were removed by centrifugation ( 5 , 000 g , 15 minutes ) . EDTA was added to the lysate to a final concentration of 5 mM . A crude membrane fraction was obtained by loading 1 . 6 ml of the bacterial lysate onto a two-step gradient consisting of 0 . 8 ml of a 60% sucrose cushion and 2 . 5 ml of 25% sucrose in 10 mM HEPES . The membranes were pelleted for 3 . 5 h at 40 , 000 rpm in an SW50 . 1 rotor at 4°C . A membrane layer visible on top of the 60% sucrose was extracted and diluted to <25% sucrose with 10 mM HEPES . The crude membrane fraction was separated by isopycnic sucrose density gradient centrifugation using a gradient consisting of a 0 . 5-ml cushion of 60% sucrose and layers of 1 ml of 55% sucrose , 2 . 4 ml each of 50% , 45% , and 40% sucrose , 1 . 4 ml of 35% sucrose , and 1 ml of 30% sucrose in 10 mM HEPES . Approximately 1 . 2 ml of the crude membrane prep was placed on top of the gradient and centrifuged in an SW41 rotor at 37 , 000 rpm for 16 h at 4°C . Fractions ( 0 . 75 ml ) were collected and analyzed by western immunoblotting using antibodies against DotC , DotD and DotH . Levels of the 28-kDa Major outer membrane protein ( MOMP ) and DotA were determined by Coomassie blue staining and western immunoblotting , respectively . Bacterial cells grown as described above in 10 ml AYE medium were washed once with cold PBS and resuspended in 10 ml of PBS . A cleavable crosslinker , dithiobis ( succinimidyl propionate ) ( DSP; final concentration: 0 . 08 mM ) was added to the suspension and incubated for 2 h on ice . The crosslinking reaction was stopped by addition of Tris pH 8 . 0 ( final 120 mM ) and cells were resuspended with 5 ml of 50 mM Tris-HCl pH 8 . 0 , 150 mM NaCl , 1 mM EDTA . A suspension was mixed with lysozyme ( 1 . 25 mg/ml ) and incubated for 30 minutes on ice . After addition of RNaseA , DNaseI and PMSF , a whole cell lysate was prepared as described above . Total proteins in 1 ml of the lysate were precipitated by final 10% TCA . The precipitates were washed three times with acetone , and dissolved in 100 µl of 50 mM Tris-HCl pH 8 . 0 , 1% SDS , 1 mM EDTA . After the heat treatment ( 100°C for 3 minutes ) , 30 µl of denatured lysate was diluted with 1 ml of Triton buffer ( 2% Triton X-100 , 50 mM Tris pH 8 . 0 , 150 mM NaCl , 0 . 1 mM EDTA ) . Insoluble materials were removed by centrifugation ( 14 , 000 rpm , 20 minutes ) and the supernatant fraction was transferred to a new tube . Indicated antibody ( 0 . 1 µg ) was added to a supernatant and incubated at 4°C for overnight with gentle rotation . Protein A resin ( GE Health Care , 10 ul of 50% suspension in Triton buffer ) was added to the mixture and incubated for further 2 h with rotation . Resins were washed twice with Triton buffer , once with 50 mM Tris-HCl pH 8 . 0 . Immunocomplexes were extracted with 50 µl of sample buffer containing a reducing agent and analyzed by western immunoblotting . The atomic coordinates have been deposited in the Protein Data Bank , www . pdb . org ( PDB ID code 3ADY ) . | Bacterial pathogens deliver virulence factors such as exotoxins and effector proteins to host cells . To accomplish this bacteria utilize specialized secretion systems such as type III and type IV secretion systems . The type IV secretion systems ( T4SS ) play a central role in pathogenesis by many important pathogens including Agrobacterium tumefaciens , Helicobacter pylori and Legionella pneumophila . T4SS is ancestrally related to the bacterial conjugation system and is divided into two subgroups , type IVA ( T4ASS ) and type IVB ( T4BSS ) , which are derived from distinct conjugation systems . In spite of its pivotal role in bacterial pathogenesis , the structural bases and molecular mechanisms of the type IVB secretion still remain largely unknown . Here we show the crystal structure of DotD , one of the core components of Legionella T4BSS . Surprisingly , the structure of DotD is not related to those of T4ASS core components . In contrast , the structure of DotD is remarkably similar to that of a subdomain of secretin family proteins , which form substrate conduits for other types of secretion systems . This finding provides intriguing insights into the nature and the evolution of bacterial secretion systems essential for pathogenesis . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
]
| [
"infectious",
"diseases/bacterial",
"infections"
]
| 2010 | Crystal Structure of Legionella DotD: Insights into the Relationship between Type IVB and Type II/III Secretion Systems |
The Long interspersed nuclear element 1 ( LINE-1 ) is a primary source of genetic variation in humans and other mammals . Despite its importance , LINE-1 activity remains difficult to study because of its highly repetitive nature . Here , we developed and validated a method called TeXP to gauge LINE-1 activity accurately . TeXP builds mappability signatures from LINE-1 subfamilies to deconvolve the effect of pervasive transcription from autonomous LINE-1 activity . In particular , it apportions the multiple reads aligned to the many LINE-1 instances in the genome into these two categories . Using our method , we evaluated well-established cell lines , cell-line compartments and healthy tissues and found that the vast majority ( 91 . 7% ) of transcriptome reads overlapping LINE-1 derive from pervasive transcription . We validated TeXP by independently estimating the levels of LINE-1 autonomous transcription using ddPCR , finding high concordance . Next , we applied our method to comprehensively measure LINE-1 activity across healthy somatic cells , while backing out the effect of pervasive transcription . Unexpectedly , we found that LINE-1 activity is present in many normal somatic cells . This finding contrasts with earlier studies showing that LINE-1 has limited activity in healthy somatic tissues , except for neuroprogenitor cells . Interestingly , we found that the amount of LINE-1 activity was associated with the with the amount of cell turnover , with tissues with low cell turnover rates ( e . g . the adult central nervous system ) showing lower LINE-1 activity . Altogether , our results show how accounting for pervasive transcription is critical to accurately quantify the activity of highly repetitive regions of the human genome .
Long interspersed nuclear element 1 ( LINE-1 ) has attracted much attention in the last decade due to its capacity to promote genetic plasticity of the human genome . LINE-1 is a DNA sequence capable of duplicating itself and other DNA sequences by mobilizing messenger RNAs ( mRNAs ) to new genomic locations via retrotransposition [1–3] . There are multiple molecular mechanisms to deactivate LINE-1 instances , most prominently , the truncation of 5’UTR due to partial retrotransposition has resulted in mostly inactive and truncated copies of LINE-1 across the human genome[3–6] . Truncated copies of LINE-1 lack their internal promoter sequence and therefore , are expected to be dead-on-arrival . Although full-length LINE-1 activity has been described in both healthy and pathogenic tissues [3 , 7 , 8] , quantifying its activity is remarkably difficult due to its repetitive nature . Until recently , LINE-1 retrotransposition was believed to occur in germ cells [9–11] and tumors [12–14] , but not in somatic tissues . However , growing evidence suggests that LINE-1 is active in the neuroprogenitor cells and in other healthy somatic tissue at low levels [15–18] . As opposed to healthy tissues , tumor and tumor derived cell lines show higher levels of LINE-1 activity [13] . LINE-1 instances are likely to be activated due to broad demethylation of LINE-1 promoter [19] . The literature describes many other factors contributing to the constraints of LINE-1 activity pre- and post-transcriptionally [20]; however , little is known about its activation and impact in tumors [21] . A major challenge to asses LINE-1 activity is the requirement of either specialized assays [22 , 23] or multiple and complementary datasets [24] , hindering estimation of autonomous LINE-1 transcription in a large number of samples . Moreover , affordable methods to quantify LINE-1 activity , such as those based on RNA [17 , 25 , 26] , are largely confounded by the high copy number nature of LINE-1 and pervasive transcription [23] , which refers to the idea that the majority of the genome is transcribed , beyond just the boundaries of known genes [27] . How much pervasive transcription influences the human transcriptome is still unclear [27–29] . Some researchers suggest that pervasive transcription is mostly derived from technical and biological noise and , therefore , might not be relevant in RNA sequencing experiments [30] . Others suggest that pervasive transcription has a stochastic nature , and if sequenced at enough depth the majority of the genome may be transcribed . With either theory , pervasive transcription should not affect quantification of the transcription of protein coding genes , which are present either in single copy or low copy numbers in the genome . However , the quantification of the transcriptional activity of transposable elements , including LINE-1 , would be greatly affected by pervasive transcription due to their multi-copy nature . The autonomous transcription of LINE-1 , on the other hand , derives from LINE-1 transcripts being fully transcribed from its internal promoter . Thus , by definition , since LINE-1 promoters are at the 5’ extremity of LINE-1 elements , autonomous transcription is more likely to derive from full length LINE-1 instances . These transcripts could derive from both intronic or intergenic full-length LINE-1 instances . This paper presents a new method to remove the effect of pervasive transcription on RNA sequencing datasets and reliably quantify LINE-1 subfamily transcriptional activity . We first show that the vast majority of reads overlapping LINE-1 elements are derived from pervasive transcription and propose a method to address this issue . We validated the LINE-1 transcription landscape in well-established human cell lines and their cell compartments . Finally , we surveyed LINE-1 activity in a variety of healthy somatic tissues . Although somatic retrotransposition has been mainly studied in the human brain , we found surprisingly little transcriptional activity in most brain regions from adults . Instead , we found LINE-1 transcriptional activity in other somatic tissues consistent with an overall trend of LINE-1 activity in cell with higher turnover .
We benchmarked TeXP by estimating the autonomous transcription of LINE-1 subfamilies in RNA sequencing experiments of well-established human cell lines [31] . Fig 2A shows the proportion of reads mapped to LINE-1 subfamilies using a naïve method ( left panel ) and proportions of reads from each signature using TeXP ( right panel ) . TeXP estimations were also compared to other transposable element quantification pipelines ( S2 Fig ) . In the naïve method ( Fig 2A; left panel ) , cytoplasmic and whole-cell polyadenylated ( polyA ) + samples had an enrichment of reads mapping to L1Hs and L1PA2 when compared to whole-cell transcripts without a polyadenylated tail ( whole-cell polyA- ) and nuclear RNA samples . The enrichment of L1Hs reads was consistent with increased transcription of full-length L1Hs ( S3 Fig ) . The estimates after applying TeXP ( Fig 2A; right panel ) revealed two major signals in MCF-7 RNA sequencing experiments: pervasive transcription and L1Hs autonomous transcription . The difference between the naïve method and TeXP suggests that reads mapped to ancient LINE-1 subfamilies , such as L1PA3 and L1PA4 , are mostly derived from pervasive transcription . TeXP also detected residual L1PA2 transcription in a small number of samples ( Fig 2A and S4 Fig ) . This result is consistent with L1Hs and L1PA2 being the only two LINE-1 subfamilies capable of autonomous transcription and autonomous mobilization in the human germline and tumors [11 , 33] . MCF-7 , a cell line derived from breast cancer , was previously described as having remarkably high levels of L1Hs autonomous transcription [17 , 24] . The transcriptome of MCF-7 and many other cell lines were carefully and consistently sequenced through the Encyclopedia of DNA elements ( ENCODE ) project . Leveraging these ENCODE cell line datasets , we assessed L1Hs autonomous transcription in distinct cell compartments ( S5 Fig and S6 Fig ) [31] . First , we found that MCF-7 whole-cell polyA+ samples had extremely high levels of L1Hs transcription ( 180 . 7 RPKM ) , in agreement with the literature . Selecting whole-cell polyA- samples reduced the signal of L1Hs autonomous transcription by 73% ( Fig 2A ) , suggesting that most of the signal was derived from mature polyA+ LINE-1 transcripts . Furthermore , we tested whether L1Hs transcripts are derived from cytoplasmic ( mature ) or nuclear ( pre-mRNA ) portions of the cell . We found that nuclear transcripts were highly enriched for pervasive transcription ( autonomous/pervasive ratio 0 . 02 ) , whereas cytoplasmic transcripts had an autonomous/pervasive ratio similar to transcripts derived from whole-cell polyA+ samples ( 0 . 45 and 0 . 51 , respectively–Fig 2A ) . Together , these results suggest that most of the LINE-1 autonomous transcription signal is derived from mature transcripts in the cytoplasm and only a small fraction of signal is derived from fragmented LINE-1 transcripts in the nucleus . Analyzing other lymphoblastic and cancer-derived cell lines such as GM12878 , SK-MEL-5 and K-562 yielded no evidence of L1Hs autonomous transcription in most cell compartments or RNA fractions , despite low levels of L1Hs autonomous transcription in whole-cell polyA+ samples ( 0 , 8 . 8 and 8 . 4 RPKM , respectively . Fig 2B and S1 Table ) . To validate the quantification of L1Hs autonomous transcription , we performed droplet digital PCR ( ddPCR ) to estimate autonomous and pervasive transcription levels on a reference panel of six cell lines: MCF-7 , K-562 , HeLa , HepG2 , SK-MEL-5 , and GM12878 . For these experiments , we assumed that expression on the 5’ end of the L1Hs transcript can be used as an approximation to autonomous transcription due to the large imbalance of 5' truncated and full-length copies . The expression on the 3’ end , on the other hand , is an approximation to the combination of autonomous and pervasive transcription . We initially designed and tested multiple assays targeting different regions of the L1Hs locus and proceeded with the two best performing assays ( S2 Table ) . The first assay targeted ORF1 , directly adjacent to the 5’UTR , representing the 5’ end of the transcript . The second assay targeted ORF2 about 1 . 5 kb upstream of the 3’ UTR , representing the 3’ end of the transcript . We completed the same design process for ORF2 to find the copy numbers of the truncated L1Hs transcripts ( i . e . , the transcripts missing the 5’ end of L1Hs ) ( Fig 2C , S3 Table ) . Since autonomous transcription results in an enrichment full-length transcript of L1Hs , we estimated an approximation to the level of autonomous transcription as the pervasive transcription created by full-length transcription subtracted from the expression of the 5’ end ( ORF1 ) . Fig 2D shows the relative quantification of L1Hs transcripts in these four cell lines using the HPRT1 5’ end as a reference . We estimate that the ddPCR analysis detected approximately 2 , 412 . 8 copies of autonomously transcribed transcripts/ng in MCF-7 cells . In agreement with our in-silico result , K562 and SK-MEL-5 had 990 . 0 and 1 , 075 . 6 copies of autonomously transcribed transcript/ng , respectively . For the GM12878 cell line , we expected to find no autonomous expression of L1Hs; however , our ddPCR assays detected low levels of autonomous transcription of L1Hs ( 233 . 0 copies of autonomously transcribed transcript/ng; Fig 2D , S3 Table ) . Overall , the quantification of L1Hs autonomous transcription using ddPCR was highly correlated with the quantification using TeXP ( Spearman correlation , rho = 0 . 9 , p-value = 0 . 01394 , S7 Fig ) . This suggests that TeXP can remove most of the noise derived from pervasive transcription , although it is insensitive to samples with little LINE-1 autonomous transcription ( S8 Fig ) . To address TeXP sensitivity in relation to noise we first tested TeXP under an ideal experimental setup . In this simulation , the number of observed reads overlapping L1 subfamilies is a simple combination of known proportions of these signals ( i . e . pervasive and autonomous transcription ) . For example , we simulate read counts where 30% of the reads derive from LIHs autonomous transcription and the remaining 70% derive from pervasive transcription . We use simulated read counts as input to TeXP and calculate the root mean square error ( rmse ) between the known proportions and estimated proportions . As observed in S15 Fig ( solid line ) , under this condition , the TeXP estimations of read counts is nearly identical to the simulated read counts ( median ( rmse ) = 0 . 0003 ) . We also tested the effect of noise in the TeXP estimations . To that end , we modeled the reads that cross-map across LINE-1 subfamilies as a noise component as a Poisson process deriving from random sequencing of cDNA fragments . For these simulations , the Poisson noise was added to read counts deriving from of known proportions of signals as described above . We tested two scenarios , one with noise equivalent to 10% of all observed reads ( dashed line ) and 20% of the of all observed reads ( long dashed lines ) . In these conditions , the median TeXP RMSEs were respectively 0 . 014 and 0 . 028 . Overall our observations suggest that , under low noise conditions , TeXP should be able to detect autonomous L1Hs transcription even when L1Hs has low transcription levels . However , we also describe the TeXP capacity to detect low levels of L1Hs autonomous transcription degrades with higher levels of noise in RNA-se data . In extreme situations , such as when 20% of the read counts are derived for a Poisson process , the RMSE can be up to 0 . 083 . Researchers have long thought that LINE-1 instances are completely silenced in most healthy somatic cells . LINE-1 is silenced by the methylation of its promoter [19] , which should preclude the transcription of mature LINE-1 mRNAs in healthy somatic tissue . To test whether LINE-1 subfamilies are completely silenced in somatic tissue , we analyzed LINE-1 transcription in 7 , 429 primary tissue samples from the Genotype-Tissue Expression ( GTEx ) project [32] ( S4 Table ) . Similar to the cell lines , we found that L1Hs was autonomously transcribed; L1P1 , L1PA2 , L1AP3 , and L1PA4 only had residual or spurious autonomous transcription in healthy tissues ( S9 Fig ) . Furthermore , we found that pervasive transcription was the major signal in most RNA sequencing datasets , accounting for 91 . 7% , on average , of the reads overlapping LINE-1 instances ( S10 Fig and S14 Fig ) . Overall , healthy tissues had a narrower range of L1Hs autonomous transcription levels than cell lines , with the peak transcription level of 47 RPKM ( Fig 2E ) versus 180 RPKM in the cell lines ( S1 Table ) . We found no or very little ( <1 RPKM ) evidence of L1Hs autonomous transcription in 2 , 520 ( 34 . 3% ) of the GTEx RNA sequencing experiments from primary tissues . Together , these results indicate that L1Hs is broadly transcribed in some healthy somatic tissues . Therefore , if post-transcriptional regulatory constraints do not completely silence LINE-1 activity , one could expect that LINE-1 to play an important role in creating genetic diversity across somatic cells within an individual . We then compared the landscape of LINE-1 subfamily transcription in Epstein-Barr virus ( EBV ) immortalized cell lines and their corresponding primary tissue to understand the changes induced by cell line immortalization . EBV immortalization causes drastic changes in the expression of cell cycle , apoptosis , and alternative splicing pathways [34–36] . Overall , we found that EBV-transformed cell lines derived from different tissues ( lymphoblastic and fibroblastic ) had distinct patterns of L1Hs autonomous transcription; lymphoblast ( blood-derived ) cell lines had no or little autonomous transcription of L1Hs ( S11 Fig ) with approximately 84% of samples having an estimated RPKM equal to zero , whereas fibroblastic ( skin-derived ) cell lines consistently had higher levels of L1Hs autonomous transcription ( median 1 . 5 RPKM ) with 58 . 7% of samples having an RPKM higher than 1 . In general , EBV-immortalized cell lines reflected their tissue of origin . While most ( 74 . 6% ) of the whole blood samples had no transcriptional activity of L1Hs , only one sample from skin had an L1Hs autonomous transcription level below 1 RPKM . We further selected patients with both primary and EBV-transformed cell lines to assess whether the EBV transformation could change L1Hs autonomous transcription . We found that both skin cells and lymphocytes had a drastic down-regulation of L1Hs autonomous transcription ( S12 Fig ) . This finding suggests that EBV-transformed cell lines partially preserve the L1Hs transcription level from their tissue of origin , potentially explaining why fibroblast-derived induced pluripotent stem cells support higher levels of LINE-1 retrotransposition [37] . Human solid tumors display increased levels of LINE-1 activity [38–40] . In order to assess if this increase is a result from pervasive transcription or an increase in autonomous transcription in LINE-1 we run TeXP in RNA-seq from healthy thyroid samples and solid thyroid tumors . We than calculate the distribution of number of reads deriving from pervasive and autonomous transcription . We first observed that there indeed a significant difference in the number of reads from LINE-1 elements ( S13 Fig ) . We also found that compared to healthy thyroid samples , tumor samples display higher levels of autonomous transcription and lower levels of pervasive transcription suggesting that autonomous transcription LINE-1 is driving most of the increase in LINE-1 expression in thyroid tumor samples . Human tissues show remarkable variability of L1Hs autonomous transcription . We found that L1Hs autonomous transcription is inversely correlated to the time it takes cells to divide ( cell turnover rate; Pearson correlation: cor = -0 . 6668968; p-value = 0 . 04 ) . No correlation was found between cell turnover and pervasive transcription ( Pearson correlation: cor = 0 . 3983474; p-value = 0 . 2883 ) . Tissues suggested to have low cell turnover , such as the human brain [41] , are amongst the tissues with the lowest levels of L1Hs autonomous transcription ( Fig 2E ) . In particular , the human cerebellum , which has no transcription of L1Hs , is likely to have strong repression of L1Hs autonomous transcription . This result seems to contradict the literature that suggests that the human brain supports high levels of somatic LINE-1 retrotransposition; however , most of these studies were based on neuro precursors that correspond to the early development stage of the human brain [15 , 42–44] . Conversely , brain samples extracted from the striatum , putamen , and caudate , all regions associated with the basal ganglia , had higher levels of L1Hs autonomous transcription compared to other brain regions ( T-test basal ganglia vs . all other brain tissues , t = -7 . 0943; p value = 9 . 867e-12 –Fig 2E ) ; importantly , these levels were still low compared to other tissues . Other tissues with low cell turnover rates , such as liver , pancreas , and spleen , also showed very little or no autonomous transcription of L1Hs ( 91 . 2% , 82 . 9% , 88 . 9% of samples , respectively , had a L1Hs RPKM < 1 –Fig 2E ) . Conversely , germinative tissues have been proposed to support somatic activity of L1Hs elements [45] . Our results suggest that this trend is more general , and most tissues associated with the reproductive system sustain higher levels of L1Hs autonomous transcription ( Fig 2E ) . In addition , we found that the tissues with the highest levels of L1Hs autonomous transcription were enriched for high cell turnover; these included the nerve ( tibia ) , skin ( both exposed and not exposed to the sun ) , prostate , lung , and vagina ( Fig 2E ) .
Prior to this study , the effects of pervasive transcription on the estimates of transposable elements activity were largely ignored . Here , we showed that most of the RNA-seq reads matching to LINE-1 instances derive from pervasive transcription , highlighting the importance of these effects . In order to account for the effect of pervasive transcription on the quantification of LINE-1 activity we developed TeXP , a method that uses the widespread nature and the mappability signatures of LINE-1 subfamilies to account for and remove the effects of pervasive transcription from L1Hs , L1P1 , L1PA2 , L1PA3 , and L1PA4 subfamilies . We compared TeXP estimates to other strategies such as naïve counts and others established methods such as SalmonTE [46] and TETranscript [47] ( S2 Fig ) . Our estimations suggest that the pervasive transcription component is frequently missed and can be a confounding factor in some quantifications , for example , we observed that over 75% of reads mapping to ancient LINE-1 subfamilies derive from inactive subfamilies ( Fig 2A and S14 Fig ) . We used TeXP to perform a comprehensive analysis of LINE-1 transcriptional activity across different cell types and somatic tissues . Previous studies suggest that LINE-1 is active in germline and tumor cells , but not in most normal somatic cells with the exception of hints of activity in neuro-precursor cells [21] . Somatic mosaicism of transposable elements , in particular LINE-1 , has been carefully characterized in the human brain and , despite some disagreement on the exact rate that LINE-1 retrotranspose in the human brain , it is clear that LINE-1 frequently create somatic copies in the human brain , in particular , in neuroprecursor , cortex and caudate nucleus cells [15 , 48 , 49] . The somatic mobilization in other tissues , nonetheless , remain obscure and lack a systematic investigation . A first step is to comprehensive characterize other tissues in terms of their LINE-1 transcriptome activity . Surprisingly , we found that LINE-1 was active in many healthy human tissues , particularly in epithelial cells . As we only detected a limited amount of LINE-1 activity in adult brain cells , our findings are in agreement with LINE-1 activity correlating with cell proliferation rate . We validated many aspects of TeXP using ddPCR probes designed to quantify pervasive and autonomous transcription of L1Hs across human cell lines . These results show that our method lacks the sensitivity to measure autonomous transcription . In particular , TeXP underestimates L1Hs transcription levels when the signal-to-noise ratio is small . One could imagine using unique regions of LINE-1 instances after removing the pervasive transcription signal to improve the transcriptional quantification . However , removing pervasive transcription from individual instances is not trivial and should be carefully investigated . As a tool , TeXP could be useful in several scopes beyond this study . Pervasive transcription should also affect the quantification of other transposable elements and repetitive regions accounting for large portions of the human genome . Our method could be further used to estimate the autonomous transcription levels of pseudogenes , SINEs ( ALUs ) and HERVs . Furthermore , TeXP could be used in model organisms to distinguish the effects of pervasive transcription beyond humans . The mouse genome , for example , has evidence of higher rates of retrotransposition but little is known about the activity of LINE-1 in somatic cells . Moreover , some of the results we describe here could be extended and uncover important biological insights . For example , assays such as induced pluripotent stem cells clones [50] , RC-seq [51] and L1-seq [52] , among others , could be used to carefully characterize of the rate of somatic retrotransposition in tissues with higher rates of autonomous transcription of L1Hs . Additionally , TeXP could be used to further investigate the biases found in individuals from different ancestral backgrounds [53] . The TeXP approach to quantify transcriptional activity by removing pervasive transcription could also be expanded to investigate the activity of LINE-1 during embryonic development or in pathological tissues , such as tumors .
TeXP models the number of reads overlapping L1 elements as the composition of signals deriving from pervasive transcription and L1 autonomous transcripts from distinct L1 subfamilies . Our model proposes that the number of reads overlapping L1Hs instances as described by the Eq 1: OL1Hs=T*GL1Hs*εpervasive+T*ML1Hs , L1Hs*εL1Hs+T*ML1Hs , L1PA2*εL1PA2+⋯+T*ML1Hs , j*εj Where OL1Hs is the observed number of reads mapping to L1Hs , T is the total number of reads mapped to L1 instances , GL1Hs defines the proportion of L1 bases in the genome annotated as L1Hs , εpervasive is the percentage of reads emanating from pervasive transcription , M is the mappability fingerprint ( defined bellow ) that describes what is the proportion of reads emanating from the signal j ∈ {L1Hs , L1P1 , L1PA2 , L1PA3 , L1PA4} that maps to L1 subfamily i ∈ {L1Hs , L1P1 , L1PA2 , L1PA3 , L1PA4} and ε is the percentage of reads emanating from the autonomous transcription L1 Subfamily j . This model can be further generalized as the Eq 2: Oi=T ( Giεpervasive+∑jMi , jεj ) We selected these five LINE-1 subfamilies based on the rates of cross-mappability of simulated data ( S3 Fig ) . In particular , we are interested in removing the effect of pervasive transcription from the estimates of L1Hs autonomous transcription . We simulated reads from L1Hs transcripts we observed that most ( >90% ) of the reads emanating from L1Hs autonomous transcription map to the L1Hs , L1PA2 , L1PA3 , L1PA4 and L1P1 subfamilies . Older subfamilies such as L1PA5 and L1PA6 , for example , correspond to less than approximately 5% of L1Hs cross-mappable reads . In contrast to L1Hs which only a fourth of the reads map back to L1Hs , older elements such as L1PA5 and L1PA6 have a much higher self-mapping rates , 60% and 70% respectively . Therefore , these subfamilies should be less affected by confounding factors deriving from pervasive transcription . The number of reads mapped to each subfamily Oi is measured by analyzing paired-end or single-end RNA sequencing experiments independently . TeXP extracts basic information from fastq raw files such as read length and quality encoding . Fastq files are filtered to remove homopolymer reads and low quality reads using in-house scripts and FASTX suite ( http://hannonlab . cshl . edu/fastx_toolkit/ ) . Reads are mapped to the reference genome ( hg38 ) using bowtie2 ( parameters:—sensitive-local -N1—no-unal ) . Multiple mapping reads are assigned to one of the best alignments . Reads overlapping LINE-1 elements from Repeat Masker annotation of hg38 are extracted and counted per subfamily . The total number of reads T is defined as T = ∑iOi . Pervasive transcription is defined as the transcription of regions well beyond the boundaries of known genes [27] . We rationalized that the signal emanating from pervasive transcription would correlate to the number of bases annotated as each subfamily in the reference genome ( hg38 ) . We used Repeat Masker to count the number of instances and number of bases in hg38 annotated as the subfamily i ∈ {L1Hs , L1PA2 , L1PA3 , L1PA4 , L1P1} . We define Pi as the proportion of LINE-1 bases annotated as the subfamily i in the Eq 3: Pi=Bi∑jBj , j∈{L1Hs , L1PA2 , L1PA3 , L1PA4 , L1P1} On the other had mappability fingerprints , which represents how reads deriving from LINE-1 transcripts would be mapped to the genome , are created by aligning simulated reads deriving from putative L1 transcripts from each L1 subfamily . For each L1 subfamily , we extract the sequences of instances based on RepeatMasker annotation and the reference genome ( hg38 ) . Read from putative transcripts are generated using wgsim ( https://github . com/lh3/wgsim-parameters:-1 [RNA-seq mean read length]–N 100000 -d0 –r0 . 1 -e 0 ) . One hundred simulations are performed and reads are aligned to the human reference genome ( hg38 ) using the same parameters described in the model session . The three-dimensional count matrix C is defined as the number of reads mapped to the subfamily i ∈ {L1Hs , L1PA2 , L1PA3 , L1PA4 , L1P1} emanating from the set of full-length transcripts j ∈ {L1Hs , L1PA2 , L1PA3 , L1PA4 , L1P1} in the simulation k . The matrix M is defined as the median percentage of counts across all simulations as in Eq 4: Mi . j=mediank∈{1 , 2 , . . , 100} ( Ci , j , k∑f∈{L1Hs , L1PA2 , L1PA3 , L1PA4 , L1P1}Ci , f , k ) We tested whether different aligners yield different mappability fingerprints . BWA , STAR , and bowtie2 yielded similar results ( S15 Fig ) . As L1 transcripts are not spliced , we decided to integrate bowtie2 as the main TeXP aligner . We further tested the effect of read length on L1Hs subfamily mappability fingerprints ( S16 Fig ) . To counter the effects of distinct read lengths TeXP constructs L1 mappability fingerprints libraries based on fastq read length . We simulated reads emanating from their respective L1 subfamily transcripts and aligned these reads to the human reference genome creating a mappability fingerprint for each L1 subfamily ( S3 Fig ) . When we analyzed the L1 subfamily mappability fingerprints we observed that younger L1 subfamilies tend to have more reads mapped to other L1 subfamilies . For example , we find that only approximately 25% of reads from L1Hs ( the most recent–and supposedly active L1 ) maps back to loci annotated as L1Hs . While older subfamilies such as L1PA4 , have a higher proportion of reads mapping back to its instances ( ~70%—S3 Fig ) . The known variables Oi , T , the vector Pi , the mappability fingerprint matrix Mi . j are used to estimate the signal proportion ε and ϵ in Eq 2 by solving a linear regression . We used lasso regression ( L1 regression ) to maintain sparsity . We used the R package penalized ( [54]—parameters: unpenalized = ~0 , lambda2 = 0 , positive = TRUE , standardize = TRUE , plot = FALSE , minsteps = 10000 , maxiter = 1000 ) . TeXP was developed as a combination of bash , R and python scripts . The source code is available at https://github . com/gersteinlab/texp . A docker image is also available for users at dockerhub under fnavarro/texp . Raw RNA sequencing datasets from healthy tissues were obtained from Database of Genotypes and Phenotypes ( DB-Gap - https://dbgap . ncbi . nlm . nih . gov ) accession number phs000424 . v6 . p1 . Raw RNA sequencing data from cell lines were obtained from the ENCODE data portal ( https://www . encodeproject . org/search ) . We selected RNA-seq experiments from immortalized cell lines with multiple cellular fractions and transcripts selection experiments . Accessions and cell lines are available in S1 Table . Mutation load and cell turn-over rate were extracted from the compilation of somatic mutation rate in Tomasetti et al [55] . More ancient elements such as DNA transposons and LINE-2 have been shown to be primarily transcribed pervasively , hitchhiking the transcription of nearby autonomously transcribed regions [32] . Therefore , we tested whether our estimation of L1Hs transcription level correlated with genes containing or adjacent to L1Hs instances . We found no significant difference between the correlation distribution of a random set of genes and genes with L1Hs in exons or introns or within 3kb upstream or 3kb downstream of L1Hs . This finding indicates that our estimation of L1Hs autonomous transcription is not significantly influenced by non-autonomous L1Hs transcription adjacent or contained by protein-coding genes’ loci . Furthermore , we tested if and enrichment of pervasive transcription deriving from intronic regions would create a background signal distinct from the pervasive transcription derived from a whole genome model . We correlated the number of LINE-1 instances from each subfamily in intergenic and intronic regions based on GENCODE v29 and found a statistically significant correlation between the number of instances in both regions ( Spearman corr = 0 . 979057 , p-value < 2 . 2e-16—S17 Fig ) . All the cell lines used in this study were obtained from the American Type Culture Collection ( ATCC ) ( Manassas , VA , USA ) . MCF-7 cells were cultured in Dulbecco’s Modified Eagle Medium: Nutrient Mixture F-12 ( DMEM/F12; Gibco ) . HeLa , SK-MEL-5 , and HepG2 cells were cultured in Dulbecco’s Modified Eagle Medium ( DMEM; Gibco ) . K562 and GM12878 cells were cultured in RPMI 1640 ( Gibco ) . All cell culture media were supplemented with 10% fetal bovine serum ( FBS ) ( Atlanta Biologics ) and 1% penicillin/streptomycin ( Fisher Scientific ) . All cells were cultured and expanded using the standard methods . RNA was extracted using the RNeasy PLUS Mini Kit and the QIAshredders ( Qiagen ) following the manufacturer’s protocol . All samples were treated with DNase I ( New England BioLabs Inc . ) to remove any remaining genomic DNA . RNA concentration was determined by Qubit 2 . 0 Fluorometer ( Invitrogen ) . RNA quality was determined by Nanodrop ( Thermo Scientific ) and 2100 BioAnalyzer with the Agilent RNA 6000 Nano kit ( Agilent Technologies ) . Approximately 5 μg of RNA was used for synthesis of the cDNA using the iScript Advanced cDNA Synthesis Kit ( Bio-Rad ) . The final cDNA product was quantified and a working solution of 10 ng/μL was prepared for the subsequent studies . Droplet Digital PCR ( ddPCR ) System ( Bio-Rad Laboratories ) was utilized to quantify the L1Hs transcript expression in the cell lines described above . Since L1Hs is a highly repetitive and heterogeneous target , we had initially designed and tested a panel of primers and probes that targeted the 5’ untranslated region ( 5’UTR ) , the open reading frame 1 ( ORF1 ) , the open reading frame 2 ( ORF2 ) , and the 3’ untranslated region ( 3’UTR ) of the L1Hs locus , respectively . After a pilot screening study , we selected the two assays covering ORF1 and ORF2 , which not only exhibited overall better performance , but also could help us to distinguish autonomous and pervasive L1Hs transcriptions . We also designed two reference assays on the housekeeping gene HPRT1 , which targeted the 5’ and 3’ ends of the transcript , respectively ( S2 Table ) . All the ddPCR primers and probes were designed based on the human genome reference hg19 ( GRCh37 ) and synthesized by IDT ( Integrated DNA Technologies , Inc . Coralville , Iowa , USA ) . The ddPCR reactions were performed according to the protocol provided by the manufacturer . Briefly , 10ng DNA template was mixed with the PCR Mastermix , primers , and probes to a final volume of 20 μL , followed by mixing with 60 μL of droplet generation oil to generate the droplet by the Bio-Rad QX200 Droplet Generator . After the droplets were generated , they were transferred into a 96-well PCR plate and then heat-sealed with a foil seal . PCR amplification was performed using a C1000 Touch thermal cycler and once completed , the 96-well PCR plate was loaded on the QX200 Droplet Reader . All ddPCR assays performed in this study included two normal human controls ( NA12878 and NA10851 ) and two mouse controls ( NSG and XFED/X3T3 ) as well as a no-template control ( NTC , no DNA template ) . All samples and controls were run in duplicates . Data was analyzed utilizing the QuantaSoft analysis software provided by the manufacturer ( Bio-Rad ) . Data were presented in copies of transcript/μL format which was mathematically normalized to copies of transcript/ng to allow for comparison between cell lines . In order to estimate the levels of pervasive and autonomous transcription using ddPCR we use the following formulation: X3′=a+p X5′=a+fp Where X3′ and X5′ are the transcript/μL measurements of the 3’ and 5’ L1Hs primers; a and p are autonomous and pervasive transcription estimates , respectively; and f is the fraction of L1Hs instances that could potentially generate full-length transcripts . We estimate that f is approximately 12% . Using this model , the estimated level of L1Hs autonomous transcription using two primers and ddPCR is: a=0 . 88*X5′−0 . 12*X3′ We designed two assays targeting the 5’ and 3’ ends of the HPRT1 transcript , respectively , and used as the reference controls in this study ( S3 Table ) . The reference gene expression level was found to be constant within each cell line , but varied between cell lines . In addition , while 4 of the 6 cell lines had similar 5’ and 3’ end expression , K562 and GM12878 both had increased 3’ end expression . This could be from different isoforms being expressed with different frequencies3 . For the 5’ end expression of HPRT , SK-MEL-5 , GM12878 , and HepG2 were all around 600 copies of transcript/ng . The remaining were all around 1200 copies of transcript/ng . When looking at the 3’ end expression , we found that SK-MEL-5 and HepG2 were around 750 copies of transcript/ng , while MCF-7 , GM12878 , and HeLa were around 1350 copies of transcript/ng , and K562 was close to 1800 copies of transcript/ng . The slight difference between the 5’ end and the 3’ end expression levels in the same cell line could be explained by a potential 3’ end bias in the cDNA synthesis . However , all the reference assays were consistent between experiments and did not affect the target expression . | Repetitive sequences , such as LINEs , comprise more than half of the human genome . Due to their repetitive nature , LINEs are hard to grasp . In particular , we find that pervasive transcription is a major confounding factor in transcriptome data . We observe that , on average , more than 90% of LINE signal derives from pervasive transcription . To investigate this issue , we developed and validated a new method called TeXP . TeXP accounts and removes the effects of pervasive transcription when quantifying LINE activity . Our method uses the broad distribution of LINEs to estimate the effects of pervasive transcription . Using TeXP , we processed thousands of transcriptome datasets to uniformly , and unbiasedly measure LINE-1 activity across healthy somatic cells . By removing the pervasive transcription component , we find that ( 1 ) LINE-1 is broadly expressed in healthy somatic tissues; ( 2 ) Adult brain show small levels of LINE transcription and; ( 3 ) LINE-1 transcription level is correlated with tissue cell turnover . Our method thus offers insights into how repetitive sequences and influenced by pervasive transcription . Moreover , we uncover the activity of LINE-1 in somatic tissues at an unmatched scale . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"sequencing",
"techniques",
"genetic",
"fingerprinting",
"cultured",
"fibroblasts",
"biological",
"cultures",
"human",
"genomics",
"dna",
"transcription",
"genetic",
"elements",
"molecular",
"biology",
"techniques",
"rna",
"sequencing",
"research",
"and",
"analysis",
"methods",
"genomic",
"signal",
"processing",
"artificial",
"gene",
"amplification",
"and",
"extension",
"gene",
"expression",
"cell",
"lines",
"molecular",
"biology",
"signal",
"transduction",
"genetic",
"fingerprinting",
"and",
"footprinting",
"cell",
"biology",
"genetics",
"transposable",
"elements",
"biology",
"and",
"life",
"sciences",
"genomics",
"mobile",
"genetic",
"elements",
"cell",
"signaling",
"polymerase",
"chain",
"reaction"
]
| 2019 | TeXP: Deconvolving the effects of pervasive and autonomous transcription of transposable elements |
In contrast to pathogenic HIV/SIV infections of humans and rhesus macaques ( RMs ) , natural SIV infection of sooty mangabeys ( SMs ) is typically non-pathogenic despite high viremia . Several studies suggested that low immune activation and relative resistance of CD4+ central memory T-cells from virus infection are mechanisms that protect SMs from AIDS . In 2008 it was reported that plasmacytoid dendritic cells ( pDCs ) of SMs exhibit attenuated interferon-alpha ( IFN-α ) responses to TLR7/9 ligands in vitro , and that species-specific amino acid substitutions in SM Interferon Regulatory Factor-7 ( IRF7 ) are responsible for this observation . Based on these findings , these authors proposed that “muted” IFN-α responses are responsible for the benign nature of SIV infection in SMs . However , other studies indicated that acutely SIV-infected SMs show robust IFN-α responses and marked upregulation of Interferon Stimulated Genes ( ISGs ) . To investigate this apparent disparity , we first examined the role of the reported IRF7 amino acid substitutions in SMs . To this end , we sequenced all IRF7 exons in 16 breeders , and exons displaying variability ( exons 2 , 3 , 5 , 6 , 7 , 8 ) in the remainder of the colony ( 177 animals ) . We found that the reported Ser-Gly substitution at position 191 was a sequencing error , and that several of the remaining substitutions represent only minor alleles . In addition , functional assays using recombinant SM IRF7 showed no defect in its ability to translocate in the nucleus and drive transcription from an IFN-α promoter . Furthermore , in vitro stimulation of SM peripheral blood mononuclear cells with either the TLR7 agonist CL097 or SIVmac239 induced an 500–800-fold induction of IFN-α and IFN-β mRNA , and levels of IFN-α production by pDCs similar to those of RMs or humans . These data establish that IFN-α and IRF7 signaling in SMs are largely intact , with differences with RMs that are minor and unlikely to play any role in the AIDS resistance of SIV-infected SMs .
In contrast to human immunodeficiency virus ( HIV ) infection of humans and experimental simian immunodeficiency virus ( SIV ) infection of Asian macaques , SIV infection of African monkeys that are natural hosts , such as sooty mangabeys ( SM ) , is typically non-pathogenic despite high virus replication [1] , [2] . Understanding the mechanisms underlying how SMs are able to avoid AIDS remains an area of active investigation [3] . Type I interferons , including IFN-α , are a family of cytokines that play a central role in the innate antiviral response mediated by different cell types , and in particular plasmacytoid dendritic cells , pDCs [4] . The production of type I IFNs is induced by numerous innate signaling pathways ( including TLRs , NLR , RLRs , etc ) and results in the expression of hundreds of antiviral effector genes that are collectively referred to as Interferon Stimulated Genes ( ISGs ) [5]–[7] . Pathogenic HIV infection of humans and SIV infection of macaques are associated with a strong type I interferon response that persists during the chronic phase of infection [8]–[11] . While exogenous IFN-α administration exerts a clear antiviral effect in vivo in both pathogenic and non-pathogenic infections [12]–[14] , an elevated type I IFN response has also been proposed as an immunopathogenic mechanism [15]–[17] and reported as a marker of poor immunologic response to antiretroviral therapy [18] . As such , the pathophysiological consequences of type I interferon responses during pathogenic HIV/SIV infections are complex and not fully understood . An article published in Nature Medicine in 2008 by Mandl et al . reported that pDCs of SMs exhibit a muted in vitro production of type I interferon ( IFN ) in response to Toll-like receptor ( TLR ) -7/9 ligands and SIV , and attributed this observation to amino acid substitutions specific to SMs in the transactivation domain of the Interferon regulatory factor ( IRF ) 7 signaling molecule [19] . Based on these findings , the authors proposed that this muted type I IFN response to SIV is the key mechanism protecting SIV-infected SMs from developing chronic immune activation in response to the virus , and ultimately from development of AIDS [19] . If confirmed , these data would be of significant scientific impact as they would indicate that the evolutionary pressure posed by SIV in the SM innate immune system has resulted in the functional amputation of a signaling axis ( i . e . , the IRF7/type I IFN ) that is evolutionarily conserved and known to play a major role in containing the replication of RNA viruses in many models of infection [20] . On the other hand , several independent studies from other groups suggested that acute SIV infection of SMs is in fact associated with a widespread innate and adaptive immune response to the virus with massive upregulation of dozens of interferon-stimulated genes ( ISGs ) [10]; production of IFN-α and IFN-β in the lymph nodes of acutely infected SMs [21]; and detection of plasma IFN-α and IP10/CXCL10 during acute SIV infection [22] . Of note , this immune response and ISG upregulation is controlled within a few weeks of the initial infection in SIV-infected SMs , while chronic immune activation and persistent type I IFN responses are present during pathogenic HIV/SIV infections of humans and rhesus macaques [8]–[10] , [17] . In this study , we sought to investigate the apparent discrepancy between the largely in vitro data reported in Mandl et al , and the in vivo data indicating a strong type I IFN response during acute SIV infection of SMs . To this end , we performed a series of experiments investigating the overall functionality , in SM-derived PBMCs and pDCs , of the innate immune pathway involving TLR7/9 stimulation , IRF7 activation , and type I IFN gene expression and protein production . We found that: ( i ) most of the SM-specific amino acid substitutions in IRF7 originally reported by Mandl et al . , were either minor alleles or an artifact of sequencing error; ( ii ) IRF7 of SMs is fully functional and retains the ability to translocate in the nucleus and transactivate IFNA genes; ( iii ) SM PBMCs upregulate type I IFN mRNA expression by 500–800 fold in response to SIVmac239 or TLR7 ligand CL097; and ( iv ) SM pDCs show a robust production of IFN-α in response to CL097 or SIVmac239 by intracellular cytokine staining . Based on these results we concluded that IFN-α and IRF7 signaling in SMs are largely intact , and that any differences with RMs are minor and highly unlikely to play any role in the AIDS resistance of SIV-infected SMs .
The SM IRF7 gene sequence published in Mandl et al . [19] was the first description of this important innate immune gene in a natural SIV host species . Of note , that sequence refers to data derived from a single animal ( FHy ) using cDNA as a template . SM-IRF7 contains 10 exons . To accurately assess the species-wide variation of the IRF7 gene in SMs , we initially sequenced all 10 exons in 16 breeder animals representative of the genetic variability in the colony housed at the Yerkes Primate National Research Center ( YNPRC ) based on pedigree . Sequencing was performed on genomic DNA in both directions , and any overlapping loci displaying ambiguity were re-sequenced . Six ( exons 2 , 3 , 5 , 6 , 7 , and 8 ) displayed variation at the amino acid level within 16 breeder animals initially sequenced . No variation was observed in four exons ( 1 , 2 , 9 and 10 ) in the breeders . To establish the colony-wide distribution of polymorphisms within variant-containing exons , we then sequenced the six exons displaying variability in the remainder of the colony ( 177 animals ) . We were able to generate sequence data in all 177 animals for exons 2 , 3 , 5 , 7 and 8 , but were only able to sequence exon 6 in 50 animals . No further sequencing efforts were undertaken for exons 1 , 4 , 9 or 10 as no variation within these regions was observed in Mandl et al [19] or within our initial sequencing of breeders . A summary of the number of animals sequenced for each exon is detailed in Supplementary Table S1 . Importantly , we also re-sequenced SM-IRF7 in the animal ( FHy ) from which the sequence reported in Mandl et al . [19] was derived . Our sequence analysis indicated that , out of the seven SM-specific amino acid substitutions reported in Mandl et al . [19] one substitution ( Ser-to-Gly in position 191 ) is a sequencing error and four others are not fixed in the SM population , with two being present in homozygosity only in a minority of animals ( Table 1 ) . It should be noted that of the seven SM-specific amino acid substitutions originally reported , the Ser-to-Gly at position 191 was the most potentially disruptive , since the serine in this position is highly conserved across species , being present in humans , macaques , mice , rats and various non-human primates . Upon re-sequencing the SM colony , we did not observe this purported genotype in any of the 16 breeder animals , nor in the 177 non-breeders . We also re-sequenced animal FHy , from which the SM-IRF7 sequence was derived in the original study by Mandl , and again , did not observe the S191G substitution . These data indicate that serine is present at position 191 in the entire SM colony . While this amino-acid substitution would have represented a good candidate for phenotypic differences in SMs , our results indicate that the observation of this polymorphism in Mandl et al . [19] is an artifact of a sequencing error . Colony-wide re-sequencing also demonstrated that other non-synonymous nucleotide changes reported by Mandl et al . [19] either represented minor variants within the colony or resulted in conservative amino acid changes . At position 252 , the SM sequence was originally reported to be valine , divergent from alanine in humans and threonine in rhesus macaques ( Table 1 ) . Upon resequencing animal FHy , this position was found to be heterozygous ( alanine/valine ) . Sequencing 50 SMs for exon 6 showed that 24% were V/V homozygous , 33% A/A homozygous , and 43% A/V heterozygous . Given the variability at this locus among SMs , this substitution is unlikely to have major effects on immune activation , as all published data indicate that the low immune activation of SIV-infected SMs is a general phenomenon among these animals . Position 256 was reported to be a Thr-to-Ala substitution in SMs . Resequencing showed that animal FHy is indeed A/A homozygous in this position . However , our sequencing of 50 animals and found that only 28% were A/A , 30% heterozygous A/T , and 42% homozygous T/T ( and therefore identical to humans and macaques ) . Position 165 was reported as an Ala-to-Gly substitution in SMs; however , G/G homozygosity was found in 78% of SMs , with 22% G/A heterozygosity . While the reported sequence of FHy for this position is correct , our data indicate that a significant polymorphism is present at the population level in SMs . At position 268 , the SMs were reported to have an Ala-to-Val substitution . Resequencing demonstrated that 73% of the animals were homozygous for V/V , 25% V/A heterozygous , and 2% homozygous A/A . In summary , our work demonstrates that , of the seven purportedly SM-specific IRF7 amino-acid substitutions described in Mandl et al . [19] , only two are actually fixed in the SM population ( i . e . , Ala-to-Gly at position 203 , which is highly conservative , and Gln-to-Arg at position 413 ) . While R at position 413 was fixed in the SM colony , the locus is not fixed in humans , in whom R is present with 28% minor allelic frequency [23] . As mentioned above , low immune activation is common to virtually all SIV-infected SMs . However , variations among individual animals have been described [24] , [25] . To determine whether the observed single nucleotide polymorphisms in the IRF7 sequence impact on the phenotype of SIV infection in SMs , we tested for the presence of statistically significant relationships between the observed SNPs and markers of SIV disease progression . We found that none of the observed IRF7 polymorphisms in SIV-infected SMs showed a significant association with viral load , CD4+ T cell counts , or immune activation ( Table 2 ) . While our extensive sequencing of SM-IRF7 did not reveal any difference with either human or rhesus macaque IRF7 genes that would predict a significant loss of function , these sequencing data do not provide per se any functional information about SM-IRF7 . For this reason , we elected to next assess the transactivation potential of SM-IRF7 on an IFNA promoter . We synthesized and cloned into the expression vector pCMV2 a full-length SM-IRF7 gene sequence identical to that of animal FFz , which contained the most representative alleles found in the SMs housed at the Yerkes colony . The ability of SM-IRF7 to initiate IFNA transcription was compared to that of RM-IRF7 , human IRF7 , and two loss-of-function mutants , using a well-described reporter system in which IRF7 , a hu-IFNA4 reporter construct and the activating kinase TBK1 are co-expressed in HEK 293 cells [26] . As shown in Figure 1A , co-expression of human , RM , and SM-IRF7 with the IFNA4 reporter construct alone yielded moderate induction of luciferase activity , consistent with the basal activity that has been reported previously for hu-IRF7 [27] . However , co-transfection of SM-IRF7 together with the IFNA4 reporter and the activating kinase TBK1 resulted in strong induction of transcriptional activity that was >600 , 000-fold over background ( Figure 1A ) . Importantly , we did not observe any significant difference between the transactivation activity of SM-IRF7 and RM-IRF7 . The observed activity of SM-IRF7 was slightly higher than that of human IRF7 ( Figure 1A ) . In addition , we compared SM-IRF7 activity against two loss-of-function IRF7 mutants , IRF7-7A , which contains Ser-to-Ala substitutions in the serine-rich domain , and IRF7-Δ , a dominant-negative mutant missing AA 7-101 of the DNA-binding domain . If the few actual species-specific substitutions in SM-IRF7 impaired its ability to initiate transcription from the IFNA promoter , we would predict that its activity would be closer to the range of the loss-of-function mutants . However , SM-IRF7 transactivation activity was >1000-fold higher than the mutants , and was in general equivalent to , or greater than , human or RM IRF7 constructs ( Figure 1A ) . Previous studies have demonstrated that viral infections can induce the nuclear translocation of IRF7 in transfected cells [28] , [29] and in purified pDCs [30] . Studies of IRF7 nuclear retention activity have shown that mutants containing deletions of AA247-415 , AA247-305 , AA417-440 , or S477A/S479A substitution mutations can abrogate nuclear translocation . As some of the species-specific amino acid substitutions in SM-IRF7 are located in these regions , we considered that the possibility that they may abrogate the ability of SM-IRF7 to enter the nucleus after activation . We thus tested this hypothesis using an assay that employs co-transfection of IRF7-GFP fusion protein with TBK1 [26] . Transfection of SM-IRF7-GFP alone resulted in cells displaying a predominantly cytoplasmic localization , with the transfected molecule being excluded from DAPI-stained nuclei ( Figure 2B ) . However , 24 hrs after co-transfection with TBK1 , the majority of cells expressing of SM-IRF7-GFP were clearly localized in the nucleus ( Figure 2B ) . Collectively , these data indicate that the observed species-specific substitutions in SM-IRF7 do not confer a defect in its ability to transactivate IFNA expression relative to humans or RM IRF7 , nor do they abrogate the ability of SM-IRF7 to translocate to the nucleus after activation . Our sequence and functional data indicated very clearly that IRF7 signaling in SMs was intact , and comparable to that of humans or RMs . To further investigate potential intrinsic defects in the type I IFN response in SMs , we examined the production of IFN-α by pDCs from uninfected SMs ( defined as Lineage-neg , HLA-DR+ , CD123+ cells , gating strategy shown in Supplementary Figure S1 ) after ex vivo stimulation with aldrithiol-2 inactivated ( AT-2 ) SIV strains ( i . e . , SIVmac239 and SIVsmmE543 . 1 ) or TLR7 ligands . We found that after overnight culture with SIVmac239 , production of IFN-α by SM pDCs was quite robust , with an average frequency of IFN-α-producing pDCs of 41 . 3% ( Figure 2A , C ) . The lowest responder SM that we observed had a frequency of 15 . 4% of IFN-α-producing pDCs in response to SIVmac239 , and the highest responder showed a frequency of 61 . 2% ( Figure 2A ) . Of note , SM pDCs incubated without virus or with microvesicles always produced <2% background production of IFN-α . When SM PBMCs were incubated with the SM-derived strain AT-2 inactivated SIVsmmE543 . 1 , we also observed robust IFN-α production in most animals , with the highest responder having a frequency of 42 . 7% pDCs staining positive for intracellular IFN-α . Average production of IFN-α in SM pDCs in response to SIVsmmE543 . 1 was 21 . 9% , with two animals showing low responses at ∼5% ( Supplementary Figure S2 ) . In the next set of experiments , we incubated PBMCs from SMs with CL097 , a derivative of the synthetic TLR7 agonist R848 , and observed a similarly robust induction of IFN-α by pDCs ( Figure 2A , B ) . Indeed , the average level of IFN-α induction in pDCs following CL097 stimulation was indistinguishable between SMs compared to RMs ( 55 . 0% and 56 . 2% of IFN-α-producing pDCs , respectively ) . While IFN-α production in response to SIVmac239 was higher in RMs than in SMs ( 65 . 5% vs 41 . 3% of IFN-α-producing pDCs p<0 . 05 ) , the type of response observed in SMs is far from “muted” and in fact is actually higher than what we observed in human pDCs incubated with AT-2 inactivated HIV ( Figure 2C ) . Similarly , the IFN-α response in pDCs to SIVsmmE543 . 1 was higher in RMs than SMs ( 41% vs 21% ) , however the response in SMs was overall quite robust . To compare the kinetics of IFN-α expression in SMs and RMs , we next conducted a series of time course experiments and measured the level of IFN-α mRNA in PBMCs at various time points after stimulation with AT-2 SIVmac239 . As shown in Figure 2D , we observed that mRNA induction was very high in both species two hours after incubation and virtually equivalent in magnitude ( average fold-change relative to unstimulated: SM 669-fold , RM 682-fold ) . However , after two hours of stimulation , the kinetics diverged between the two species , with RMs peaking at four hours post-stimulation whereas SMs started to decline . After overnight stimulation , IFN-α levels returned to baseline in the SM ( 1 . 6-fold ) but remained elevated in the RMs ( 18-fold ) . Interestingly , these data are consistent with the strong pattern of IFN-α mRNA induction in SMs described in Mandl et al [19] . It is tempting to speculate that the observation of a similarly strong but more transient IFN-α upregulation in SMs as compared to RMs may play a role in the rapid down-modulation of the type I IFN response that occurs in vivo during acute SIV infection of SMs . During the acute phase of SIV infection , SMs produce high levels of type I IFNs and show massive upregulation of ISGs , which can be induced by any type I IFN ( α , β , ω , λ ) [10] , [21] . To test the possibility that SMs can produce other non-α type I IFNs in response to TLR7 stimulation , we measured the level of IFN-β gene expression at the RNA level in PBMCs derived from both SMs and RMs after in vitro stimulation with the TLR7 ligand CL097 as well as SIVmac239 . It should be noted that , in these experiments , we could not link IFN-β production to pDCs due to lack of an IFN-β-specific monoclonal antibody that can be used for flow cytometric analysis . As shown in Figure 3 , both the TLR7 ligand CL097 and SIVmac239 induced a marked upregulation of IFN-β gene expression . Concurrently , we also observed robust upregulation of IFN-β protein in the supernatants of PBMCs stimulated with CL097 . While multiple cell types may be contributing to IFN-β production , these data are consistent with previously published results showing high levels of IFN-β production by pDCs in the lymph nodes of SIV-infected SMs during the acute phase of infection [21] , and indicate that , in SMs , the ability to produce additional type I IFNs after TLR7 stimulation is largely intact .
In stark contrast to pathogenic HIV and SIV infections of humans and macaques , respectively , SIV infections of natural host species such as the sooty mangabeys ( SM ) are typically non-pathogenic despite high viremia . The mechanisms by which SIV-infected SMs avoid AIDS are complex and only partly understood , but appear to be related to the absence of chronic immune activation and the relative preservation of CD4+ central-memory T cells from direct virus infection [31] . Of note , the attenuated immune activation of chronically SIV-infected SMs occurs as a result of a rapid down-modulation of a vigorous innate and adaptive immune response that is observed during the first weeks of infection [10] , [32] , [33] . An influential study published in 2008 by Mandl et al [19] reported that SMs exhibit a near-complete deficiency in the ability of pDCs to produce IFN-α in response to SIV or TLR7 ligands in vitro . In this study , the defective IFN-α production was attributed to species-specific differences at the amino acid level of the IRF7 signaling molecule between SMs and “pathogenic” hosts such as humans and rhesus macaques [19] . The authors concluded that a “muted” type I IFN response to SIV by pDCs is the main reason why SIV-infected SMs avoid chronic immune activation and do not progress to AIDS . This study drew considerable interest , as it would have represented , if confirmed , the first example of a species-specific adaptation to a chronic virus infection that occurs through a genetically determined abrogation of a key innate immune pathway , i . e . , the TLR7/9-IRF7-type I IFN pathway . However , subsequent studies of the in vivo immune response to SIV during acute infection of SMs showed a vigorous , albeit transient , type I IFN response in the blood and lymph nodes of these animals [10] , [21] . In the current study , we have found that several of the published experimental results could not be replicated , and have provided extensive evidence that , in fact , SMs exhibit an intact IRF7 sequence and function , and most importantly , are capable of strong type I IFN production in response to SIV and TLR-7 ligands . We first set out to establish the genetic distribution of the reported IRF7 polymorphism within the SM population housed at the Yerkes Center ( total of 177 animals ) , and found that the most potentially disruptive mutation , S191G , reported by Mandl et al [19] was a sequencing error , and other reported species-specific amino acid substitutions in SM IRF7 are allelic variants whose presence or absence has no impact whatsoever on the phenotype of SIV infection in terms of virus replication , CD4+ T cell counts , level of immune activation or disease progression . The inaccurate and/or incomplete information on the gene sequence of SM IRF7 presented in Mandl et al [19] reflects the fact that only one animal was sequenced , and that the sequence was generated from cDNA derived from mRNA rather than direct sequencing of genomic DNA . The reasons why such a strong conclusion was reached based in an extremely limited set of data are unclear . In any event , the current set of data reflects a truly comprehensive analysis of the SM-IRF7 gene sequence and allelic polymorphisms that should from now on represent the reference dataset for those interested in this field of investigation . In this study , we were able to confirm that two of the reported substitutions located in a region of IRF7 that , based on human studies , may impact on pDC function [28] were indeed fixed in the SM colony ( G203A and Q413R ) . While the G203A allele is highly conservative and likely to be of limited impact , the Q413R substitution is more interesting . Of note , the R allele is expressed at an estimated 28% prevalence in the human population [23] , although its presence was not linked to diminished type I IFN production by pDCs [34] . The Q413 allele was also examined in the context of systemic lupus erythematosus ( SLE ) , in which one report suggested an association between this allele and an increased susceptibility to SLE [35] , and subsequent work did not replicate these findings [36] . Given these findings , we felt that it was important to directly test the ability of SM-IRF7 to transactivate an IFNA promoter . We found that the consensus SM-IRF7 is perfectly capable of efficiently initiating IFNA transcription and that the level of transactivation is equivalent or higher to that of humans and RMs . Of note , our assay system utilized TBK1 to activate the transfected IRF7 constructs . IRF7 is primarily activated through two pathways: endosomal TLR7/9 and the ‘intrinsic’ or RIG-I related pathway [5] . pDCs exclusively use the TLR7/9 pathway to sense RNA viruses [37] , in which IRF7 is phosphorylated by IKKα in association with TRAF6 , MYD88 and other accessory molecules [38] , [39] . Non-pDC cell types predominantly use the RIG-I pathway , which activates IRF7 via TBK1 [37] , [40] . Our choice of in vitro system was based on comparative robustness; although IKKα is more directly relevant to pDCs , it requires co-transfection of additional adaptor proteins and typically results in a modest signal [27] , [38] . The TBK1 reporter system we chose does not require multiple accessory proteins system [26] , [38] and has a stronger signal than other adaptor proteins such as MYD88 or TRAF6 [39] . Nevertheless , these data indicate that SM-IRF7 maintains its primary function , namely induction of transcriptional activity . During a viral infection , rapid induction of innate immunity is necessary to limit viral burden and spread , until an effective adaptive response can be mounted . Equally important is the ability to ‘rein in’ the response , as unabated activation of the TLR and IFN systems is deleterious to the host , and may lead to immunopathology [41] . While our data has demonstrated that SM-IRF7 maintains the ability to efficiently transactivate IFNA promoter activity , it is possible that the SM-specific amino acids that are divergent from humans and RMs could influence other aspects of IRF7 activity , in particular its negative-regulation . Although no studies to date have examined the role of human genetic polymorphisms in the regulation of IRF7 activity , recent work has uncovered molecular strategies by which viruses inhibit IRF7 activity and , conversely , by which mammalian cells self-regulate the IFN response by modulating IRF7 at the transcriptional and post-transcription levels [42] . Litvak et al demonstrated that the transcription factor FOXO3 acts to suppress IRF7 transcription by enhancing deacetylation of the IRF7 promoter [43]; and Lee and colleagues have demonstrated that regions in the 5′ UTR of IRF7 mRNA are targeted for degradation by the ISG OASL in a negative feedback cycle [44] . In previous work , we showed that in vivo SIV infection of SMs is characterized by an initial , widespread IFN response that is rapidly down-regulated within a few weeks from the initial infection [10] . In this study , we observed in vitro that the production of IFNA mRNA in response to SIV also declines more rapidly in SMs than RMs , suggesting that the resolution of the ISG response observed during SIV infection may be due to a more rigid control of IRF7 activity . While this study focused on the coding regions of SM-IRF7 , it will be interesting to evaluate if interspecies differences in the 5′ UTR of IRF7 alter the regulatory feedback loops controlling its activity . Having established that the IFNA transactivation activity of SM-IRF7 is intact , we next investigated the ability of SMs to produce type I IFNs in response to SIV , HIV or the TLR7 ligand CL097 . We found that SM pDCs were able to robustly produce high levels of IFN-α in response to CL097 , SIVmac239 and SIVsmE543 . 1 . We observed that SM pDCs produced similar levels of IFN-α compared to RMs when cultured with CL097 , and were only slightly reduced for SIVmac239 . These data are in direct contrast with the results of Mandl et al , who found an average of ∼5% pDCs producing IFN-α in response to SIVmac239 [19] . One possible explanation for this discrepancy is that the addition of BFA after 8 hours of stimulation has resulted , in Mandl et al , in an underestimation of the frequency of pDCs producing IFN-α in SMs . Indeed , in our hands addition of BFA after 2 or 4 hours of stimulation makes the detection of IFN-α-producing pDCs more sensitive in both SMs and RMs ( data not shown ) . Interestingly , we were able to replicate the strong up-regulation of IFN-α mRNA levels that Mandl et al observed in similarly stimulated SM PBMCs ( ∼500-fold over baseline , see Mandl et al , Figure S3 [19] ) . Furthermore , we found that SM PBMCs show levels of IFN-β mRNA upregulation in response to TLR7 stimulation that are similar or even higher than those observed in RMs . The reasons why Mandl et al . attributed their initial finding of a “muted” IFN-α response at the protein level ( which occurred in association with a robust mRNA upregulation ) to purported IRF7 genetic defects , rather than post-transcriptional mechanisms , remain unclear . It is important to note that , in addition to the data presented here , in which we have shown robust production of IFN-α in pDCs in response to SIV directly , Harris and colleagues have also demonstrated that high levels of IFN-α production by pDCs in the LNs of SIV-infected SMs during acute infection are detectable using immunohistochemical techniques [21] . Taken together , these data indicate quite clearly that production of type I IFNs by SM pDCs is largely intact in SMs . While the observed production of IFN-α by SM pDCs in response to SIVmac239 was quite robust , we found that it was significantly lower than in RMs in terms of frequency of pDCs producing IFN-α in response to SIVmac239 . However , the relatively modest magnitude of this difference ( 41% vs 65% ) makes it unlikely to play any major role in the divergent pathogenicity of SIV infection in these two species . This possibility becomes even more remote if one considers that the production of IFN-α by SM pDCs in response to SIVmac239 , is in fact higher than that of human pDCs in response to HIV as observed by us ( Figure 2C ) and others [34] , [45] . Of note , the difference in the frequency of pDCs producing IFN-α between SMs and RMs may be accounted for by the different kinetics of RNA production between the two species , with SMs resolving their IFN-α RNA upregulation more rapidly than RMs ( Figure 2 ) , and a consequently larger accumulation of intracellular IFN-α protein in RM pDCs during an overnight time course . In this regard , the current set of in vitro data is consistent with previously published in vivo data showing that acute SIV infection of SMs is associated with a massive , albeit transient , up-regulation of ISGs in both blood and lymphoid tissues [10] , and confirms that a more effective down-modulation of type I IFN signaling , rather than an intrinsic genetic inability to produce this cytokine in response to TLR-7 ligands , is a candidate mechanism for the absence of chronic innate immune activation observed in SIV-infected SMs . This model is also consistent with the observation by several groups that acute SIV infection of another natural host species , the African green monkeys , is also associated with a robust but transient type I IFN response in vivo [46]–[48] . In our preliminary experiments , we also found that the ability of SM pDCs to produce IFN-α was highly sensitive to processing time , with a delay of even a few hours from collection to stimulation rendering these cells much less responsive to SIV ( data not shown ) . This observation may in part explain the discrepancy between our results and those published by Mandl et al . In a subsequent report report [49] , the same Authors describe a “muted” IFN-α response to yellow fever virus ( YFV ) both in vitro and in vivo . It will be important to establish if technical factors are also underlying these in vitro data , while the lower YFV replication in SMs as compared to RMs is the most parsimonious explanation for the lower in vivo IFN-α response to YFV observed in SMs in that experiment [49] . The data presented herein support a model in which the ability of SMs to avoid SIV-induced chronic immune activation is due , at least in part , to a rapid control of the type I IFN response occurring during acute infection . It is important to note , that a significant amount of work by several groups has demonstrated that SMs and AGMs have likely evolved multiple , non-mutually exclusive immunological strategies to avoid AIDS-defining chronic immune activation ( reviewed in [3] , [50] ) . While these mechanisms are only partly understood , they appear to be related to ( i ) an absence of chronic immune activation and ( ii ) preservation of key CD4+ lymphocyte subsets/function . The mechanisms underlying the maintenance of low immune activation of SIV-infected SMs are complex , and involve ( i ) the ability of SIVsmm Nef to effectively down-modulate the CD3-TCR complex from the surface of infected cells [51] , [52]; ( ii ) the rapid down-modulation of the innate and adaptive immune activation associated with acute SIV infection [10] , [21] , coincident with a rapid up-regulation of PD-1 expression in LNs [53]; and ( iii ) the preservation of mucosal immune function , with normal levels of CD4+ Th17 cells in the intestine and absence of microbial translocation [54] , [55] . In addition , we proposed that the observed lower levels of virus infection in CD4+ TCM and in lymph nodes of SIV-infected SMs [31] , [53] contributes to the low immune activation by compartmentalizing in vivo virus replication and antigen production away from the secondary lymphoid tissues in which most antiviral immune responses are initiated [3] . Similiarly , SIVsmm infected SMs also exhibit lower frequency of infection of lymph node resident CD4+ TFH cells compared to SIVmac239 and SIVE543 infected RMs [53] . Preservation of CD4+ T helper function in natural hosts may also occur via CD4-negative ‘surrogates’ . AGMs maintain a population of CD4-CD8αDIM T cells demonstrating CD4+ T helper functionality [56] , and SMs harbour a population of CD3+CD4-CD8- ‘double-negative’ T cells that exhibit a predominantly effector phenotype and share functional and transcriptomic features with TH1 , TH2 , TH17 and TFh [57] . This “double negative” lymphocyte population is preserved even in the rare SIV-infected SMs with extremely low ( <50 cell/ul ) CD4+ T cell counts [58] and may explain the lack of AIDS in these animals . While considerable progress has been made in our understanding of the virology and immunology of natural SIV infections , much more work is needed before a comprehensive and exhaustive model emerges of how SMs and other natural hosts avoid AIDS . In conclusion , the data presented here demonstrate very clearly that SM pDC produce robust levels of type I IFNs in response to SIV , and that SM-IRF7 maintains intact IFNA transactivation and nuclear translocation activity . We believe that these findings resolve a previous significant theoretical contradiction in the field of natural SIV infection and provide a solid premise for future studies aimed at defining the molecular mechanisms by which innate immune responses to SIV are rapidly down-modulated in natural host species despite ongoing virus replication . Ultimately , it is hoped that these advances may help designing interventions that target the chronic innate immune activation that is associated with HIV infection of humans .
Ten healthy HIV-uninfected individuals were recruited for this study for blood draws . All individuals who participated in this study provided informed consent in writing in accordance to the protocol approved by the Institutional Review Board of Emory University , IRB#00045821 , entitled “Phlebotomy of healthy adults for the purpose of evaluation and validation of immune response assays” . The protocol adheres to international guidelines established in the Declaration of Helsinki by the World Medical Association . Blood draws were obtained from sooty mangabeys and rhesus macaques housed at the Yerkes National Primate Research Center , which is accredited by American Association of Accreditation of Laboratory Animal Care . 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 , a national set of guidelines in the U . S . and also to international recommendations detailed in the Weatherall Report ( 2006 ) . This work received prior approval by the Institutional Animal Care and Use Committees ( IACUC ) of Emory University ( IACUC protocol #2000793 , entitled “Comparative AIDS Program” ) . Appropriate procedures were performed to ensure that potential distress , pain , discomfort and/or injury was limited to that unavoidable in the conduct of the research plan . The sedative Ketamine ( 10 mg/kg ) and/or Telazol ( 4 mg/kg ) were applied as necessary for blood draws and analgesics were used when determined appropriate by veterinary medical staff . 16 sooty mangabeys ( Cercocebus atys atys ) were identified from different families housed at the Yerkes National Primate Research Center ( YNPRC ) that cumulatively have over 50 offspring within the group and thus represent a large proportion of the genetic variance within the sooty population . Added to this group was animal FHy , from which the IRF7 sequence in the Mandl et al . [19] study was originally derived . Exonic sequence of the IRF7 gene was obtained from UCSC genome browser ( http://genome . ucsc . edu/cgi-bin/hgGateway ) using the rheMac2 assembly of the Indian rhesus macaque genome , and primers were designed to span the entire exon by Primer3 ( http://frodo . wi . mit . edu ) . PCR was performed using standard amplification reactions on ABI 9700 thermal cyclers using MgCl2 concentrations of either 1 . 5 mM or 2 . 0 mM . PCR products were checked for expected size by electrophoresis on agarose gels . Shrimp alkaline phosphatase and Exonuclease I were added to remove single stand DNA . Direct Sanger sequencing reactions were performed using Applied Biosystem Big Dye terminator protocol on an ABI 9700 thermal cycler . The reaction was purified by EDTA/EtOH protocol , and sequencing reactions performed on an ABI 3730 genetic analyzer . Subsequent analysis was done using Sequencher 4 . 7 genetic software . Results were aligned to the corresponding rhesus genomic location . Sequence products of all exons mapped to the expected genomic locations . In secondary analysis we sequenced all 177 animals that currently make up the Yerkes NPRC sooty mangabey colony for the exons exhibiting variation at the amino acid level using the same techniques ( Smentary Table S1 ) . All reported sequences were submitted to NCBI GenBank ( Supplementary Table S2 ) . One-way ANOVA was used to compare the effect of each SNP genotype upon the listed phenotypic measures of immune function . N-terminal FLAG fusion proteins of human-IRF7 and the 6D , 7A , Δ12-101 mutants , TBK1 and the IFNA4-luciferase reporter have been described previously [26] , [27] . The full-length coding region of sooty mangabey IRF7 containing the most representative alleles ( animal FFz ) within the colony was synthesized by Invitrogen ( Carlsbad , CA ) , rhesus IRF7 was PCR amplified from cDNA; and both were cloned into the pFLAG-CMV2 vector using the HindIII and BamHI sites with n-terminal FLAG sequences . The nt sequence of FFz IRF7 has been posted to GenBank under accession # JX438328; the nucleotide , amino acid translation , and alignment of FFz smIRF7 aligned with rhesus and human IRF7 is available in the Supplementary Data Files S1 , S2 , S3 . The pGL3 reporter plasmid ( SV40-luciferase ) was purchased from Promega ( Madison , WI ) . eGFP-SM-IRF7 was made by subcloning sooty mangabey IRF7 into a parental vector of pEGFP-C1 Hu-IRF7 described previously [28] . Subcloning of IRF7 was performed by the Emory Custom Cloning Core Facility . HEK-293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum ( FBS ) . For transient transfection , the HEK-293 cells were cotransfected in 24-well plates with 50 ng of IFNA promoter reporter plasmid , 50 ng of TBK1 and 50 ng of various IRF7 plasmids . All transfections were done with Lipofectamine 2000 ( Invitrogen ) as detailed by the manufacturer using 150 ng of plasmid in 70 ul transfection cocktails . The pGL3 plasmid served as positive control and empty vector was used to equalize the total amount of DNA . Luciferase assays were performed with the Luciferase Assay System ( Promega ) . Cells were harvested 24 hrs posttransfection and lysed in passive lysis buffer and frozen at −80 C . 20 ul of lysates were mixed with Promega Luciferase Assay Substrate and read by luminometer . Readings for a media control were subtracted from relative light unit readings for individual samples . For IRF7-GFP visualization studies , COS cells were maintained in DMEM with 10% FBS and plated to approximately 80% confluency in 6-well plates . 4 ug total plasmid was transfected using 500 ul of 2% Lipofectamine in RPMI . After 24 hrs , cells were washed with PBS , fixed with ice cold 1∶1 methanol∶acetone at −80 C for 10 min and reconstituted in PBS with ProLong Gold Antifade Reagent with DAPI and transferred to slides . Localization was visualized on an AxioVision image capture system ( Carl Zeiss ) fluorescent microscope at 200× and 630× magnification . PBMCs were obtained via density gradient centrifugation using Lymphocyte Separating Media ( Lonza ) according to manufacturer's instructions . Immediately following isolation , cells were cultured in a 48-well plate at 4×106 cells/ml in media containing RPMI ( CellGro ) , 10% defined fetal bovine serum ( HyClone ) penicillin-streptomycin ( CellGro ) , and L-glutamine ( CellGro ) , and incubated with 1 ng/ml IL-3 ( R&D Systems ) for 30 min at 37°C and 5% CO2 . Cells were then stimulated with 10 µM CL097 ( Invivogen ) or virus , as described . All viral stimulations ( SIVmac239 , SIVsme543 . 1 or HIV-1 ADA ) were conducted using aldrithiol 2 ( AT-2 ) inactivated strains at a concentration of 3 µg/ml ( provided by Dr Jeff Lifson , NCI , NIH ) overnight at 37°C and 5% CO2 . For intracellular cytokine experiments , Brefeldin-A ( 10 µg/ml ) was added after 2 hrs of stimulation unless otherwise indicated . Intracellular cytokine staining was performed following the protocol described by Lamoreaux et al [59] . Briefly , PBMCs were incubated with Aqua Live/Dead amine dye-AmCyan ( Invitrogen ) for 10 min followed by 30 min incubation with antibodies specific for surface markers . For cytokine detection PBMCs were fixed and permeabilized using Cytofix/Cytoperm reagents ( BD Biosciences ) according to manufacturer's instructions . Cells were stained with monoclonal antibodies to the following proteins: anti-CD123 PE-Cy7 ( clone 6H6 ) ( eBioscience ) , anti-CD20 Pacific Blue ( clone 2H7 ) ( BioLegend ) , anti-CD14 PE-Texas Red ( clone RM052 ) ( Beckman Coulter ) , anti-IFNa-2 PE ( clone 225 . c ) ( Chromaprobe ) , Aqua LiveDead ( Invitrogen ) , and anti-CD3 Pacific Blue ( clone SP34-2 ) , anti-CD11c APC ( clone SHCL-3 ) , anti-HLA-DR PerCP-Cy5 . 5 ( clone L243/G46-6 ) ( all from BD Biosciences ) . At least 1 , 500 , 000 single , live lymphocyte events were acquired on a LSRII flow cytometer ( BD ) . Analysis was performed with FlowJo ( Treestar ) . pDCs were classified through the following gates: singlet , lymphocyte , CD3-CD20-HLADR+CD14-CD11c-CD123+ as shown in Supplementary Figure S1 . Total RNA from PBMCs was purified using RNeasy mini kits ( QIAGEN ) according to manufacturer's protocol utilizing on-column DNAse digestion . RNA quantity was estimated using Nanodrop analysis . RNA samples ( 0 . 5–1 . 0 µg ) were reverse transcribed in a volume of 20 µl as previously described [9] and 0 . 1 µl of cDNA was used for real time PCR analysis using an ABI 7900 HT instrument ( Applied Biosystems ) . Primers specific for GAPDH mRNA were used to normalize samples . Fold-changes were calculated using the relative standard curve method . IFN-α mRNA was measured using TAQman probes from Applied Biosystems and Universal MasterMix and IFN-β was quantitated using SYBR green . ABI TAQman Probes were: rhesus IFNa2 ( Rh029027494 ) ; rhesus GAPDH ( Rh02621745 ) ; Primer sequences for SYBR green PCR were: IFNB-forward 5′-TTC GCT CTG GCA CAA CAG GTA GTA -3′ , IFNB-reverse 5′-AGC CTT CCA TTC AAT TGC CAC AGG-3′; GAPDH-forward 5′-GAA GGT GAA GGT CGG AGT C , GAPDH-forward 3′-CAA GCT TCC CGT TCT CAG CC . PBMCs from RMs and SMs were incubated with 10 µM CL097 for the times indicated , after which 100 µl of supernatant was harvested . IFN-β levels were quantitated using the Simian Interferon Beta kit ( USCN ) according to manufacturer's instructions . Samples were incubated on pre-coated plates in duplicate . Plates were then read at 405 nm within 15′ on a VMax Kinetic microplate reader ( Molecular Devices ) . A standard was used to quantity protein levels and blanks were used to determine background absorbance . MPMIII ( Biorad ) was used to analyze data . | Sooty mangabey ( SM ) monkeys are an important model for studying HIV disease processes because they do not develop AIDS when infected with SIV , a primate version of HIV . The reasons why SIV-infected SMs remain healthy are not completely understood , but are related to reduced activation of the immune system , and to their ability to shut off a strong antiviral response ( Interferon-alpha ) , that is present during early infection but disappears over time . The interferon-alpha response is an essential component of the host immune response to viral infections , but may contribute to AIDS progression if it persists indefinitely , as occurs in HIV-infected humans and AIDS-susceptible macaque monkeys . In this study , we found that , contrary to a previous report , SMs had an intact interferon response , and SM plasmacytoid dendritic cells ( pDCs ) , a cell-type specialized in producing interferon , were not deficient in their responses to SIV . We also show that IRF7 , a molecule essential to initiate the interferon-alpha response , maintains function in SMs compared to humans and macaques . These data provide novel information on how SIV-infected SMs avoid AIDS despite high levels of virus , and support the hypothesis that immuno-regulatory mechanisms underlie their ability to rapidly shut-off the interferon-alpha that occurs during early infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"immunopathology",
"medicine",
"infectious",
"diseases",
"inflammation",
"cytokines",
"immunity",
"hiv",
"innate",
"immunity",
"retrovirology",
"and",
"hiv",
"immunopathogenesis",
"immunology",
"biology",
"viral",
"diseases",
"immune",
"system"
]
| 2013 | Intact Type I Interferon Production and IRF7 Function in Sooty Mangabeys |
Genotype I ( GI ) Japanese encephalitis virus ( JEV ) that replaced GIII virus has become the dominant circulating virus in Asia . Currently , all registered live and inactivated JEV vaccines are derived from genotype III viruses . In Taiwan , the compulsory JEV vaccination policy recommends that children receives four doses of formalin-inactivated Nakayama ( GIII ) JEV vaccine . To evaluate the influence of genotype replacement on the post-vaccination viral neutralizing ability by GIII and GI viruses , the small panel of vaccinated-children serum specimens was assembled , and the reciprocal 50% plaque-reduction neutralizing antibody titers ( PRNT50 ) were measured against Nakayama vaccine strain , CJN GIII human brain isolate and TC2009-1 GI mosquito isolate . The seropositivity rate ( PRNT50≥1∶10 ) and geometric mean titers ( GMT ) against the TC2009-1 virus were the lowest among the three viruses . The protective threshold against the CJN and TC2009-1 viruses could only be achieved when the GMT against Nakayama virus was ≥1∶20 or ≥1∶80 , respectively . Using undiluted vaccinees' sera , the enhancement of JEV infection in K562 cells was observed in some low or non-neutralizing serum specimens . Our preliminary study has shown that neutralizing antibodies , elicited by the mouse brain-derived and formalin-inactivated JEV Nakayama vaccine among a limited number of vaccinees , have reduced neutralizing capacity against circulating GI virus , but more detailed studies are needed to address the potential impact on the future vaccine policy .
South and Southeast Asia are Japanese encephalitis ( JE ) endemic areas in which approximately 10% of the susceptible populations are infected with JE virus ( JEV ) each year , based on the ratio of asymptomatic to symptomatic infections of 200 to 1 [1] , [2] , [3] . The most cost-effective control strategy for JE is vaccination , and there are several licensed vaccines , including live-attenuated , chimeric live-attenuated and inactivated SA14-14-2; inactivated Nakayama; P3 and Beijing-1 vaccines [3] , [4] , [5] , [6] . In Taiwan , compulsory vaccination has been implemented since 1968 using the mouse-brain derived and formalin-inactivated Nakayama vaccine , and since then clinical JE cases have decreased dramatically to 20–30 cases each year [7] . It has been estimated that vaccine effectiveness is in the range of 85% to 90% after immunization with two doses of inactivated Nakayama vaccine [7] , [8] . We have witnessed dramatic changes in the molecular epidemiology of circulating JEV in the past two decades . Historically , genotype III ( GIII ) viruses were the most widely distributed JEV in South and Southeast Asia [9] . However , genotype I ( GI ) JEV , having emerged in the 1970s in Thailand/Cambodia , has replaced GIII as the dominant circulating virus in JE endemic/epidemic regions since the 1990s [10] . Genotype I viruses first appeared in Japan , and by the 1990s the majority of Japanese JEV isolates belonged to GI [11] . Subsequently , the phenomena of genotype replacement were observed in many countries , including Korea , Vietnam , Thailand , and China [12] , [13] , [14] . Genotype I JEV was first detected in Taiwan in 2008 , and became the dominant circulating genotype island-wide within a year [15] , [16] . The nucleotide and amino acid variation between the envelope ( E ) glycoproteins of GIII and GI JEV is in approximately 12% and 3% , respectively [9] . All licensed JEV vaccine strains , including SA14-14-2 , Nakayama , P3 , and Beijing-1 , belong to GIII . The reduced capacity of neutralizing antibody against field-isolated GIII viruses had been reported among vaccinated human serum samples [17] , [18] . Thus , strain-specific neutralizing antibodies elicited by GIII JEV vaccines in vaccine recipients need to be assessed against GI virus . The protective efficacy of inactivated JE-VAX ( suckling mouse brain-derived Nakayama vaccine ) and P3 , and live-attenuated SA14-14-2 vaccines has been evaluated in a mouse model . Beasley et al . have shown that mice that received JE-VAX vaccine or were passively transferred JE-VAX-vaccinated mouse sera had lower neutralizing antibody titers and were less protected against GI virus as compared to GIII virus , but the strain-dependent protection could not be excluded [19] . However , Liu et al . showed that the live-attenuated SA14-14-2 and inactivated P3 vaccines protected vaccinated mice equally against GIII and GI viruses [20] . In a series of ex vivo studies , Van Gessel et al . eloquently demonstrated that mice receiving passively transferred immune sera collected from adult human volunteers vaccinated with JE-VAX or SA14-14-2-derived IC51 ( tissue culture-derived inactivated ) vaccine were protected against GIII and GI viruses , and firmly established for the first time that the strain-specific correlate of protection ( PRNT50≥10 ) could be accurately estimated by measuring the reciprocal neutralizing antibody titer at the time of viral challenge [21] . The cross-neutralization and protection elicited by GIII JEV vaccines against GI viruses are not consistent in mouse models . More importantly , no study has been conducted using vaccinated children's serum specimens collected from the general population . In the present study , the panel of specimens collected from children immunized with mouse brain-derived and formalin-inactivated Nakayama vaccine in Taiwan were assembled and used to assess the potency of neutralizing antibodies against the vaccine strain , a GIII local human isolate , and a newly introduced GI mosquito isolate .
The nucleotide and amino acid sequences of JEV GI , GIII , and Nakayama viruses were retrieved from GeneBank and analyzed using BioEdit version 7 . 1 . 3 [16] , [22] . To localize the amino acid substitutions , the protein structure of JEV E glycoprotein was downloaded according to recent report [23] and analyzed by Swiss-Pdb Viewer 3 . 7 structure analysis software . The antibody-accessible amino acids should have at least a 35% surface accessibility potential based on the result of structure analysis [24] . In order to assess the impact of amino acid substitutions on the impact of E protein , the stability calculation was performed for all amino acid substitutions by using the Prediction of Proteins Mutations Stability Changes server ( http://babylone . ulb . ac . be/popmusic/index . html ) . In Taiwan , a compulsory vaccination program was implemented in 1968 which utilizes mouse-brain derived and formalin-inactivated Nakayama vaccine . Children receive an initial dose of vaccine at one-and-one-half-years of age , and subsequent doses two weeks later , one year later , and in the first grade of elementary school . We also collected serum samples from children who received varying numbers of doses of vaccine and at different time post final booster vaccination . All the serum samples used in this study were collected from an already-existing collection in two hospitals , the Mennonite Christian Hospital in Hulian and the Tungs' Taichung Metroharbor Hospital in Taichung , in 2010 and were anonymized . For serum sample collection , the clinical protocols were reviewed and approved by the institutional review boards of these two hospitals ( 10-03-007-ER and 99006 ) . Serum was obtained from whole blood after clotting at room temperature and then centrifuged at 3 , 000 rpm for 10 min . The aliquot of serum was stored at −20°C until use . Five JEV strains were used in the serological assay , including: the GIII , cluster III Nakayama vaccine strain; the GIII , cluster I CJN strain isolated from human brain in 1998; the GIII , cluster II T1P1 strain isolated from a mosquito pool in 1997 [15]; and the GI , cluster I TC2009-1 strain and GI , cluster II YL2009-4 strain isolated from mosquito pools in 2009 [16] . Viruses were amplified in C6/36 mosquito cells and stored in aliquots at −80°C until use . Plaque reduction neutralization tests ( PRNT ) are the most suitable method for assessing neutralizing antibodies against JEV . The PRNT protocol used is similar to that in our previous report with some modification [17] . The BHK-21 cells were dispensed into each well of 6-well polystyrene plates ( Costar , Cambridge , MA , USA ) . The plates were incubated at 37°C for 36 h to form a monolayer . Serum samples were inactivated at 56°C for 30 min before a two-fold serial dilution was carried out . A target dose of about 100 plaque forming units ( PFUs ) of JEV were then incubated with the previously diluted serum samples . The mixture of diluted test serum and control virus was added onto the BHK-21 monolayers . After adsorption for 1 h at 37°C , the infected cells were overlaid with 4 ml/well of 1 . 1% methyl cellulose ( Sigma ) in MEM containing 2% FBS and 1% antibiotics . After an additional incubation for 3 . 5–4 days , the cells were fixed with 10% formalin and stained with 1% crystal violet . The PRNT titer was obtained from the reciprocal of the dilution of serum that reduced the plaque number by at least 50% relative to the virus-only control . To determine the potential for antibody-dependent enhancement ( ADE ) of different genotypes of JEV , K562 cells were used to measure viral yield and infection rate [25] . The undiluted serum samples were used in the assay . Briefly , 100 µl serum was mixed with 2×105 K562 cells and 2×104 PFUs virus ( MOI = 0 . 1 ) , and incubated with gentle rotation at 37°C for 2 h . After being washed twice , the cells were incubated with RPMI-FBS medium at 37°C for one day . The culture supernatants were collected and the viral yield was determined by plaque count using BHK-21 cell monolayers; also , infected K562 cells were collected , stained with mouse anti-JEV HIAF , and the cell infection rate was estimated by flow cytometry . The monoclonal antibody 4G2 ( obtained from Dr . Chang GJ of US CDC , Fort Collins , CO ) was diluted 1∶100 and was used as the positive ADE control . A cut-off value to determine seropositivity of neutralizing antibody titer was defined as PRNT50≥10 . Seropositive subjects were defined as those having a reciprocal antibody titer above or equal to the cut-off value; seronegative subjects , those falling below the cut-off value . Antibody titers below the cut-off value of 1∶10 were given an arbitrary value of 5 for geometric mean titer ( GMT ) calculation . The cut-off value for virus yield and cell infection rate for differentiating neutralization and enhancement was calculated from the average of five repeat measurements plus two times the standard deviations ( SD ) of the viral yield and cell infection rate obtained with the negative controls . In the negative controls , RPMI-FBS was substituted for the serum specimen in the virus–serum–cell preparation . Virus yield and cell infection rate of at least 2SDs above the negative control was determined to represent a statistically significant event . Student's t rest was used for all analyses , and statistical significance was defined as a P value < . 05 .
The GIII Nakayama strain has been used exclusively for the manufacture of formalin-inactivated suckling mouse brain-derived human vaccine in Taiwan since 1968 . An extensive field investigation and molecular epidemiological study outlined in our previous report indicates that the replacement of JEV GIII by GI occurred in 2009 in Taiwan [16] . JEV envelope ( E ) protein is the primary antigen eliciting protectively neutralizing antibodies . The amino acid differences in the E protein region between Nakayama vaccine strain and GI and GIII sequences of virus isolated from Taiwan are shown as Table S1 . Amino acid differences in the E protein between the vaccine strain and strains used in this study are highlighted in Figure 1 . There are eight amino acid differences between Nakayama and the human brain isolate GIII CJN virus , including E protein amino acid positions 33 , 51 , 83 , 176 , 227 , 242 , 276 , and 290 ( Figure 1 ) . Among these residues , position 33 , 176 , and 290 are located in E domain I ( EDI ) ; 51 , 83 , 227 , 242 , and 276 are located in EDII; and none are in EDIII . There are 13 amino acid variations between Nakayama and mosquito isolate GI TC2009-1 virus , including E protein amino acid positions 33 , 51 , 83 , 123 , 129 , 176 , 222 , 227 , 242 , 276 , 290 , 327 , and 366 ( Figure 1 ) . Among these residues , positions 33 , 176 , and 290 are located in EDI; 51 , 83 , 123 , 129 , 222 , 227 , 242 , and 276 are located in EDII; and 327 and 366 are located in EDIII . The unique differences distinguish the CJN and GI strains from the vaccine Nakayama strain were at E protein amino acid positions 33 , 51 , 83 , 176 , 227 , 242 , 276 , and 290 . A pilot experiment was conducted to determine the difference of neutralizing antibody titers against five JEVs , GIII Nakayama ( cluster III ) , GIII T1P1 ( cluster II ) , GIII CJN ( cluster I ) , GI TC2009-1 ( cluster I ) , and GI YL2009-4 ( cluster II ) , were evaluated using the small panel of serum samples ( Figure S1 ) [15] , [16] . The correlation of PRNT50 between T1P1 and CJN or TC2009-1 and YL2009-4 was 0 . 75 and 0 . 92 ( Figure S1 ) , respectively . Due to insufficient amount of some serum samples , the GIII Nakayama , GIII CJN , and GI TC2009-1 viruses were selected as viral strains for further analysis . A total of 157 serum samples from vaccinated children were collected and grouped based on age ( 0–15 years ) and dosage of vaccination ( 0–4 doses ) at the time of sampling ( Table 1 ) . The seroprotection or seropositivity rate , defined by the PRNT50≥10 using BHK-21 cells against homologous Nakayama strain , was 66 . 7% after primary vaccination , peaking after the receipt of the 4th dose of vaccine ( 80 . 0% for 8–9-years-old ) , but decreased gradually to 60 . 0% in children aged 14–15 years ( Figure 2 ) . The trend in strain-specific seroprotection rate against the GIII CJN strain was similar to that of Nakayama virus , but the seroprotection rate themselves were generally lower ( ranged from 0 to 21 . 5% ) as compared to those for the Nakayama strain . The seroprotection rate against GI TC2009-1 virus were much lower as compared to other two viruses , especially in the 14–15-years-of-age group ( P< . 05 , Figure 2 ) . The geometric mean titers ( GMT ) of strain-specific PRNT50 against the GIII vaccine and CJN viruses and the TC2009-1 GI virus are shown in Table 1 . The GMT of neutralizing antibodies against homologous Nakayama strain was 25 . 2 after primary vaccination and reached a peak ( 38 . 0 ) following the third dose of vaccine , but decreased gradually after the final booster and dipped below the protective threshold of 10 in children aged 14–15 years . The strain-specific GMT of neutralizing antibodies against the GIII CJN strain trended similarly to those against the Nakayama virus , but the titers were lower . Interestingly , the GMT of antibodies neutralizing field-isolated GI TC2009-1 virus were significantly lower than those for the other two viruses ( P< . 05 ) and the titers were below the presumptive protective threshold of 10 with the exception of children aged 2 . 5–4 years who had received a third dose of vaccine . Neutralizing antibodies elicited by the mouse brain-derived , formalin-inactivated Nakayama vaccine could be protective against circulating GIII and GI virus in Taiwan ( Figure 2 and Table 1 ) . Serum specimens capable of neutralizing Nakayama virus were selected and stratified into groups with PRNT50 titers of 10 , 20 , 40 , 80 , 160 , and ≧320 . The strain-specific GMT from each group was calculated for the human brain GIII CJN and mosquito GI TC2009-1 viruses ( Table 2 ) . The grouping results suggest that the GMT reach the presumptive protective threshold ( PRNT50 = 10 ) against CJN and TC2009-1 viruses when the neutralizing titer against Nakayama virus is 1∶20 or 1∶80 , respectively . Antibodies elicited by the inactivated Nakayama vaccine are less potent in neutralizing GI virus , and the majority of vaccinees' PRNT50 titers are below the protective threshold ( Table 2 ) . The enhancement of virus infection resulting from vaccination has been suggested [26] . Thus , 26 serum samples from vaccinated children were selected , and the undiluted serum samples were used to estimate the potential for ADE of GI JEV infection in K562 cells by measurement of viral yield ( Figure 3 ) . The three serum samples most highly neutralizing against Nakayama ( PRNT50≧80 ) also strongly inhibited the infectivity of GI JEV . Of the 16 weakly neutralizing serum samples ( PRNT50 = 10 to 40 against Nakayama ) , only three exhibited some increase in virus yields as compared to the control serum . However , most ( 6/7 ) of the non-neutralizing serum samples ( PRNT50<10 against Nakayama ) enhanced GI virus infection . The risk of ADE of GI JEV infection in K562 cells was also analyzed by flow cytometry to determine the infection rate of cells infected with serum-treated virus . The results of selected samples , not all samples were included due to insufficient amount individual serum , are shown in Figure 4 . At an MOI = 0 . 1 , the untreated K562 infection rate by TC2009-1 was 6 . 9% . Thus , an infection rate of less than 6 . 9% can be interpreted as neutralization , and greater than 6 . 9% as enhancement . The flavivirus group-cross reactive murine monoclonal antibody 4G2 , used as an ADE control at a 1∶100 dilution , resulted in an infection rate of 20 . 5% , significantly higher than that of the virus control . The T36 ( PRNT50 = 20 ) and C86 ( PRNT50 = 160 ) samples neutralized GI virus , and the cell infection rate were reduced to 1 . 4–1 . 5% , significantly lower than that of the virus control ( P<0 . 05 ) . The B19 serum neutralized Nakayama and CJN viruses , but not TC2009-1 virus . Treatment with B19 serum resulted in some degree of enhancement of TC2009-1 infection ( 10 . 5% infection ) ( P> . 05 ) . Sera I6 , 113 , D7 , D15 , and I5 , non-neutralizing against all three viruses , showed significant enhancement of GI JEV infection ( P< . 05 ) . In addition , we conducted an ex vivo ADE experiment with vaccinees serum samples , and the result was shown in Supplemental Figure 2 . Due to the availability of vaccinee's serum , the undiluted serum from TC36 , C86 , D7 , and I5 only were preincubated with 1000 PFU of TC2009-1 virus and injected intraperitoneally into suckling mouse ( four mice per serum sample ) . The brain of infected suckling mouse was collected at four days post-inoculation; and the virus titer was determined by the plaque forming assay in BHK21 cells . The TC36 and C86 serum samples almost completely neutralized TC2009-1 virus replication in suckling mouse . However , compared to virus control , the mouse brain virus titers were significantly increased to 5 . 53- and 5 . 79-fold when the TC2009-1 virus was pre-incubated with D7 or I5 serum samples , respectively . The result of this ex vivo ADE experiment was consisted with the result of flow analysis ( Figure 4 ) .
Long-term evolution , geographical barriers and host-parasite interactions have resulted in many flaviviruses evolving into multiple genotypes , including dengue virus serotypes 1 through 4 , West Nile virus , and JEV [27] , [28] , [29] , [30] , [31] , [32] . Genotype replacement in dengue virus serotypes 3 and 4 in Sri Lanka and Puerto Rico , respectively , has resulted in increased transmissibility and epidemic potential of the new viruses [33] , [34] . Vaccination is the most effective strategy to control flavivirus epidemics [35] . Genotype I JEV has replaced GIII as the dominant genotype throughout Asian countries , and human clinical cases due to GI virus infection have been observed in China [10] , [36] . The GIII mouse brain-derived , inactivated Nakayama vaccine and live-attenuated SA14-14-2 vaccine are the most widely distributed vaccines in Asian countries , including Japan , Korea , Taiwan , Vietnam , Thailand , Malaysia , China , India , and Nepal [3] , [4] . The replacement of GIII by GI virus provides the opportunity to evaluate the contribution of genotype replacement to the strain-specific vaccine effectiveness of JEV . The reported protective efficacies of GIII JEV vaccines against GI virus are not consistent due to differences in evaluation models used , including virus strain , passive or active immunization , and challenge dosage ( 17 , 18 , 19 ) . The Nakayama JEV was isolated in 1935 , and the vaccine derived from it was developed in 1956 [37] . Differences between vaccine strain and circulating GIII viruses are expected , but the vaccine is proven to be highly protective [38] . The efficacy of formalin-inactivated Nakayama JEV vaccine in vaccinees who received one , two , and three doses of immunization is 85 . 59% , 91 . 07% and 98 . 51% , respectively [8] . The trial of mouse brain-derived inactivated JEV vaccine , including Nakayama or Nakayama plus Beijing-1 strains , was conducted in Thailand in 1984–1985 , and the estimated efficacy was 91% . But , several JE-confirmed cases were diagnosed among vaccinated group indicating the potential of primary vaccine failure . The JEV genotype I was first isolated in the Southeast Asian countries; it has been suggested that the switch from genotype III to I in the 1980s might contribute to the primary vaccine failure without confirmed virological evidence [13] , [39] . Our study also showed that the vaccine , used in Taiwan , offered strain-specific neutralization against field-isolated GIII CJN virus . This result suggests and supports that the amino acid variations , located in EDI and EDII , between these two viruses , are not critical in eliciting neutralizing antibody as compared to residues located in domain III [40] . Our previous report demonstrated GI JEV replaced GIII in Taiwan in 2009 , and here we conducted the first study using serum samples from vaccinated children to systematically evaluate strain-specific neutralizing antibodies elicited by the GIII Nakayama vaccine [16] . Previous report indicated that the GMT titers against GIII isolates , including T1P1 , CC27 , CJN , and CH1392 , were 2-fold lower than that against Nakayama strain; and Shyu et al . , reported that only 37 . 9% of vaccinees sera in the 15–19 year-old group can actually neutralize JE5 Taiwanese isolate [17] , [18] . In presented study , the neutralizing titer against GI virus was 8-fold lower ( Table 2 ) than against Nakayama strain , and the seroprotection threshold against GI virus was 10% of vaccinees in the vaccinated 14–15 year-old group . Thus , the lower titers of antibody against GI in general in Taiwan might relate directly to overall antigenic variability between GIII and GI , but strain-dependent neutralization could not be totally excluded . The seroprotection rate should be appropriately estimated against currently circulating JE strains rather than against vaccine strain . There were only five informative amino acids variations between GIII and GI JEV in the E protein: residues 123 , 129 , 222 , 327 , and 366 . However , the antibodies elicited by the GIII JEV vaccines were only weakly neutralizing against circulating GI virus as compared to the human GIII isolate . Three of the five amino acid differences occur in domain II , which is involved in weakly or non-neutralizing antibody binding [41] , and residues 129 and 222 are not accessible for antibody binding . Thus , the remaining two amino acid variations , residues 327 and 366 , might play an important role in lower strain-specific neutralization against GI virus as compared to the GIII vaccine and human isolates . These two residues are accessible for antibody binding and are located in domain III of the E protein , which has been shown to be the most important region eliciting neutralizing antibody . Antibodies targeting domain III of E protein make up a relatively small proportion of the polyclonal human antibody response against flaviviruses , thus more detail studies are need to clarity the role of residues 327 and 366 [42] . GI virus-specific residues at position 327 and 366 located in the BC and DE loops on domain III of E protein , respectively , may play an important role in eliciting genotype-specific neutralizing antibodies . Genotype- and strain-specific neutralizing MAbs have been characterized against dengue virus serotypes 1 , 2 and 3 [36]–[38] . Monoclonal antibody E104 , derived from mice immunized with GII of DENV-1 virus , is a genotype-specific MAb , which recognizes residues 328 , 330 , 361 and 362 in the BC and DE loops of EDIII . Residues 328 and 329 , located in the BC loop , are recognized by another DENV-3-derived genotype-specific MAb . Thus , the amino acids located in the BC and DE loops on domain III of E protein may involve in the induction of genotype-specific neutralizing antibodies . Non-neutralizing antibodies are a possible risk factor for ADE in dengue pathogenesis during infection with heterologous or homologous dengue viruses , but the role of ADE contributes to JE disease is unclear [43] . Vaccine-induced enhancements of virus infection have been documented for members of different virus families as well [26] . Our preliminary in vitro and ex vivo studies suggest that the potential of ADE with vaccinees serum specimens may be increased due to GIII to GI replacement . Prior to the genotype replacement in Taiwan , the average mortality rate of confirmed cases of JEV infection was 7 . 8% ( 2000–2008 ) . However , during 2009–2010 , after GI JEV became the dominant circulating virus , the average mortality rate increased to 14 . 2% ( Official statistics of the Department of Health , Taiwan ) . The neurovirulence of GI and GIII viruses was similar and might not directly associate with an increasing case fatality rate of JE cases [19] , [20] . The potential for ADE , measured by in vitro assays and supplemented by the limited number of ex vivo assays , increases dramatically when the neutralization titer of vaccinee serum decreases to below the protective threshold of 1∶10 against the vaccine strain . The statistical differences we have seen in older children , reflecting waning of neutralizing antibody to GI virus to less than seroprotective threshold ( Figure 2 ) . The presence of memory B cells , CD4+ and CD8+ T cells , and anamnestic response has been indicated after received JEV inactivated vaccine [44] , but rapidly decline of seroprotection rate against field-isolated JEVs also has been suggested [45] . Thus , the duration of immunity among vaccinated adults should be evaluated comprehensively ten years after the final booster vaccination . The conclusion of current study is limited by the small sample size and the volume of vaccinee's serum . With the Institution Approved Protocol in the further , we plan to increase the sample size and collect larger serum volume to increase more selection of GIII and GI strains in the analysis . The most cost-effective control strategy for JE is vaccination , but genotype replacement in JEV endemic/epidemic regions may reduce the efficacy of traditional GIII virus-based vaccines . The efficacy of GIII JEV vaccines should be closely monitored at the national or regional level . The potential impact due to genotype replacement could be overcome in the future by 1 ) increasing the effective immunogenic dose or incorporating a novel adjuvant in the vaccine formulation to improve the immunogenicity of the current vaccine , or 2 ) replacing the GIII vaccine strain with a dominant GI isolate and conducting a non-inferiority study of GI strain-specific immune responses . | Genotype I ( GI ) Japanese encephalitis virus ( JEV ) that replaced GIII virus has become the dominant circulating virus in Asia; however , all available JEV vaccines are derived from genotype III viruses , and no study has been conducted on the cross-neutralization and protection elicited by GIII JEV vaccines against GI viruses using vaccinated children’s serum specimens collected from the general population . Genotype I virus was first detected in Taiwan in 2008 , and became the dominant circulating JEV , and was island-wide within a year . In the present study , the small panel of GIII virus vaccinated-children serum specimens were not only showed lower strain-specific neutralization against GI virus as compared to the GIII vaccine and human isolates but also observed the enhancement of GI virus infection in K562 cells in some low or non-neutralizing serum specimens . These preliminary results indicated the reduced neutralization potency due to genotype replacement should be closely monitored in the JE epidemic/endemic regions in the future . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"viral",
"vaccines",
"clinical",
"immunology",
"immunity",
"virology",
"immunology",
"biology",
"microbiology"
]
| 2012 | Partially Neutralizing Potency against Emerging Genotype I Virus among Children Received Formalin-Inactivated Japanese Encephalitis Virus Vaccine |
MicroRNAs are short , noncoding RNAs that play important roles in post-transcriptional gene regulation . Although many functions of microRNAs in plants and animals have been revealed in recent years , the transcriptional mechanism of microRNA genes is not well-understood . To elucidate the transcriptional regulation of microRNA genes , we study and characterize , in a genome scale , the promoters of intergenic microRNA genes in Caenorhabditis elegans , Homo sapiens , Arabidopsis thaliana , and Oryza sativa . We show that most known microRNA genes in these four species have the same type of promoters as protein-coding genes have . To further characterize the promoters of microRNA genes , we developed a novel promoter prediction method , called common query voting ( CoVote ) , which is more effective than available promoter prediction methods . Using this new method , we identify putative core promoters of most known microRNA genes in the four model species . Moreover , we characterize the promoters of microRNA genes in these four species . We discover many significant , characteristic sequence motifs in these core promoters , several of which match or resemble the known cis-acting elements for transcription initiation . Among these motifs , some are conserved across different species while some are specific to microRNA genes of individual species .
MicroRNAs are endogenous single-stranded RNAs ranging from 19–25 nt in length . They are generated from long precursors , which fold into hairpin structures , and are known to repress post-transcriptional gene expression in both animals and plants [1 , 2] . The two well-understood microRNAs , lin-4 and let-7 , were discovered in the 1990s , and proved to regulate developmental timing in C . elegans by repressing the translation of a family of key mRNAs [3–5] . Since then , several hundred microRNAs have been identified in viruses , plants , and animals , and their important post-transcriptional regulatory functions have been discovered . The biogenesis of microRNAs is complex . Most microRNAs are encoded in their own genes situated in intergenic regions or located on the antisense strands of annotated genes [6–8] . The intergenic microRNA genes are believed to be transcribed independently and to form a new gene family , whereas the intronic ones and the ones interspersed with mobile elements Alu in the human genome can be transcribed with their host genes [9 , 10] . Our knowledge of post-transcriptional processing of microRNAs has greatly expanded in recent years through various studies [11–14] . However , we have limited understanding of the transcription of microRNA genes , which is the first , and an important , step of microRNA biogenesis . In this study , we are interested in the known microRNA genes that contain their own transcriptional units . Many pieces of evidence have indirectly suggested that microRNA genes are class-II genes ( i . e . , genes transcribed by RNA polymerase II ( pol II ) ) . For instance , primary transcripts of some microRNA genes contain poly ( A ) tails , or the cap structure [15 , 16] . Expressions of some microRNA genes are regulated by enhancers [17 , 18] or hormones [19] . Lee et al . reported the first direct evidence from an experiment on a single polycistronic microRNA gene , mir-23a∼27a∼24–2 , showing that it can be transcribed by pol II [20] . They also determined the promoter and terminator regions of this gene . However , their results , especially those on the promoter of mir-23a∼27a∼24–2 , do not match very well with our knowledge of pol II promoters . Specifically , the promoter of mir-23a∼27a∼24–2 appears to lack the known common promoter elements required for initiating transcription , such as the TATA-box , initiator element , downstream promoter element ( DPE ) , TFIIB recognition element ( BRE ) [20] , or the proximal sequence element ( PSE ) . Additionally , they also found that a large portion of a given pri-microRNA ( the primary transcript of an microRNA gene ) does not contain a 5′ cap or a poly ( A ) tail [20] . Another piece of experimental evidence was from a M . musculus polycistronic microRNA gene , mmu-mir-290∼291∼292∼293∼294∼295 . Houbaviy et al . found a canonical TATA-box , located at −35 , of capped and polyadenylated pri-microRNA of this gene , and showed that this upstream region was also conserved in a H . sapiens homologous gene , hsamir-371∼372∼373 [21] . Furthermore , Xie et al . identified the promoters of 52 A . thaliana microRNA genes , and showed that most of them have TATA-boxes in their core promoters [22] . All these results are fundamentally important; they have provided direct evidence that a microRNA gene can be transcribed by pol II . However , a few critical questions remain unanswered . One of them is whether all known microRNA genes of different species are class-II genes . Although more than 50 A . thaliana microRNA genes have been shown to be transcribed by pol II , our knowledge of the transcription of microRNA genes in animals is still limited . We consider this important issue through a genome-wide computational analysis on four model species , C . elegans , H . sapiens , A . thaliana , and O . sativa . Our overall strategy is based on the following perspective on transcriptional regulation . Class-II genes and class-III genes ( genes transcribed by RNA polymerase III ) must have distinctive features in their promoter regions , including transcription factor binding motifs , to recruit the right transcriptional machineries to initiate their transcription . Based on this perspective and supported in part by the results in [20–22] , we first assume that the core promoters of intergenic microRNA genes share common sequence features with the core promoters of the known class-II or class-III genes . We then build computational models to separate the core promoters of class-II and class-III genes as well as random sequences . Using these models , we test all known intergenic microRNA genes in the four species to determine what types of promoters they have . We subsequently answer the question: which RNA polymerase is responsible for the transcription of these microRNA genes ? The promoter of a gene is a crucial control region for its transcription initiation [23 , 24] . To understand the mechanism and conditions of the activation of microRNA genes , it is required to locate their core promoter regions . One practical way to identify core promoters of microRNA genes is to first apply a promoter prediction method to predict their core promoters , and then to verify the predictions by wet lab experiments . Developing the promoter identification algorithm is a very challenging problem . Although computational methods have been developed for predicting core promoters of protein-coding genes , their performances are far from satisfactory . The main reason is that our understanding of the transcription process is incomplete . The situation with microRNA genes is even worse . All existing promoter prediction methods for protein-coding genes may not be suitable for microRNA genes , since they were not built based on the core promoters of microRNA genes . Furthermore , the promoters of most microRNA genes in all species remain undefined . For H . sapiens , only the promoters of two microRNA genes , hsa-mir23a∼27a∼24–2 [20] and hsa-mir-371∼372∼373 [21] , have been identified so far . The promoter of hsa-mir-23a∼27a∼24–2 has been located by biological experiments [20] , while the promoter of hsa-mir-371∼372∼373 [21] has been identified by a comparative genomic analysis . The 52 microRNA genes in A . thaliana studied in [22] are not sufficient to build a good predictive model . Core promoter regions contain essential components for the regulation of gene transcription [23 , 24] . The basal transcription machinery , comprising the multisubunit RNA polymerase and several auxiliary factors , is thought to interact directly with core promoter elements [23 , 24] . Thus , revealing functional regulatory binding sites in promoter regions is important for determining promoter structures and characterizing transcriptional regulation . However , core promoter elements are highly variable , requiring sophisticated techniques for their detection . Discovering key cis-elements of microRNA genes is more difficult , since our knowledge about the transcription of this novel family of genes is limited . Lee et al . located the promoter of mir-23a∼27a∼24–2; however , none of the canonical promoter elements were discovered in this promoter [20] . TATA-box was found in mmu-mir-290∼291∼292∼293∼294∼295 [21] . However , the deletion of this putative TATA-containing promoter region had almost no effect on the expression level of mir292 and the precursor to mir292 in transfected cell lines [21] . Ohler et al . scanned the 1 , 000-bp upstream sequences of Drosophila microRNA genes for known promoter motifs , but did not detect a consistent preference for any known motifs that are enriched in protein-coding genes [25] . In this study , we propose a novel promoter prediction approach , CoVote ( common query voting ) , for predicting microRNA core promoters . Using CoVote , we investigate core promoter regions of microRNA genes in C . elegans , H . sapiens , A . thaliana , and O . sativa , and further analyze sequence motifs in the putative core promoters that may be involved in the transcription of microRNA genes . Our objectives are to ( 1 ) identify characteristic motifs in core promoters of known microRNA genes in these four species , and ( 2 ) compare the potential promoter structure of microRNA genes in different species . We examine the presence and distribution of conserved motifs in these species , and also investigate species-specific motifs .
Two discriminative models were built and used in our study . The first model ( the three-class model , discussed in Discriminative Models of Pol II and Pol III Promoters ) is for discriminating the promoters of genes transcribed by RNA polymerases II ( pol II promoters ) and the promoters of genes transcribed by RNA polymerases III ( pol III promoters ) , as well as random sequences . To build this model , we prepared training sequences of three different types: known pol II core promoter sequences , known pol III core promoter sequences , and random sequences . The numbers of these sequences are listed in Table 1 . The second model is for identifying putative promoters of microRNA genes . This model only needs to separate pol II promoter sequences and random sequences ( see The CoVote Algorithm for Locating Core Promoter Regions of MicroRNA Genes ) . Therefore , we only used these two types of sequences as training data . The pol II sequences were downloaded from the Web as of March 2005 . The C . elegans core pol II promoters were retrieved from C . elegans promoter database ( CEPDB ) ( http://rulai . cshl . edu/cgi-bin/CEPDB/home . cgi ) . The H . sapiens pol II promoters were downloaded from the Eukaryotic Promoter Database ( EPD ) ( http://www . epd . isb-sib . ch/seq_download . html ) . The plant core pol II promoters were obtained from Plant Promoter Database ( PlantProm ) ( http://mendel . cs . rhul . ac . uk/mendel . php ? topic=plantprom ) . All these sequences are 250 bp long and cover the regions from −200 bp to +50 bp with respect to the corresponding transcription start sites . The known core promoter sequences of A . thaliana and O . sativa are not sufficient to build a discriminative model . As shown in Table 2 , we thus included the pol II promoter sequences from 44 dicotyledonous and seven monocotyledonous plants in our study . Both the discriminative model for pol II and pol III promoters and the promoter prediction model trained with these sequences were applied to A . thaliana and O . sativa . For each species , the pol III promoter sequences that we used included the promoter sequences of tRNAs , U6 snRNAs , 7SL RNAs , and 7SK RNAs ( Table 3 ) . The promoter of each tRNA covered the complete coding region of the tRNA and its upstream sequence with a total length of 250 bp . The promoters of U6 snRNA , 7SL RNA , and 7SK RNA included 200-bp upstream sequences and 50-bp downstream sequences , relative to their transcription start sites ( TSSs ) . The sequences of these ncRNAs were downloaded from the ncRNA database ( http://noncode . bioinfo . org . cn/showclass . php ? class=snRNA ) . Since availability of known pol III promoters is limited , we randomly chose 50 pol III promoter sequences from C . elegans , H . sapiens , and plants , respectively , as independent test sets for corresponding discriminative models . We generated 1 , 000 random sequences of 250 bp length to represent intergenic sequences other than pol II and pol III core promoter sequences . For each species , we used the nucleotide composition of intergenic regions of its genome to generate these sequences . We did not use intergenic sequences from a genome for this purpose because it is difficult to ensure that intergenic sequences do not overlap with real promoter regions . Three independent test sets for each species studied were used to validate the three-class discriminative model . The first set included 1 , 000-bp upstream sequences of 1 , 000 randomly chosen coding genes . These sequences were obtained from RSA Tools ( http://rsat . ulb . ac . be/rsat/ ) . The second set contained the 50 pol III promoters not used in training . The last set of sequences included 1 , 000 randomly generated sequences of 2 , 000 bp length . We applied the nucleotide composition of pol II and pol III promoter sequences to generate 500 sequences , respectively , for each species . Two independent sets were also prepared to validate the promoter prediction model . The first set includes 4 , 189 H . sapiens pol II promoters , downloaded from the Database of Transcriptional Start Sites ( DBTSS ) ( http://dbtss . hgc . jp/samp_home . html ) . The second set contained 4 , 000 sequences randomly chosen from H . sapiens protein coding regions . For each species studied , the upstream sequences of pre-microRNAs ( hairpin precursors ) of the intergenic microRNA genes were obtained as follows . First , when a pre-microRNA and its upstream gene were unidirectional ( same direction ) , if the distance between them was longer than 2 , 400 bp , the 2 , 000-bp sequence upstream of the pre-microRNA was retrieved; otherwise , the sequence between 400 bp downstream of the upstream gene and the precursor was used . Second , when a pre-microRNA and its upstream gene were convergent ( opposite directions ) , if the distance between them was longer than 4 , 000 bp , the 2 , 000-bp sequence upstream of the precursor was obtained; otherwise , the sequence from the precursor and the middle point between the upstream gene and the precursor was retrieved . Some C . elegans and H . sapiens microRNA genes are polycistronic , in which case only upstream sequences of the 5′ pre-microRNAs were considered in our study . In addition to intronic microRNA genes , the ones in human that are interspersed and transcribed with Alu elements were excluded from our analysis . Our overall approach depends on building accurate discriminative models of transcriptional regulation , which in turn rely on sequence features . We may simply use all possible k-mers , with reasonable values of k , as such features . However , not all k-mers have the same amount of information , and the number of k-mers increases exponentially with k . The key then is to find a sufficient number of statistically overrepresented motifs in the sequences of interest . We used the WordSpy algorithm developed by Wang et al . [26 , 27] to find significant motifs , for several reasons . Statistical modeling and word counting methods have been integrated in WordSpy; it is able to build a dictionary of a large number of statistically significant motifs . WordSpy adopts a strategy of steganalysis , which is a technique for discovering hidden patterns and information from a medium such as strings , so that it does not have to rely on additional background sequences and is still able to find motifs of nearly exact lengths . It is believed that Pol II and Pol III transcribe different types of genes whose promoters are intrinsically different from each other and from other genomic sequences [23] . Therefore , it is viable to assume that the core promoters of these two classes of genes have discriminative sequence features that separate them from each other and from the other genomic sequences . Consequently , a discriminative model can be built using the known promoters of these two types of genes , and be used to determine if query sequences are pol II promoters , pol III promoters , or other intergenic sequences . Specifically , we built a three-class discriminative model , or classifier , to distinguish pol II promoters , pol III promoters , and random intergenic sequences for each of the four species that we studied , i . e . , C . elegans , H . sapiens , A . thaliana , and O . sativa . We extracted statistically overrepresented sequence motifs of 5–10-bp length from each training set separately , using the WordSpy motif-finding algorithm [26] . With these sequence motifs as features , we represented each promoter sequence as a vector , where an entry in the vector was the number of occurrences of a motif in the sequence . We then built two classifiers for each species , one using a decision tree [28] , the other using a support vector machine ( SVM ) [29] to separate the three types of sequences . We adopted these two well-studied classification methods to ensure that our analysis of microRNA genes is not skewed by the computational methods used . We applied the SVM implementation in the WEKA software package [30] under its default setting . We tested linear , polynomial , and radial kernels [29] . Although the cross-validation accuracies of the polynomial and radial kernels were slightly better than that of the linear kernel , we used the linear kernel due to its simplicity . For the decision tree learning , we applied the J48 program in WEKA [30] , which is an implementation of the well-known C4 . 5 algorithm [28] . To prevent overfitting , we required each leaf node to have at least five sequences . The accuracies of the discriminative models were estimated using a 10-fold cross-validation . In this process , a training set was randomly partitioned into ten roughly equal-sized subsets . Each subset was then used in turn as a test set to estimate the prediction quality of the model built with the other nine subsets . The average quality of these tests was the final accuracy measure . To measure prediction quality , we calculated recall , precision , and overall accuracy for each type of sequence . The recall for pol II promoters ( respectively , III ) was defined as the ratio of the number of correctly predicted pol II ( respectively , III ) sequences versus the total number of pol II ( respectively , III ) sequences tested . The precision was defined as the ratio of the number of correctly predicted pol II ( respectively , III ) sequences versus the total number of predicted pol II ( respectively , III ) sequences . The overall accuracy was defined as the number of correctly predicted sequences versus the total number of sequences tested . When we applied the discriminative models to predict the type of promoter that a query gene may have , the upstream sequence of the query gene was fragmented using a sliding window of 250 bp , with an increment of 50 bp . Each segment was then tested by the discriminative models separately . The experimental results were organized in five categories . The first category contained the upstream sequences in which at least one of the 250-bp segments was classified as pol II promoter and none of the rest were predicted as pol III promoter . This class , called definitive pol II class , provided the definitive evidence for class-II genes . The second category had the sequences in which some of the segments were classified as pol II and some as pol III promoters , but there were more pol II segments than pol III segments . We called this category possible pol II class , since we simply classified a sequence to be a pol II promoter based on the majority prediction for its segments . The next category , called possible pol III class , was similar to the second , but the number of pol III segments was greater than the number of pol II segments . The fourth category , called definitive pol III class , had sequences in which at least one segment was a pol III promoter but none of the rest was predicted as a pol II promoter . The last category , called random class , contained sequences with all segments classified as random promoters . Our method , which we called common query voting , shorthanded as CoVote , is based on the following understanding of the promoters of the microRNA gene . MicroRNA genes have the same type of promoters as other class-II genes , as shown in this paper and in [20–22] . Therefore , there must be characteristic sequence features in the core promoters of microRNA genes with respect to random sequences that have the same nucleotide compositions of intergenic sequences . Moreover , compared with other upstream regions , core promoters should be the most similar upstream regions among most , if not all , microRNA genes . Although the promoters of microRNA genes have some similar , or even the same , features as promoters of the known class-II genes , they may have their own unique features that have not been discovered . Compared with many existing promoter prediction methods , CoVote not only takes into account the features that the training instances have , but also captures potential common features in many query instances . The CoVote algorithm runs as follows . We applied the WordSpy algorithm to identify significant motifs from putative core microRNA promoters . Furthermore , in addition to WordSpy , we also applied the popular MEME algorithm [31] with its default parameters to find 20 top-ranking degenerate motifs for each species considered . It is critical to ensure that the motifs from putative core microRNA promoters are indeed specific to promoters . For this purpose , we used a whole-genome Monte Carlo simulation to measure the specificity and significance of a motif in the putative promoters , which we call target set , with respect to a set of different sequences , which we call reference set . A reference set can be drawn from other regions of a genome . For example , in this research , we randomly chose reference sets from open reading frames ( ORFs ) and other genome regions . Given a motif of interest , we computed its Z-score with respect to other regions of the genome as follows . We first obtained the average number of occurrences per target sequence for the motif , denoted as Nt . We then randomly generated a large number of reference sets and computed the average number of occurrences of the motif , Nr , and its standard deviation , σr , over the reference sets . The Z-score was then calculated as Z = ( Nt/Nr ) = σr . Here , we set the size of a reference set to be the same as that of the target set . Therefore , all the reference sets can be considered as independently and identically distributed , and follow a normal distribution when the number of samples is large . Consequently , the Z-score simply measures the normalized difference between the average occurrence of the motif in the target set and the sample mean in the reference sets . For example , if the Z-score is 2 , the specificity of the motif to the target set is two times the standard deviation to the example mean of the reference sets .
We evaluated the quality of the three-class discriminative models in terms of recall , precision , and accuracy ( see Discriminative Models of Pol II and Pol III Promoters ) . Table 4 lists the 10-fold cross-validation results of the SVM and decision tree–based classifiers . The results show that these discriminative models are fairly accurate , with the minimum accuracy greater than 96% for the SVM models and greater than 87% for the decision tree models . The SVM models are marginally better than the decision-tree models . To further examine the accuracy of the models , we assessed the error rates by control experiments on independent test sets ( see Datasets ) . The decision-tree models have comparable but slightly worse classification accuracies than the SVM models , so the results are omitted . For each of the three SVM-based models , their accuracies were examined on three independent test sets . The first set includes promoter sequences of randomly chosen protein coding genes . Since the protein coding genes contain pol II promoters , the percentage of protein coding genes predicted to have pol III promoters will reflect the error rates of these discriminative models . The error rates of the SVM models are shown in Table 5 . Among 1 , 000 coding genes , only a handful of them were predicted to have possible pol III or definitive pol III promoters ( i . e . , eight C . elegans genes , 25 H . sapiens genes , and 31 plant genes ) . The second independent set contains 1 , 000 random sequences of 2 , 000 bp length . Half of these sequences have the same nucleotide composition as pol II promoter sequences , while the other half have the same nucleotide composition as pol III promoter sequences . We used randomly generated intergenic sequences instead of real intergenic sequences , since it is difficult to ensure that the intergenic sequences do not to overlap with real promoter regions . As shown in Table 5 , the error rates of the discriminative models on randomly generated sequences for C . elegans , H . sapiens , and plants are 6 . 4% , 10 . 8% , and 7 . 7% , respectively . Moreover , since experimentally verified pol III promoters are very limited , we saved 50 pol III promoter sequences from C . elegans , H . sapiens , and plants , respectively , as independent test sets . As shown in Table 5 , for the discriminative models on pol III promoters from C . elegans , H . sapiens , and plants , the error rates are 2% , 0% , and 2% , respectively . Based on the cross-validation and these three independent tests , we can conclude that ( 1 ) pol II and pol III promoters can be separated from each other and are also distinguishable from random intergenic sequences , and ( 2 ) the quality of the discriminative models that we developed is sufficiently high . To determine the promoter types of the known intergenic microRNA genes of the four model species , we conducted two experiments using the three-class discriminative models that we developed . We considered separately the precursors ( pre-microRNAs ) and primary transcripts ( pri-microRNAs ) of known microRNAs . We analyzed upstream sequences up to 2 , 000 bp of these transcripts . As described in the section Discriminative Models of Pol II and Pol III Promoters , these upstream sequences were fragmented using a sliding window of 250 bp , with an increment of 50 bp . Each segment was then tested by the discriminative models separately , and the experimental results were organized into five categories: definitive pol II class , possible pol II class , possible pol III class , definitive pol III class , and random class , as discussed in Discriminative Models of Pol II and Pol III Promoters . Table 6 shows the results on the four species using the SVM models . The results from the decision tree models were similar . We tested 73 C . elegans , 109 H . sapiens , 112 A . thaliana , and 114 O . sativa pre-microRNAs that are in intergenic regions according to the genome annotation as of March 2005 . Among them , 67 ( 91 . 8% ) C . elegans , 81 ( 74 . 3% ) H . sapiens , 81 ( 72 . 3% ) A . thaliana , and 92 ( 80 . 7% ) O . sativa microRNAs have definitive pol II class promoters . These results suggest that most microRNA genes in the four species have the same promoters as protein coding genes . However , six ( 8 . 2% ) , 24 ( 22% ) , 17 ( 15 . 2% ) , and 12 ( 10 . 5% ) microRNAs of these species have possible pol II class promoters , respectively . One H . sapiens , three A . thaliana , and one O . sativa microRNA genes were predicted to have possible pol III promoters . In the upstream regions of these microRNA genes , some segments were predicted to be pol II promoters while some were predicted to be pol III promoters . Combining the microRNAs in these two categories , 73 ( 100% ) C . elegans , 105 ( 96 . 3% ) H . sapiens , 98 ( 87 . 5% ) A . thaliana , and 104 ( 91 . 2% ) O . sativa microRNA genes have pol II promoters . Importantly , none of the microRNA genes were predicted to have a definitive pol III promoter , and only one H . sapiens , three A . thaliana , and one O . sativa microRNA genes were predicted to have possible pol III promoters . Similar results , shown in Table 6 , were obtained on H . sapiens and A . thaliana pri-microRNAs . We expected the results based on pri-microRNAs to be more definitive than those from pre-microRNAs . However , we were only able to find 13 pri-microRNAs for H . sapiens and 19 pri-microRNAs for A . thaliana . It is difficult to draw a meaningful conclusion based on such limited samples . Nevertheless , as shown in Table 6 , nine out of 13 ( 69 . 2% ) H . sapiens microRNAs and 16 out of 19 ( 84 . 2% ) A . thaliana microRNAs were predicted to have definitive pol II promoters . These results provided genome-wide evidence that most microRNA genes are class-II genes and have pol II promoters . This is consistent with the previous study on a polycistronic H . sapiens microRNA gene , mir-23a∼27a∼24–2 [20] , and the report on some A . thaliana microRNA genes [22] . In this research , we developed a novel computational , sequence-centric method , CoVote , for identifying the core promoter regions of microRNA genes , as described in the section The CoVote Algorithm for Locating Core Promoter Regions of MicroRNA Genes . Using CoVote , we predicted putative core promoters for most known microRNA genes of the four species . Specifically , we predicted promoters for all of the 73 tested C . elegans microRNA genes , 107 ( 98 . 2% ) of 109 tested H . sapiens microRNA genes , 95 ( 84 . 8% ) of 112 tested A . thaliana microRNA genes , and all of the 114 tested O . sativa microRNA genes . Among the microRNA genes whose promoters were identified by CoVote , some were predicted to contain multiple core promoter regions . Figure 1 shows the distributions of the positions of putative promoters with respect to corresponding microRNA foldbacks ( the first foldbacks of polycistronic microRNA genes ) . In short , 70 ( 95 . 9% ) of 73 C . elegans microRNA genes , 100 ( 93 . 5% ) of the 107 H . sapiens microRNA genes , 80 ( 84 . 2% ) of 95 A . thaliana microRNA genes , and 109 of 114 ( 96 . 6% ) O . sativa microRNA genes have putative promoters within 500 bp of upstream regions . This distribution pattern may imply that real core promoters of most microRNA genes are close to pre-microRNA hairpins . Recently , Xie et al . experimentally identified 65 core promoters of 52 A . thaliana microRNA genes ( multiple transcription start sites were reported for some of these genes ) [22] . As shown in Table 7 , CoVote correctly identified 51 ( 78 . 5% ) of these 65 known core promoter sequences . For 40 out of these 52 ( 76 . 9% ) A . thaliana microRNA genes , CoVote predicted at least one core promoter region correctly . This analysis shows that our new promoter prediction method is fairly accurate . In comparison , TSSP ( SoftBerry , http://www . softberry . com ) , which is one of the best promoter prediction methods for plants , only identified 39 ( 60% ) promoters for 34 ( 65 . 4% ) of these microRNA genes . Therefore , CoVote outperformed TSSP in this study . Using a comparative genomics approach , Ohler et al . studied the flaking sequences of 43 pairs of orthologous C . elegans and C . briggsae pre-microRNAs , and reported ∼250 bp conserved regions located around 200 bp upstream of the foldbacks [25] . In this study , we found that these conserved regions significantly overlapped with our predicted core promoter regions . In addition , the promoters of two microRNA genes in H . sapiens , hsa-mir-23a∼27a∼24–2 , and hsa-mir-371∼372∼373 , reported in [21 , 20] , were also correctly predicted in our analysis . The accuracy and false positive rate of CoVote were also assessed by known H . sapiens core promoters from DBTSS [32] ( positive test set ) and coding sequences ( negative test set ) . The known core promoters of 4 , 189 H . sapiens protein-coding genes in the positive set were all correctly predicted . Ideally , we should evaluate false positive rates of these models with intergenic sequences that do not contain any promoters . However , it is difficult to obtain such intergenic sequences . Thus , we randomly chose 4 , 000 coding sequences as a negative control . For these , 4 , 000 negative test sequences , 1 , 325 ( 33 . 1% ) were predicted to be core promoters , which gives the false positive rate of this method , although some of the predictions may be real . To further characterize the predicted microRNA core promoters and gain a deep insight into microRNA transcriptional regulation , we performed a motif analysis to identify statistically significant and biologically meaningful motifs in the putative promoters . As shown in Figure 1 , most putative promoters are located within the 500-bp upstream regions of pre-microRNA foldbacks . Therefore , for the microRNA genes that have multiple predicted promoter regions , we chose those promoters within the 500-bp upstream proximal regions of pre-microRNA hairpins for motif analysis . For those genes that do not have putative promoters within the 500-bp upstream regions , the promoters closest to the precursors were used . In our study , we first applied two motif-finding algorithms , MEME [31] and WordSpy [26 , 27] , to identify statistically overrepresented motifs . MEME is a statistical model–based algorithm for finding degenerate motifs , while WordSpy is a dictionary-based algorithm for finding a large number of exact motifs of high fidelity . We then conducted a whole-genome , Monte Carlo analysis to assess the biological relevance and specificity of the identified motifs to the core promoter regions of interest ( see Motif Analysis ) . The motifs with Z-scores smaller than 3 . 0 were discarded , since they may also be prevalent in coding regions and/or other intergenic regions . The remaining ones are core promoter–specific motifs and likely to be biologically relevant to the transcriptional regulation of microRNA genes . Figure 2 lists some significant motifs that were identified by both motif-finding approaches and that were also reported in the literature as significant motifs in promoters of protein-coding genes . The whole list of motifs from WordSpy is given at http://cic . cs . wustl . edu/microrna/promoters . html . Many motifs from WordSpy match well with the motifs from MEME . In C . elegans , one of the most significant motifs identified by MEME has a consensus TTTCAATTTTTC ( motif 1 , Figure 2 ) , which appears in 69 of the 73 predicted promoters . This motif matches the Inr ( initiator ) element , which has a weak consensus PyPyPyCANPyPyPyPyPy [23 , 24] . MEME also identified a significant motif in H . sapiens microRNAs that resembles the Inr element . This motif has a consensus CCCCACCTCC ( motif 3 , Figure 2 ) , which appears in 78 putative promoters of H . sapiens microRNA genes . Wordspy also discovered several Inr-like motifs in both species . TATA-box , which is one of the most well-known motifs in the core promoters of eukaryotic class-II genes , was discovered in A . thaliana and O . sativa ( motifs 6 and 10 , Figure 2 ) . Among the 95 A . thaliana microRNA genes whose promoters were predicted by CoVote , 81 ( 85 . 3% ) contain TATA-box . This observation is consistent with the experimental result in [22] . Specifically , Xie et al . reported that 42 ( 86 . 5% ) of 52 A . thaliana microRNA genes contained TATA-box in their promoters [22] . In O . sativa , 84 of 114 ( 73 . 7% ) microRNA genes contain TATA-box in their promoters . Although MEME did not report TATA-box in the promoters of C . elegans and H . sapiens microRNA genes , WordSpy identified it as a significant motif . We further scanned the putative promoters of C . elegans and H . sapiens microRNA genes with the TATA-box weight matrix curated in the Eukaryotic Promoter Database ( EPD ) ( http://www . epd . isb-sib . ch ) . Including hsa-mir-371∼372∼373 , whose promoter regions were analyzed by Houbaviy et al . [21] , 35 ( 33% ) of 107 H . sapiens microRNA genes and 34 ( 47% ) of 73 C . elegans microRNA genes contain the canonical TATA-box in their promoters . The Z-scores of TATA-box in the promoters of microRNA genes in H . sapiens and C . elegans are 8 . 4 and 3 . 38 , respectively , showing that TATA-box is a significant motif in the promoters of microRNA genes in these two species . Note that the frequency of TATA-box in plant microRNAs is nearly twice of that in animal microRNAs . This discrepancy deserves some further investigations . Interestingly , CT-repeat microsatellites are significant motifs in the putative promoters of all four species ( motifs 2 , 4 , 5 , 7 , 8 , 9 , 11 , 12 , and 13 , Figure 2 ) . To elucidate the significance of CT repeats in microRNA gene promoters , we performed several additional analyses . First , we analyzed the occurrences of CT repeats in the 2 , 000-bp upstream sequences of pre-microRNAs in all four species . As shown in Figure 3 , in all four species tested , most microRNA genes have CT repeats in the 500-bp upstream regions of microRNA foldbacks . Second , we estimated the expected frequencies of CT repeats in the whole genomes of these species by a Monte Carlo simulation . Briefly , for each species , we randomly sampled n sequences with a length of 500 bp from its genome , where n was the number of microRNA genes whose upstream regions were analyzed for occurrences of CT repeats . Both strands of the genome sequences were scanned with the matrices of CT-repeat motifs listed in Figure 2 and other predefined CT-repeat sequences , including ( CT ) n , ( CCT ) n , ( CTT ) n , ( CCTT ) n , ( CGCT ) n , ( CCTCG ) n , ( CCTCT ) n , ( CGTCT ) n , and ( CTCTT ) n [33–36] . We then calculated the percentage of these sequences that contain CT repeats . We repeated the sampling 10 , 000 times , and computed the average percentage and the standard deviation of CT-repeat occurrences . As shown in Figure 3 , in each of these four species the expected frequency in the whole genome is much lower than that in the promoter regions of microRNA genes . We also analyzed the distribution of CT repeats in the experimentally identified promoters of the 52 A . thaliana microRNA genes [22] , and calculated the distances between the CT repeats and the TSSs . As shown in Table 8 , 40 of these 52 genes contain CT repeats; in 30 of these 52 genes , the distances between CT repeats and TSSs are less than 100 bp . Additionally , the experimentally identified promoter regions of two H . sapiens microRNA genes , hsa-mir-23a∼27a∼24–2 [20] and hsa-mir-371∼372∼373 [21] , contain CT repeats . The −56 to −34 upstream region of has-mir-23a∼27a∼24–2 is CTCTCTCTCTCTTTCTCCCCTCC [20] . The −43 to −34 upstream region of hsa-mir-371∼372∼373 , which is located closely nearby in the upstream of the reported TATA-box , contains a shorter CT repeat , CTCTCACCCT [21] . It has been shown that CT repeats are functional elements in the promoters of protein-coding genes in many mammalian species [37–40] , Gallus gallus [41–43] , and Drosophilia melanogaster [34 , 44 , 45] . Similar CT-repeat microsatellites in the core promoter regions of protein coding genes were also reported recently in A . thaliana and O . sativa [33 , 35 , 36] . Furthermore , initiator elements are pyrimidine-rich and contain CT repeats [45 , 42] . From a structure viewpoint , CT repeats can form non–B-DNA , which may potentially play important roles in gene transcription activation [46 , 47] . The frequent occurrence and the conservation across all four tested species suggest that CT repeats may play an important role in the transcription of microRNA genes . A CpG island is one of the significant characteristics in the promoters of Eukaryotic class-II genes . We analyzed the presence of CpG islands in the upstream sequences of pre-microRNAs in all four species , as well as in the upstream sequences of 49 C . briggsae and 113 M . musculus microRNA genes . C . briggsae and M . musculus microRNA genes were included in order to form three pairs of evolutionarily closely related species , C . elegans versus C . briggsae , H . sapiens versus M . musculus , and A . thaliana versus O . sativa , for conservation analysis . We first identified CpG islands with CpGProD [48] and further confirmed the results with CpGPlot ( http://bioweb . pasteur . fr/seqanal/interfaces/cpgplot . html ) . As shown in Table 9 , a small number of microRNA genes in these species , except A . thaliana , have CpG islands in their upstream regions . The list of microRNA genes that contain CpG islands in their upstream sequences is given at http://cic . cs . wustl . edu/microrna/promoters . html . Two interesting observations are worth mentioning . First , CpG islands are often located close to pre-microRNA hairpins . Second , for most CpG-island–containing microRNA genes , their corresponding orthologous genes in closely related species also contain CpG islands in the upstream sequences . This may imply that CpG islands are evolutionarily conserved to a certain degree in these microRNA genes , and may be involved in the regulation of microRNA genes . However , none of the A . thaliana microRNA genes contain CpG islands , whereas 25 O . sativa microRNA genes do . It has been estimated that , in mammals , CpG islands are associated with approximately half of the promoters of protein coding genes [23] . CpG islands are frequently associated with ubiquitously expressed housekeeping genes [23]; thus , their roles in the regulation of those microRNA genes require further study . Besides these conserved motifs , we also discovered several significant motifs that are specific to one of the four species studied . Two motifs ( motifs 1 and 2 , Figure 4 ) are specific to C . elegans microRNA genes , which match the consensus sequences of two motifs ( CTCCGCCC and GCGTGGCS , S = C or G ) conserved in the upstream of 43 pairs of C . elegans and C . briggsae orthologous microRNA genes [25] . A novel motif ( motif 7 , Figure 4 ) appears specifically in promoter regions of 61 A . thaliana microRNA genes . We further analyzed the distribution of this motif in the experimentally identified promoters of the 52 A . thaliana microRNA genes [22]: 24 promoters of 20 microRNA genes contain this motif . In most of these 24 promoters , the distance between this motif and TSS is smaller than 100 bp . Among four motifs that are specific to the promoters of O . sativa microRNA genes , motifs 9 and 10 in Figure 4 are known plant motifs reported in the literature . Motif 9 is an RY-repeat , which is conserved in the promoters of seed-specific genes in both monocot and dicot species [49–51] . Motif 10 has been found in the promoters of some anaerobic genes involved in the fermentative pathway of different plant species [52] . Motif 11 has been reported to be the binding site of HNF6 ( Hepatocyte nuclear factor-6 ) in human and mouse by ChIP–chip experiments [53] , while its function in plants remains unknown . There are two additional interesting observations on the motifs specific to O . sativa microRNA genes . First , all O . sativa motifs in Figure 4 have repetitive patterns in their consensus . Motif 8 has two copies of GCTA , motif 9 contains two copies of CATG , motif 10 can be viewed as CTG-repeats , and motif 11 has two copies of CGAT . Second , motifs 8 , 9 , and 11 are palindromic . Since palindromic patterns have been shown in binding sites of some transcription factors such as nuclear receptors in mammalian species [54] , it may suggest that these three motifs are involved in the transcription of microRNA genes . In additional , four novel motifs discovered in the putative promoters of H . sapiens microRNA genes are all functionally unknown and need further study . Sequence similarities in promoters of Arabidopsis-specific microRNA genes have been addressed [55] . Therefore , although the functions of these species-specific motifs remain unclear , they will be important assets for future research , such as developing a new method for genome-wide identification of novel microRNA genes and conducting a wet lab microRNA analysis . In summary , we extensively analyzed the promoters of the known intergenic microRNA genes in four model species , C . elegans , H . sapiens , A . thaliana , and O . sativa . The genome-wide evidence from these four species showed that most , if not all , microRNA genes have the same type of promoters as protein-coding genes , and therefore are very likely to be transcribed by pol II . Our study extended the results on a small number of individual microRNA genes in H . sapiens [21 , 20] and A . thaliana [22] to all known microRNA genes in the four model species . Moreover , with a new promoter identification method , we also located the core promoter regions of most known microRNA genes of these four species . The position distribution of putative promoters with respect to microRNA hairpins suggests that the core promoters of most microRNA genes are close to corresponding pre-microRNA hairpins ( in the case of polycistronic microRNA genes , core promoters are close to the first pre-microRNA hairpins ) . Furthermore , our extensive motif analysis of these putative promoters identified many cis-elements that are essential to the initiation of gene transcription . CT-repeat microsatellites were found to be conserved in all four species . Inr-like elements , which are relatively common in the promoters of protein-coding genes , were also discovered in the microRNA genes of C . elegans and H . sapiens . On the other hand , our results indicated that TATA-box does not seem to be necessary for most microRNA genes in C . elegans and H . sapiens , although most studied microRNA genes of A . thaliana and O . sativa contain TATA-box . Finally , CpG islands were discovered in a small portion of C . elegans and H . sapiens microRNA genes and their orthologues in C . briggsae and M . musculus , respectively . However , none of the A . thaliana microRNA genes contained CpG islands , although their O . sativa orthologues were found to contain CpG islands in their upstream sequences . Additionally , some motifs were discovered to be specific to individual species studied . We expect our results on the putative promoters and the sequence motifs to be useful for future microRNA prediction and for elucidating the details of the regulation of microRNA gene transcription . Additional supporting results and data files are available at http://cic . cs . wustl . edu/microrna/promoters . html . | MicroRNAs are a class of short RNA sequences that have many regulatory functions in complex organisms such as plants and animals . However , our knowledge of the transcriptional mechanisms of microRNA genes is limited . Here , we analyze the upstream sequences of known microRNA genes in four model species , i . e . , C . elegans , H . sapiens , A . thaliana , and O . sativa , and compare them with the promoter sequences of protein-coding genes and other classes of RNA genes . This analysis provides genome-wide evidence that microRNA genes have the same type of promoter sequences as protein-coding genes , and therefore are likely transcribed by RNA polymerase II ( pol II ) . Second , we present a novel computational method for promoter prediction , which is then applied to locate the core promoters of known microRNA genes in the four model species . Furthermore , we present an analysis of short DNA motifs that appear frequently in the predicted promoters of microRNA genes , and report several interesting motifs that may have some functional meanings . These results are important for understanding the initiation and regulation of microRNA gene transcription . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results/Discussion"
]
| [
"oryza",
"caenorhabditis",
"arabidopsis",
"computational",
"biology",
"homo",
"(human)",
"genetics",
"and",
"genomics"
]
| 2007 | Characterization and Identification of MicroRNA Core Promoters in Four Model Species |
The transduction of sound in the auditory periphery , the cochlea , is inhibited by efferent cholinergic neurons projecting from the brainstem and synapsing directly on mechanosensory hair cells . One fundamental question in auditory neuroscience is what role ( s ) this feedback plays in our ability to hear . In the present study , we have engineered a genetically modified mouse model in which the magnitude and duration of efferent cholinergic effects are increased , and we assess the consequences of this manipulation on cochlear function . We generated the Chrna9L9′T line of knockin mice with a threonine for leucine change ( L9′T ) at position 9′ of the second transmembrane domain of the α9 nicotinic cholinergic subunit , rendering α9-containing receptors that were hypersensitive to acetylcholine and had slower desensitization kinetics . The Chrna9L9′T allele produced a 3-fold prolongation of efferent synaptic currents in vitro . In vivo , Chrna9L9′T mice had baseline elevation of cochlear thresholds and efferent-mediated inhibition of cochlear responses was dramatically enhanced and lengthened: both effects were reversed by strychnine blockade of the α9α10 hair cell nicotinic receptor . Importantly , relative to their wild-type littermates , Chrna9L9′T/L9′T mice showed less permanent hearing loss following exposure to intense noise . Thus , a point mutation designed to alter α9α10 receptor gating has provided an animal model in which not only is efferent inhibition more powerful , but also one in which sound-induced hearing loss can be restrained , indicating the ability of efferent feedback to ameliorate sound trauma .
In bringing information about the world to an individual , sensory systems perform a series of common functions . Each system responds with some specificity to a stimulus , and each one employs some specialized receptor cells at the periphery to translate specific stimuli into electrical signals that all neurons can use . That initial electrical event begins the process by which the central nervous system constructs an orderly representation of , for example , sounds , odors , tastes , and objects . Thus , basic sound detection begins when sound waves strike the eardrum , which transmits that physical stimulus to the organ of Corti within the cochlea , the sensory epithelium of the mammalian inner ear . Here , the primary receptor cells known as inner hair cells ( IHCs ) transform the information into electrical signals that are sent to the central nervous system by the auditory nerve [1] . However , unlike vision , touch , and the chemical senses , sound processing is modulated by efferent signals that travel in reverse , from the brain back to the inner ear [2] . One fundamental question in auditory neuroscience is what role ( s ) this feedback plays in our ability to hear . The medial olivocochlear ( MOC ) efferents ( Figure 1A ) originate in the medial portion of the superior olivary complex and project to outer hair cells ( OHCs; Figure 1B ) of the organ of Corti , where large synaptic contacts are formed [2] . In contrast to IHCs , OHCs act as biological motors to amplify the sound-evoked motions of the organ of Corti through a type of somatic electromotility generated by a specialized membrane protein known as prestin . Activation of the MOC pathway , either by sound or by shock trains delivered to the bundle at the floor of the IVth ventricle ( Figure 1A ) , reduces cochlear sensitivity through the action of the neurotransmitter acetylcholine ( ACh ) on nicotinic receptors ( nAChRs ) at the base of OHCs ( Figure 1C ) [3] . Although significant progress has been made in defining the cellular mechanisms of hair cell inhibition [3] , the functional role ( s ) of this sound-evoked feedback system , including control of the dynamic range of hearing [2] , improvement of signal detection in background noise [4–6] , mediating selective attention [7 , 8] , and protection from acoustic injury [9] , remain controversial . In addition to their significance in auditory processing , efferent cholinergic synapses provide a unique opportunity to assess the function of neuronal nAChRs . In contrast to nAChR-expressing cells in the central nervous system [10] , hair cells are compact and isopotential , and receive no other synaptic inputs . Moreover , compared to most other neuronal nAChRs , whose subunit composition remains challenging to elucidate [11] , the oligomeric structure of the hair cell nAChR has been defined: α9 and α10 subunits arrange into a pentameric assembly with a likely ( α9 ) 2 ( α10 ) 3 stoichiometry [12–15] . Current data support the notion that activation of the α9α10 nAChR leads to an increase in intracellular Ca2+ and the subsequent opening of small conductance Ca2+-activated K+ SK2 channels , thus leading to hyperpolarization of hair cells [3 , 16 , 17] , ( Figure 1C ) . In the present study , we have generated a genetically modified mouse model in which the magnitude and duration of the MOC efferent effect is increased , and assessed the consequences of this manipulation on auditory thresholds and susceptibility to noise-induced hearing loss . We have substituted a threonine for a leucine at position 9′ ( L9′T ) of the second transmembrane domain of the α9 subunit and examined the neurotransmitter responses , synaptic efficiency , and cochlear function of Chrna9wt/wt , Chrna9wt/L9′T , and Chrna9L9′T/L9′T mice . We show that the L9′T mutation produced an increase in sensitivity to ACh , and decreased rates of ACh-induced desensitization . More significantly , synaptic currents were dramatically prolonged in the OHCs of Chnra9L9′T/L9′T mice . Consistent with these effects , Chnra9L9′T/L9′T mice had elevated acoustic thresholds , and shock-evoked MOC activation produced both enhanced and prolonged cochlear suppression . This enhanced efferent inhibitory drive attenuated sound-induced , permanent acoustic injury . Our results establish efferent feedback inhibition's function in preventing acoustic trauma , and provide preliminary characterization of a mouse model to explore additional physiological functions of MOC innervation . In addition , it motivates further exploration of the efferent synapse as a model system to study fast neurotransmission mediated by a neuronal nAChR .
In Xenopus oocytes , α9α10 nAChRs assembled from mutant α9L9′T subunits exhibit an increased sensitivity to ACh , a reduced desensitization rate , and an increased channel mean open time [18] . To determine whether similar changes occurred in hair cells of L9′T mutant mice , we examined the response of neonatal IHCs to exogenous ACh . IHC nAChRs , which are expressed transiently before the onset of hearing , are of the α9α10 subtype and are functionally indistinguishable from those that arise later in OHCs [19 , 20] . Since recordings are more frequently and reliably obtained from neonatal IHCs than OHCs , the former provided the bulk of the data presented . As shown below , a subset of our findings has been repeated in OHCs of Chrna9L9′T mutant mice . As observed in Figure 3A , IHCs of Chrna9wt/L9′T and Chrna9L9′T/L9′T mice responded to ACh at P6–P10 . In each case , ACh dose-response curves of IHCs were left-shifted when compared to wild type , with a concomitant decrease in the half maximal effective concentration ( EC50 ) . The EC50 was significantly different ( p < 0 . 01 ) between homozygous mutant and wild-type IHCs ( Chrna9wt/wt: EC50 , 90 . 8 ± 18 . 2 μM , Hill coefficient 1 . 6 ± 0 . 1 , n = 4 animals , 7 cells; Chrna9wt/L9′T: EC50 , 57 . 2 ± 6 . 5 μM , Hill coefficient 2 . 2 ± 0 . 1 , n = 4 animals , 9 cells; and Chrna9 L9′T /L9′T: EC50 , 41 . 7 ± 5 . 7 μM , Hill coefficient 2 . 1 ± 0 . 2 , n = 4 animals , 7 cells ) . During prolonged ( >60 s ) application of 1 mM ACh at −90 mV ( Figure 3B ) , the current was entirely inward since the ACh response was recorded in isolation from the SK2 component ( see below ) . The evoked membrane current in wild-type IHCs decayed with a complex time course . In IHCs of both Chrna9wt/L9′T and Chrna9L9′T/L9′T mice , ACh-evoked membrane current was more sustained than that seen in wild-type mice ( i . e . , showed reduced desensitization , as previously shown with recombinant α9α10 nAChRs expressed in oocytes [18] ) . The fraction of current remaining after a 30-s ACh application was 0 . 29 ± 0 . 04 in Chrna9wt/wt ( n = 10 animals , 15 cells ) , 0 . 45 ± 0 . 05 in Chrna9wt/L9′T ( n = 9 animals , 12 cells , p < 0 . 05 ) , and 0 . 48 ± 0 . 06 in Chrna9 L9′T/L9′T ( n = 6 animals , 10 cells , p < 0 . 05 ) . Following efferent transmitter release in vitro , spontaneous inhibitory postsynaptic currents ( sIPSCs ) can be measured in neonatal IHCs . As determined for MOC synapses located on OHCs , the IHC inhibitory postsynaptic currents ( IPSCs ) result from ion flux through α9α10 nAChRs and calcium-activated SK2 channels [19 , 20] , ( Figure 1C ) . To only measure currents through α9α10 nAChRs , responses were recorded using pipettes filled with KCl-BAPTA ( thereby preventing calcium from activating SK2 channels ) and 5 nM apamin ( an SK channel blocker ) in the extracellular solution . As shown in Figure 4A and detailed in Table 1 , these “nAChR-only” sIPSCs of mutant mice were dramatically prolonged compared to wild type . There was a 6-fold increase in the τdecay , no difference in τrise , and a 3-fold increase in the duration at half amplitude ( halfwidth ) . In addition , the sIPSCs in Chrna9L9′T/L9′T IHCs had significantly smaller amplitudes than those of wild-type IHCs . Despite the amplitude reduction , overall charge transfer ( as indicated by the area value ) of the “nAChR-only” synaptic current was more than doubled in Chrna9L9′T/L9′T when compared to Chrna9wt/wt IHCs . When sIPSCs were recorded with KCl-EGTA in the pipette , currents were inward at −90 mV and outward at −40 mV ( Figure 4B ) . As noted for wild-type IHCs ( [19 , 20] and present results ) , the polarity shift of the sIPSC indicates that SK2 channel gating occurs in the Chrna9L9′T mutant mice . At −90 mV , sIPSCs of Chrna9L9′T/L9′T mice were dramatically prolonged compared to Chrna9wt/wt , with a 4- and 3-fold increase in the τdecay and halfwidth , respectively . At −40 mV , τdecay and halfwidth were 3-fold greater than in wild type , and a slight increase in τrise also was observed . The peak current at −40 mV , carried by SK2 channels , was equivalent in Chrna9wt/wt and Chrna9L9′T/L9′T IHCs . Nonetheless , the smaller amplitude nAChR-only current observed in IHCs of mutant mice activates SK2 channels as effectively as in wild-type IHCs . Indeed , the area under the Chrna9L9′T/L9′T SK2 current was nearly three times greater than that of the Chrna9wt/wt SK2 current . Presumably , this enhancement results from the increased total calcium ion flux during the prolonged activity of the nAChRs in mutant mice . As shown in Figure 4C , synaptic activity evoked by 40 mM K+ in IHCs was reversibly blocked by the α9α10 blocker strychnine ( 300 nM ) [13] , indicating that the synaptic activity observed in Chrna9L9′T/L9′T was in fact mediated through mutant α9α10 nAChRs . Since hair cell sIPSCs were infrequently observed , synaptic currents and the effects of exogenous ACh were measured in P10–P11 OHCs using high-potassium buffer ( 40 mM K+ plus EGTA , minus apamin ) to depolarize the efferent terminals ( Figure 5A and 5B ) . Under these recording conditions , the synaptic waveform includes current through both the nAChR and associated SK2 channels ( entirely inward at −90 mV since 40 mM K+ shifts the equilibrium potential of this cation to −33 mV ) . Analysis of waveforms revealed a 4- and 3 . 5-fold increase in τdecay and halfwidth , respectively , and a 0 . 6-fold reduction in amplitude ( Figure 5C and Table 2 ) in Chrna9L9′T/L9′T IPSCs compared to Chrna9wt/wt , much as was observed in IHCs . Despite a drop in amplitude , the overall charge transfer ( as indicated by the area value ) was 2-fold larger in Chrna9L9′T/L9′T compared to Chrna9wt/wt OHCs . Under these recording conditions , synaptic currents reflect ion flux through both nAChRs and associated SK2 channels . The presence of an ACh-evoked SK2 current was demonstrated directly by application of ACh to Chrna9L9′T/L9′T OHCs held at different membrane potentials . As shown in Figure 5D , ACh responses were outward at −40 mV , confirming the activation of potassium currents in mutant OHCs . We examined cochlear responses in Chrna9L9′T mutant mice via auditory brainstem responses ( ABRs ) and distortion product otoacoustic emissions ( DPOAEs ) . Wave 1 of the ABR , which is recorded from scalp electrodes in response to short tone pips , represents the summed activity of cochlear nerve fibers projecting from IHCs ( Figure 1B ) to the cochlear nucleus , the first central nucleus in the ascending auditory pathway [21] . The DPOAEs are sounds created within the cochlea , amplified by the action of OHCs and propagated through the middle ear back to the ear canal , where they can be measured with a microphone [22] . Given that the generation of DPOAEs requires neither IHCs nor cochlear nerve fibers [23] , a comparison of ABRs and DPOAEs can provide insight into the locus of any dysfunction . As shown in Figure 6A , mean ABR thresholds were elevated by 5–15 dB in both Chrna9wt/L9′T ( F ( 1 , 55 ) = 4 . 30 , p = 0 . 043 ) and Chrna9L9′T/L9′T ( F ( 1 , 79 ) = 7 . 87 , p = 0 . 06 ) when compared to Chrna9wt/wt mice . DPOAE thresholds ( Figure 6B ) were also elevated by 5–15 dB in both Chrna9wt/L9′T ( F ( 1 , 53 ) = 14 . 82 , p < 0 . 001 ) and Chrna9L9′T/L9′T ( F ( 1 , 77 ) = 9 . 20 , p = 0 . 003 ) . Threshold shifts were larger at frequencies above 16 kHz and were similar in magnitude whether measured by ABRs or DPOAEs , suggesting that all dysfunction can be explained by changes in OHC contributions to cochlear amplification . Since MOC activity in vivo decreases OHC-based amplification of cochlear responses [2] , we asked whether the increased cochlear thresholds might arise from enhanced synaptic currents through mutant α9-containing channels in response to the normal low level of spontaneous MOC activity [24] . To test this hypothesis , we injected mice with strychnine ( 30 mg/kg , intraperitoneally [i . p . ] ) , a potent blocker of α9α10 channels [13 , 25] . Indeed , as shown in Figure 7 , strychnine improved acoustic thresholds in Chrna9L9′T/L9′T mice by almost 9 dB , thus restoring the baseline thresholds to that seen in the wild type . Activation of MOC efferents normally decreases the OHC contribution to cochlear amplification . Thus , the electrical stimulation of the MOC fibers at the floor of the IVth ventricle ( Figure 1A , black arrow ) results in a decrease in the amplitude of the DPOAEs [25] . To assess MOC function in vivo , DPOAEs were measured before , during , and after a 70-s train of shocks to the olivocochlear bundle . In wild-type mice , DPOAE suppression is seen immediately after shock-train onset ( Figure 8A , black arrowhead ) and then adapts to a steady state , which is maintained throughout the shock epoch ( Figure 8A and 8B ) . In Chrna9L9′T mutant mice , suppression had a much slower onset ( Figure 8A , white arrowhead ) , but continued to grow during the shock train ( Figure 8A and 8B ) until responses disappeared into the noise floor . To determine the full suppression magnitude , we raised the level of the acoustic stimulus ( Figure 8C ) : peak suppression in Chrna9L9′T/L9′T reached approximately 17 dB , whereas for Chrna9wt/wt , the maximum suppression was less than approximately 5 dB . In homozygous mutants , suppression persisted for almost 10 min after the end of the shock train , with a slow recovery to baseline ( Figure 8B ) . Moreover , there is a prominent “overshoot , ” which , except for a slower time course , is reminiscent of the post-shocks enhancement seen in wild-type mice [25] . As shown in Figure 8A and 8B , responses in Chrna9wt/L9′T were intermediate between those of homozygous mutant and wild-type mice . As with wild-type mice [25] , suppression could be blocked by strychnine ( 10 mg/kg ) , leaving only the postshock enhancement ( Figure 9A , grey arrowhead ) , suggesting that the increased suppression in Chrna9L9′T mutants was due to the enhanced activity of α9α10 receptors . With a lower dose of strychnine ( 3 mg/kg ) , suppression in Chrna9L9′T/L9′T resembled the normal Chrna9wt/wt response ( i . e . , reduced suppression magnitude but with rapid onset and decay; Figure 9B , grey arrowhead ) . Given that MOC activation most likely results in a decrease in acoustic injury [26–28] , we investigated the susceptibility of Chrna9L9′T mutant mice to intense noise . Since the mechanisms underlying reversible and irreversible noise-induced threshold shifts are different [29 , 30] , we used two different noise exposures to induce either a temporary ( TTS ) or permanent ( PTS ) threshold shift . Homozygous mutant mice showed increased resistance to permanent acoustic injury when compared to wild type ( Figure 10A , 5 . 0–45 . 2 kHz , Chrna9L9′T/L9′T: F ( 1 , 8 ) = 8 . 894 , p = 0 . 018; Chrna9wt/L9′T: F ( 1 , 7 ) = 6 . 263 , p = 0 . 041 ) . With a lower intensity exposure designed to produce a TTS , there was no significant difference in vulnerability between wild-type and mutant mice , when examined 12 h after exposure ( Figure 10B ) .
Understanding neurotransmission mediated by neuronal nicotinic receptors is a fundamental challenge in neuroscience , since decline , disruption , or alteration of nicotinic cholinergic mechanisms in the central nervous system contributes to dysfunctions such as epilepsy , schizophrenia , Parkinson's disease , autism , dementia with Lewy bodies , Alzheimer's disease , and addiction [10] . Although fast , direct nicotinic synaptic activity drives neurotransmission in autonomic ganglia , only rare cases of fast nicotinic transmission mediated by neuronal nAChRs have been reported in the mammalian brain . Because cholinergic neurons in the brain are usually loosely distributed and often sparsely innervate broad areas , it is experimentally difficult to stimulate a large number of those neurons and to record selectively from their postsynaptic targets . Indeed , it is likely that authentic fast nicotinic transmission is present at low densities in more neuronal areas than the few that have been reported [10] . Furthermore , the variable transcription of genes coding for nAChR subunits and the possible combinatorial assembly of these subunits produces a wide structural diversity of receptor types . This imposes additional challenges for studying native synaptic neuronal nAChRs [31] . Thus , the cholinergic synapse between MOC terminals and cochlear hair cells provides a valuable model for the study of fast neurotransmission mediated by nAChRs of known composition [12 , 13] . Targeted mutations can be introduced and the phenotypic consequences can be analyzed at the synaptic , whole-organ , and systems level . When analyzed at the level of nAChR function , the α9 L9′T mutant mice reproduce what has been previously described for the recombinant receptor expressed in oocytes , i . e . , a decrease in EC50 for ACh and a reduced desensitization kinetics [18] . Both these effects probably derive from the fact that the L9′ position of nAChRs is critical for channel gating [32] and that hydrophilic substitutions at this position lead to increased mean open times [18 , 33] . Moreover , these changes in channel gating properties translate into increased synaptic efficacy , as seen from the prolonged synaptic currents observed in IHCs and OHCs of mutant mice . In addition , the inhibitory signature of the efferent synapse is conserved , since the nAChR currents remain coupled to the associated SK2 currents , which were also substantially prolonged . In fact , the increase in overall charge transfer during sIPSCs between wild-type and knockin mice was more pronounced when the secondary SK2 currents were measured , than for the nAChR-only currents ( Table 1 ) , pointing to an additional amplified step , possibly calcium-induced calcium release from the nearby synaptic cistern [34] . Finally , the finding that the L9′T mutation does not lead to hair cell death , is distinct from effects of similar mutations in α4 and α7 nAChR subunits , which lead to death of substantia nigra dopaminergic neurons [35] and apoptotic cell death throughout the somatosensory cortex [36] , respectively , most likely due to Ca2+ excitotoxicity . The observation that α9α10 receptors are indeed highly permeable to Ca2+ [37 , 38] points toward an efficient Ca2+ buffering system in hair cells . In fact , proteinaceous calcium buffers ( e . g . , parvalbumin-β ) are expressed in OHCs at high levels , similar to those found in skeletal muscle [39] . The main peripheral effect of the MOC activity is to inhibit cochlear responses by decreasing the gain of the cochlear amplifier [2] . MOC neurons comprise the effector arm of a sound-evoked negative feedback system that , in a quiet environment , normally has little effect on cochlear sensitivity , since MOC neurons have low levels of spontaneous activity and do not respond to sound until levels exceed threshold by 15–20 dB [24] . The fact that the baseline threshold elevation in L9′T mutants can be “rescued” via strychnine , the potent α9α10 nAChR blocker , suggests that it arises from an enhancement of cholinergic effects on OHCs . Therefore , compared to Chrna9 and Chrna10 knockouts , in which baseline cochlear thresholds were normal [40 , 41] , the present knockin strategy reveals cholinergic MOC effects under resting conditions . This baseline inhibition could arise from an increased sensitivity of mutant receptors to normal low levels of spontaneous ACh release from MOC terminals or to the increased probability of spontaneous channel openings of L9′T mutant receptors in the absence of ACh [18] . The appearance of spontaneous channel openings , as described for the mutant receptors [18] , may also explain how low-dose strychnine can speed the onset kinetics of the in vivo response in mutant mice ( Figure 9B ) , i . e . , by restoring mutant channels to the closed state , from which rapid ACh-mediated channel openings can occur . Hair cell recordings from the L9′T mutant mice in the present study show ( 1 ) ACh-evoked currents with a greatly reduced desensitization rate , and ( 2 ) spontaneous miniature synaptic currents with slower activation and decay kinetics . These changes in ACh response kinetics provide likely explanations for some aspects of the electrically evoked suppression of DPOAEs . The doubling of onset time for the ACh-evoked SK2 current in mutant hair cells ( Table 1 ) must contribute to the slowed onset of MOC-mediated suppression in vivo , by reducing linear summation of postsynaptic effects . However , the slowing of suppression onset is so dramatic ( <1 s in wild type vs . >20 s in mutant mice ) that other possible factors , e . g . , alterations of release probability or slowed facilitation in the efferent terminals [42] , might also contribute to this delayed time course . In addition to rising more slowly , MOC-mediated cochlear suppression was also larger and longer lasting in L9′T mutant mice . The normal adaptation of suppression during continuous MOC activation was absent in the presence of slowly desensitizing α9L9′T mutant receptors , consistent with the idea that receptor desensitization is a key factor in the decay of the wild-type response after shock-train onset . However , the dramatic prolongation of cochlear suppression after shock-train offset ( 500 s in mutants vs . <5 s in wild types ) is too large to be explained only by the 2- to 3-fold prolongation of synaptic currents observed in the mutant hair cells . Another contributing factor may be the observation that choline , the metabolite produced when acetylcholinesterase degrades ACh released at the synapse , is a full agonist of the mutant α9α10 receptor , and only a weak partial agonist of the wild-type receptor [18] . Thus , in vivo , the suppression can only decay after choline is taken up by the synaptic terminal or otherwise diffuses away . Treatment with a low dose of the α9α10 antagonist strychnine ( 3 mg/kg ) supports this interpretation . This treatment completely abolished the prolonged postshock suppression in mutant mice ( and unmasked a fast onset ) . Since choline has 4-fold lower affinity for α9L9′T mutant receptors than ACh [18] , a low concentration of strychnine may be differentiating the effects of the two agonists ( i . e . , blocking the effects of the agonist with a lower potency , choline , but leaving unchanged the effects of the high-affinity agonist , ACh ) , as well as modulating the resting level of desensitization . Beyond the possibility of metabolite activation , extended inhibitory effects could involve more than changes in membrane conductance . For example , it has been shown that exposure to ACh alters the stiffness and motility of OHCs isolated from the gerbil cochlea over tens of minutes [43] . Stiffness and motility of OHCs depends , at least in part , on the motor protein prestin , a molecule that has selectively evolved in mammals to subserve somatic electromotility and amplification [44 , 45] . Thus , it remains to be determined whether prolonged Ca2+ influx through mutated α9α10 receptors leads to changes in prestin structure or function resulting in prolonged MOC efferent effects . The presence of mutant α9L9′T receptors , designed to increase the magnitude of MOC effects on OHCs , also increased the protective action of the MOC system in vivo . This is consistent with previous work showing that overexpression of wild-type α9 channels , which more modestly increased the magnitude of MOC-mediated DPOAE suppression , also increased the resistance of the ear to acoustic injury [26] . Prior in vivo studies of electrically evoked MOC activity have described both fast and slow effects of ACh on cochlear neural responses [46 , 47] and cochlear mechanical vibrations [48] . The fast suppression , with an onset time course of approximately 100 ms , arises from the increased K+ conductance in neighboring SK2 channels and the effects of the resultant OHC hyperpolarization on the magnitude of electromotility and thus on cochlear vibration amplitude . The slow suppression , with an onset time course of approximately 10 s , may require a wave of calcium-induced calcium release along the OHC membrane and appears to reflect a change in OHC stiffness [48] that also reduces cochlear vibrations . The observation that overexpression of SK2 channels does not increase resistance to acoustic injury [49] , although as with α9 overexpression , it also increases the magnitude of MOC-mediated DPOAE suppression , suggests that it is the slow effects of ACh that are responsible for its protective action , rather than the hyperpolarization-induced decrease in electromotility . There are two fundamentally different ways in which slow effects of α9 activation could reduce acoustic injury: ( 1 ) by reducing mechanical vibration of the sensory epithelium , or ( 2 ) via intracellular modifications of OHCs arising as other downstream effects of the calcium entry through α9α10 receptors . The further observation that in the α9L9′T mutants , protection was seen only for exposures intense enough to produce irreversible damage and permanent threshold shifts , and not in less traumatic exposures producing only reversible changes , argues against a general reduction of vibration amplitude as the underlying mechanism . Most of the permanent threshold elevation underlying this type of noise exposure arises from damage to the hair cell stereocilia bundles , including disarray loss and/or fusion of these modified microvilli [50] , which house the mechanoelectric transduction channels . Further insight into the mechanisms linking calcium entry through nAChRs and protection from acoustic injury will require a clearer delineation of the molecular events leading up to this type of noise-induced stereocilia damage . This work shows that a point mutation in the hair cell's nAChR produces dramatic prolongation of efferent MOC inhibitory effects at both cellular and systems levels . This alteration provides enhanced protection from permanent acoustic trauma , indicating that cholinergic synaptic feedback is not only necessary , but sufficient for this effect . In addition , the enhanced activity of the modified nAChR revealed a novel tonic inhibitory effect , raising baseline acoustic thresholds over those of wild-type littermates , confirming the inhibitory nature of the MOC efferent system . Thus , this α9L9′T knockin presents new insights into the cellular mechanisms of cholinergic inhibition , as well as a promising model in which to probe the functional role of MOC efferents . In addition , the hair cell's efferent synapse , much like the neuromuscular junction , can provide insights into cholinergic signaling , and promises to be an equally informative model for studying activity-dependent synaptic function .
A HindIII-NotI ( the latter derived from the construction of the library ) restriction endonuclease fragment of approximately 9 , 500 bp encoding Chrna9 exons 1–4 ( see GenBank accession number NT_039305 . 7 and Figure 2A ) was obtained from a mouse strain 129S4/SvJae genomic DNA library ( kindly provided by Dr Bernhard Bettler , University of Basel ) and subcloned into the vector pKO-Select DT ( Lexicon Genetics ) . A 2-kbp neomycin resistance cassette flanked by two loxP sites ( loxP-neo-loxP ) was inserted in the NcoI restriction site within the intron located between exons 3 and 4 and the Chrna9L9′T mutation ( * ) introduced via site-directed mutagenesis using the QuickChange Site-Directed Mutagenesis kit ( Stratagene ) and amplimers A9sense ( 5′-CTCTGGGAGTGACCATCCTAacGGCCATGACTGTATTTCAGC-3′ ) and A9antisense ( 5′-GCTGAAATACAGTCATGGCCgtTAGGATGGTCACTCCCAGAG-3′ ) . This targeting vector was used to electroporate 129S4/SvJae embryonic stem ( ES ) cells , and homologous recombinants were obtained following gentamicin ( G418 ) selection and Southern blot hybridization analyses . Genomic DNA was purified from G418-resistant ES cell clones , digested with HindIII and KpnI , electrophoresed on 0 . 8% agarose , and hybridized to a 32P-labled DNA 962-bp SacII-KpnI fragment probe prepared from the DNA fragment shown in Figure 2A . Based on the sequence of the mouse Chrna9 subunit gene ( see GenBank accession number NT_039305 . 7 ) , wild-type ES cell DNA yielded a fragment of 13 , 800 bp , whereas ES cells that have undergone homologous recombination yielded a 7 , 300-bp fragment ( see Figure 2B ) . Transfection of the linearized targeting vector into murine ES cells resulted in the insertion of the L9′T mutation into the α9 nAChR subunit . The frequency of homologous recombination events in ES cells was 42% . Six independent ES cell lines carrying the mutation were injected into blastocysts to generate germline chimeric males and were then implanted into pseudopregnant females . Chimeric male progeny mice were backcrossed to strain C57BL/6J females , and agouti coat color was used to assess the germline status of the targeted allele . Founder males were backcrossed to wild-type 129S4/SvJae females and heterozygous N1 females mated to cre-expressing transgenic males ( FVB/N-Tg ( EIIa-cre ) C5379Lmgd/J , stock 003314; Jackson Laboratory ) to remove the “floxed” neomycin resistance cassette . The excision of the neo cassette and subsequent segregation and loss of the cre transgene were monitored via PCR . The single loxP site footprint and flanking regions of the targeted allele in neo-deleted mice were confirmed by DNA sequencing . The Chrna9L9′T mutant allele has been maintained in congenic FVB . 129P2-Pde6b+ Tyrc-ch/AntJ ( stock number 004828; Jackson Laboratory ) strain . C57BL/6J mice develop a marked and progressive late-onset hearing loss characterized by cochlear degeneration . The FVB background lacks this hearing loss . Moreover , the wild-type Pde6b allele avoids blindness due to retinal degeneration typical of the FBV strain . All experiments reported in this paper were performed using neo-deleted Chrna9wt/wt , Chrna9wt/L9′T , or Chrna9L9′T/L9′T mutant mice backcrossed with congenic FVB . 129P2-Pde6b+ Tyrc-ch/AntJ stock for four to five generations ( i . e . , N4–N5 ) . Routine genotyping of Chrna9 mice was performed using tail biopsy tissue DNA samples ( Wizard Genomic DNA Purification kit; Promega ) , amplimers A9LOXP . 1 ( 5′-TAC TGG CTA TCC TCC AGA CAG AGC-3′ ) and A9LOXP . 2 ( 5′-AGG AGC GAG CAG AGG TCC TAA ATT-3′ ) ( see Figure 2A ) , and Failsafe PCR System Kit with buffer D ( Epicentre ) as described by the manufacturer . PCR cycle parameters were: 95 °C , 0 . 5 min; 55 °C , 1 . 0 min; and 72 °C , 2 min for a total of 35 cycles . Reaction products were electrophoresed on 1 . 5% agarose , stained with ethidium bromide , and photographed . The PCR fragment length for the wild-type Chrna9 allele is 203 bp and 269 bp for the mutant allele ( see Figure 2C ) . In each breeding pair , the mutant status of the Chrna9L9′′T was assessed by sequencing of the PCR fragment obtained with amplimers L9T5′ ( 5′-CTCTCTGACTTCATTGAAGACG-3′ ) and L9T3′ ( 5′-CCGCACACATACAGGGTTCGAT-3′ ) ( Figure 2D ) . Mice were sacrificed by decapitation . All experimental protocols were carried out in accordance with the American Veterinary Medical Associations' AVMA Guidelines on Euthanasia ( June 2007 ) . Apical turns of the organ of Corti were excised from mice at postnatal ages P6–P11 for IHCs and P10–P11 for OHCs , and used within 3 h . Day of birth was considered postnatal day 0 ( P0 ) . Cochlear preparations were mounted under a Leica DMLFS microscope ( Leica Microsystems ) and viewed with differential interference contrast ( DIC ) using a 40× water immersion objective and a Hamamatsu C7500–50 camera with contrast enhancement ( Hamamatsu ) . Methods to record from IHCs and OHCs were essentially as described previously [17 , 19] . Briefly , hair cells were identified visually with the 40× objective and during recordings , by the size of their capacitance ( 7 to 12 pF ) , by their characteristic voltage-dependent Na+ and K+ currents [51] . Some cells were removed to access IHCs , but mostly , the pipette moved through the tissue under positive pressure . The extracellular solution was as follows ( in mM ) : 155 NaCl , 5 . 8 KCl , 1 . 3 CaCl2 , 0 . 9 MgCl2 , 0 . 7 NaH2PO4 , 5 . 6 d-glucose , and 10 Hepes buffer ( pH 7 . 4 ) . Two different pipette solutions were used . One contained ( in mM ) : 150 KCl , 3 . 5 MgCl2 , 0 . 1 CaCl2 , 5 ethyleneglycol-bis ( β-aminoethyl ether ) -N , N , N′ , N′- teraacetic acid ( EGTA ) , 5 Hepes buffer , 2 . 5 Na2ATP , pH 7 . 2 ( KCl-EGTA saline ) . In the other , EGTA was replaced with 10 mM bis ( 2-aminophenoxy ) ethane-N , N , N′ , N′-tetra-acetic acid ( BAPTA ) ( KCl-BAPTA saline ) . The latter solution was used where indicated , in order to record the current through the ACh receptor in isolation from the coupled SK2 current ( nAChR-only ) . In addition , in the latter condition , the SK blocker apamine ( 5 nM ) was added to the recording solution . Solutions containing ACh or high K+ were applied by a gravity-fed multichannel glass pipette ( ∼150 μm tip diameter ) , positioned about 300 μm from the recorded cell . All working solutions containing either ACh or elevated K+ or both , were made up in a saline containing low Ca2+ ( 0 . 5 mM ) and no Mg2+ so as to optimize the experimental conditions for measuring currents flowing through the α9α10 receptors [38] . sIPSCs were recorded immediately after rupturing into the cell , in the extracellular saline containing 1 . 3 mM Ca2+ and no Mg2+ . All experiments designed to record synaptic currents , either spontaneous or evoked with high K+ , were done using an extracellular solution containing 1 . 3 mM Ca2+ and no Mg2+ . sIPSCs were recorded immediately after rupturing the cell . Glass pipettes , 1 . 2-mm I . D . , had resistances of 5–8 MΩ . Currents in both IHCs and OHCs were recorded in the whole-cell patch-clamp mode with an Axopatch 200B amplifier , low-pass filtered at 2–10 kHz , and digitized at 5–20 kHz with a Digidata 1322A board ( Molecular Devices ) . Recordings were made at room temperature ( 22–25 °C ) . Holding potentials were not corrected for liquid junction potentials or for the voltage drop across the uncompensated series resistance . IPSCs were analyzed with Minianalyis ( Synaptosoft; Jaejin Software ) . IPSCs were identified using a search routine for event detection and confirmed by eye . The individual events of each cell were aligned and averaged using Minianalysis . Rise time was used as the criterion to align . For further analysis , Prism 4 ( GraphPad Software ) was used . The τrise and τdecay were fit to a monoexponential . Prolonged agonist application responses and concentration-response curves to ACh in IHCs were performed in the nAChR-only condition , normalized to the maximal agonist response , and iteratively fitted with the equation: I/Imax = An/ ( An + EC50n ) , where I is the peak inward current evoked by agonist at concentration A; Imax is current evoked by the concentration of agonist eliciting a maximal response; EC50 is the concentration of agonist inducing half-maximal current response , and n is the Hill coefficient . Mice were anesthetized with xylazine ( 20 mg/kg i . p . ) and ketamine ( 100 mg/kg i . p . ) . DPOAEs at 2f1–f2 were recorded with a custom acoustic assembly consisting of two electrostatic drivers ( TDT EC-1; Tucker-Davis Technologies ) to generate primary tones ( f1 and f2 with f2/f1 = 1 . 2 and f2 level 10 dB < f1 level ) and a Knowles miniature microphone ( EK3103 ) to record ear-canal sound pressure . Stimuli were generated digitally , while resultant ear-canal sound pressure was amplified and digitally sampled at 4 μs ( 16-bit DAQ boards , NI 6052E; National Instruments ) . Fast Fourier Transforms were computed and averaged over five consecutive waveform traces , and 2f1–f2 DPOAE amplitude and surrounding noise floor were extracted . Iso-response curves were interpolated from plots of amplitude versus sound level , performed in 5-dB steps of f1 level . “Threshold” is defined as the f1 level required to produce a DPOAE with amplitude = 0 dB sound pressure level ( SPL ) . For measurement of ABRs , needle electrodes were inserted at vertex and pinna , with a ground near the tail . ABRs were evoked with 5-ms tone pips ( 0 . 5-ms rise-fall , cos2 onset , at 35/s ) . The response was amplified ( 10 , 000× ) , filtered ( 100 Hz–3 kHz ) , and averaged with an A-D board in a LabVIEW-driven data-acquisition system . Sound level was raised in 5-dB steps from 10 dB below threshold to 80 dB SPL . At each level , 1 , 024 responses were averaged ( with stimulus polarity alternated ) , using an “artifact reject” whereby response waveforms were discarded when peak-to-peak amplitude exceeded 15 μV . Upon visual inspection of stacked waveforms , “threshold” was defined as the lowest SPL level at which any wave could be detected , usually the level step just below that at which the response amplitude exceeded the noise floor ( ∼0 . 25 μV ) . For amplitude versus level functions , wave-I peak was identified by visual inspection at each sound level and the peak-to-peak amplitude computed . Mice were anesthetized with urethane ( 1 . 20 g/kg i . p . ) and xylazine ( 20 mg/kg i . p . ) . A posterior craniotomy and partial cerebellar aspiration were performed to expose the floor of the IVth ventricle . To stimulate the MOC bundle , shocks ( monophasic pulses , 150-μs duration , 200/s ) were applied through fine silver wires ( 0 . 4-mm spacing ) placed along the midline , spanning the olivocochlear decussation . Shock threshold for facial twitches was determined , muscle paralysis induced with α-d-tubocurarine ( 1 . 25 mg/kg i . p . ) , and the animal connected to a respirator via a tracheal cannula . Shock levels were raised to 6 dB above twitch threshold . During the MOC suppression assay , f2 level was set to produce a DPOAE 10–15 dB or 20–25 dB greater than the noise floor . To measure MOC effects , repeated measures of baseline DPOAE amplitude were first obtained ( n = 56 ) , followed by a series of 70 contiguous periods in which DPOAE amplitudes were measured with simultaneous shocks to the MOC bundle and additional periods during which DPOAE measures continued after the termination of the shock train . Animals were exposed free-field , in a small reverberant chamber . Acoustic trauma consisted of a 2-h exposure to an 8–16-kHz octave band noise presented at 100 dB SPL ( for permanent injury ) or a 15-min exposure to the same noise at 94 dB SPL ( for temporary injury ) . For the higher level exposure , animals were anesthetized ( ketamine and xylazine , exactly as for the ABR and DPOAE testing ) , because many mutant animals experienced audiogenic seizures as soon as the high-level noise was turned on . The exposure stimulus was generated by a custom white-noise source , filtered ( Brickwall Filter with a 60-dB/octave slope ) , amplified ( Crown power amplifier ) , and delivered ( JBL compression driver ) through an exponential horn fitted securely to a hole in the top of a reverberant box . Sound exposure levels were measured at four positions within each cage using a 0 . 25'' Bruel and Kjaer condenser microphone: sound pressure was found to vary by less than 0 . 5 dB across these measurement positions . Statistical analyses of in vitro electrophysiology experiments were carried by the Student t-test in the case of pairwise comparisons or a one-way ANOVA followed by a Dunnet test for multiple comparisons . In the case of in vivo data , a two-way ANOVA was performed . A p < 0 . 05 was selected as the criterion for statistical significance . Mean values are quoted as means ± the standard error of the mean ( S . E . M . ) . ACh chloride , strychnine HCl , Na2ATP , BAPTA , and all other reagents were from Sigma Chemical . EGTA and Na2ATP were dissolved at the moment of preparing the intracellular solutions . All experimental protocols were carried out in accordance with the National Institutes of Health guide for the care and use of laboratory animals as well as Instituto de Investigaciones en Ingeniería Genética y Biología Molecular ( INGEBI ) , Tufts University , and Massachusetts Eye and Ear Infirmary Institutional Animal Care and Use Committee ( IACUC ) guidelines , and best practice procedures . | Nicotinic cholinergic receptors are essential to higher order brain function . Structurally , these operate through a myriad of ligand-gated pentameric arrangements of different homologous subunits . Here , we report progress in understanding the structural properties of a neuronal nicotinic receptor at the synapse . Receptors assembled from two nicotinic cholinergic subunits ( α9 and α10 ) serve exclusively at the synapse between central nervous system descending fibers and cochlear hair cells . This enabled us to show direct causality between a point mutation of the α9 subunit , and predicted alterations in the synaptic strength in sensory hair cells of the cochlea of α9 point mutant mice . Furthermore , this single mutation results in profound enhancement of central nervous system feedback to the cochlea . And finally , as a consequence , mutant mice possessing this altered receptor have substantially improved resistance to traumatic sound . Thus , central neuronal feedback on cochlear hair cells provides an opportunity to define one specific role that nicotinic receptors can play in the nervous system , enabling study from biophysical to behavioral levels and promoting a target for the prevention of noise-induced hearing loss . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"neuroscience",
"physiology"
]
| 2009 | A Point Mutation in the Hair Cell Nicotinic Cholinergic Receptor Prolongs Cochlear Inhibition and Enhances Noise Protection |
The cellular response elicited by an environmental cue typically varies with the strength of the stimulus . For example , in the yeast Saccharomyces cerevisiae , the concentration of mating pheromone determines whether cells undergo vegetative growth , chemotropic growth , or mating . This implies that the signaling pathways responsible for detecting the stimulus and initiating a response must transmit quantitative information about the intensity of the signal . Our previous experimental results suggest that yeast encode pheromone concentration as the duration of the transmitted signal . Here we use mathematical modeling to analyze possible biochemical mechanisms for performing this “dose-to-duration” conversion . We demonstrate that modulation of signal duration increases the range of stimulus concentrations for which dose-dependent responses are possible; this increased dynamic range produces the counterintuitive result of “signaling beyond saturation” in which dose-dependent responses are still possible after apparent saturation of the receptors . We propose a mechanism for dose-to-duration encoding in the yeast pheromone pathway that is consistent with current experimental observations . Most previous investigations of information processing by signaling pathways have focused on amplitude encoding without considering temporal aspects of signal transduction . Here we demonstrate that dose-to-duration encoding provides cells with an alternative mechanism for processing and transmitting quantitative information about their surrounding environment . The ability of signaling pathways to convert stimulus strength into signal duration results directly from the nonlinear nature of these systems and emphasizes the importance of considering the dynamic properties of signaling pathways when characterizing their behavior . Understanding how signaling pathways encode and transmit quantitative information about the external environment will not only deepen our understanding of these systems but also provide insight into how to reestablish proper function of pathways that have become dysregulated by disease .
Many substances , such as hormones , neurotransmitters and a variety of pharmaceuticals , affect cellular behavior by binding to membrane receptors and activating intracellular signaling pathways . These pathways transmit information from the plasma membrane to selected cellular components to generate an appropriate response to the environmental cue . However , signaling networks are not simply passive relay systems , but actively modulate the transmitted signals . For example , cross inhibition is used to avoid spurious crosstalk between pathways . Similarly , negative feedback allows pathways to adapt or desensitize to persistent stimuli [1] , [2] . In many cases , the nature of the response depends on the dose of the stimulus . Thus , in addition to relaying qualitative information ( e . g . the presence or absence of a stimulus ) , signaling pathways must also transmit quantitative information about the intensity of the stimulus . Many signaling pathways consist of a cell surface receptor , G protein transducer , and a series of protein kinases , including a mitogen activated protein kinase ( MAPK ) . This architecture is widely employed in mammalian cells , but is also found in single-cell eukaryotes such as yeast [3] . The pheromone response pathway of the yeast Saccharomyces cerevisiae provides an instructive example in which the elicited cellular response depends on the concentration of the stimulus . At low pheromone levels , cells continue vegetative growth . At intermediate concentrations , cells develop an elongated morphology and in the presence of a pheromone gradient the growth is directed to the source of the stimulus , a process known as chemotropic growth [4]–[7] . Finally , at high pheromone concentrations cells initiate a mating program that eventually leads to growth arrest and the development of mating projections ( for a review see [3] ) . Therefore , for yeast to make the correct developmental decision , quantitative information about the pheromone concentration must be reliably transmitted to the appropriate cellular components . Here we use mathematical modeling to investigate the different ways this information can be transferred . The results of this analysis taken together with our recently published data demonstrate that the pheromone pathway uses a strategy in which the agonist dose is encoded as the duration of the signal . Because the yeast pheromone response pathway consists of a G-protein coupled receptor and MAP kinase cascade , the results of our investigations should have direct implications for signal transduction in mammalian cells .
In the simplest scenario of a linear signaling pathway subject to a sustained stimulus ( Figure 1A ) , quantitative information about the dose of the stimulus can only be transmitted as the activity level of the signaling proteins that make up the pathway . We refer to this mode of signal transduction as “amplitude encoding” because information about the stimulus is contained in the amplitude of the propagated signal [8] . For linear pathways , the dynamic range , ( i . e . , the range of stimulus levels to which the pathway can respond in a dose-dependent manner ) is limited when the activity of a pathway component becomes saturated . Often , a downstream component saturates before the receptor . Thus the agonist concentration required to achieve the maximum downstream response may be less than the concentration needed to saturate the receptors , causing the dose-response curve of the pathway to shift to the left of the receptor-occupancy curve ( Figure 1B ) . Pharmacologists refer to this phenomenon as “amplification” or “receptor reserve” . Signaling networks are rarely linear , however , and often include combinations of feedback and feed-forward loops that positively and negatively regulate pathway activity . Among other things , these regulatory mechanisms allow signaling systems to respond transiently ( adapt ) to a persistent stimulus . We show that transient pathway activation provides the possibility of “dose-to-duration” encoding . That is , information about the stimulus concentration is transduced as the duration of the propagated signal rather than the amplitude . We demonstrate that an advantage of dose-to-duration encoding is that it provides a mechanism for increasing the dynamic range of signaling systems by allowing them to respond in a dose-dependent manner even after pathway components have become saturated , even at agonist concentrations that saturate the receptor ( Figure 1C ) . In the following sections , we analyze dose-to-duration encoding as a means for relaying quantitative information about the extracellular environment and discuss simple pathway architectures capable of carrying out this conversion . The key information transfer in this strategy occurs during the transient activation of pathway components rather than through their steady state levels of activation [9] . Next , we present data for MAP kinase activity that demonstrates dose-to-duration encoding is used in the pheromone response pathway of yeast . Finally , we present a mathematical model of the pathway based on the mechanisms discussed here that is consistent with experimental observations . Dose-to-duration encoding requires the propagated signal to act transiently . That is , at least one component of the pathway must return to its pre-stimulus level on a time-scale significantly shorter than that of the physiological response . This transient activity can result from the stimulus itself acting transiently or arise because the pathway contains regulatory elements that convert a sustained input into a transient output . The first case is commonly observed in inter-cellular signaling , where the duration of pathway activity often is regulated by the slow degradation or elimination of the agonist ( e . g . , reuptake of a neurotransmitter ) [10] , [11] . We focus on the second case , that is , adaptive systems that possess the ability to convert the intensity of the input signal into duration of the output signal in the presence of a persistent stimulus . Figure 2 shows schematically how dose-to-duration encoding works . In this example , a fixed agonist concentration quickly activates receptors in the plasma membrane . The steady-state level of active receptor ( input ) causes the activation of a signaling module ( “encoder” , grey box ) that generates a transient activation of the signaling protein A . We use an asterisk to denote the active form of a protein ( e . g . , A* ) . The output of the encoding module is a signal of constant amplitude but dose-dependent duration ( A* in Figure 2 ) . At each stimulus dose , the amplitude of A* rapidly saturates , but information about the level of receptor occupancy is preserved in the duration of the A* signal . Mechanisms capable of such dose-duration transformations are the subject of the next section . Figure 2 shows two possible scenarios for how A* activates its downstream targets . In the first scenario , species B is slowly activated by A* . This causes the activity of B ( B* ) to increase during the entire period of A's transient activation . If the kinetics for the deactivation of B also are slow , B activity remains elevated for a significant amount of time after the A* has returned to its basal level . In this case B effectively works as a decoder , transforming the duration of A activity into the amplitude of B activity . In other words , slow kinetics makes B an integrator capable of measuring how long the upstream signal has been on . In the second scenario depicted in Figure 2 , species C has fast activation and deactivation kinetics . As a result , the C* concentration closely mimics the behavior of A* , reaching a quasi-equilibrium level soon after the signal is received and rapidly returning to pre-stimulation levels once A activity ceases . In this case , quantitative information about the stimulus is preserved even when C* is saturated because it is encoded as signal duration . We note that dose-to-duration encoding does not place restrictions on what types of responses a cell can initiate . For example , positive feedback acting downstream of either components B or C can be used to convert transient pathway activation into a permanent developmental switch [12] . In this section we discuss mechanisms that can achieve dose-to-duration encoding . As previously mentioned , we are focusing on cases involving a sustained input , and therefore need to consider systems capable of adaptation or desensitization . In order to work as a dose-to-duration transducer , the duration of the output has to increase with the concentration of the stimulus . As we illustrate below , this is not a general property of adaptive systems . Figure 3 shows a number of architectures capable of performing the dose-to-duration transformation . The two pathway architectures depicted in Figure 3A consist of incoherent feed-forward loops [13] in which the upstream stimulus activates both a positive and negative regulator of the signaling protein K . For the system to show transient activity , negative regulation must occur on a slower time scale than the activation rate of K . As shown in the figure , this can be achieved if the negative regulation is mediated by an intermediate species X . This species can operate either by inhibiting activation of K by KK or by promoting deactivation of K . This type of architecture occurs in ERK signaling networks in which agonists , such as epidermal growth factor , causes transient extracellular signal-regulated kinase ( ERK ) activation by triggering rapid Ras activation followed by slow recruitment of its negative regulator , Ras GTP-ase regulating protein ( Ras-GAP ) , to the membrane [14] . Figure 3B shows two simple pathway architectures involving negative feedback loops that can exhibit adaptive behavior . In these examples , the signaling molecule activates its own negative regulator . In the first case , the negative regulator X increases the deactivation rate of K and in the second case X decreases the activation rate of K . Both strategies produce qualitatively similar behavior . Similar to the case of feed forward regulation , adaptive behavior in these systems requires the negative feedback to operate on a slower time scale than that of activation of K . This type of architecture plays a role in the regulation of ERK signaling by the enzymatic activation of members of the MAPK phosphatase group ( MKP's ) [15]–[17] , and in the regulation of cytokine signaling by the induction of suppressor of cytokine signaling ( SOCS ) proteins [18] . We focus primarily on the negative feedback system depicted by the left diagram in Figure 3B , but the results that follow easily generalize to the other architectures . The equations that describe this model are given in the Methods . To understand how this system performs the dose-to-duration transformation , it is helpful to consider the steady-state response curve of K as a function of the activity of the upstream component KK in the absence and presence of the negative regulator X . In Figure 4A , the left curve corresponds to the case in which X has been deleted . When present , the effect of the negative regulator X is to shift the signal-response curve to higher active KK concentrations . Accordingly , the right curve shown in Figure 4A corresponds to the case in which X is maximally activated . The response of the system now can be understood by considering how the signal-response curve shifts in time ( Figure 4B ) . This approach is valid because the requirement of slow negative feedback implies that the K and K* concentrations are in quasi-equilibrium with respect to the current X* concentration . Figure 4C ( left ) shows time series for an example in which the level of KK* is given by the blue vertical line in Figure 4B . Upon activation , KK* quickly drives the level of K activity to the steady-state value expected in the absence of the negative regulator X , which for this example corresponds to full activation . Subsequent to the initial rapid rise of K* , the activation and deactivation rates roughly balance ( Figure 4C , right ) and the relative ratios of K and K* remain in a quasi-equilibrium determined by the current level of X* . As the level of X* slowly increases , the signal-response curve gradually shifts to the right ( Figure 4B ) , and eventually , the EC50for K activation moves beyond the available concentration of KK . At this point the concentration of X* is such that the stimulus cannot counteract the level of negative regulation and K* activity returns to near basal levels ( Figure 4C , left ) . The two properties required for the dose-to-duration transformation are that ( i ) the activation rate of K is proportional to the stimulus concentration ( KK* concentration in the model under consideration ) and ( ii ) the kinetics of the negative regulator are slow . It is also important that the negative regulator induces a reversible change in K rather than on irreversible change , such as degradation or irreversible desensitization . Under these conditions , by slowly increasing the deactivation rate of K* , the system is actually “measuring” the activation rate of K rather than the K* concentration . The readout is the time necessary to produce enough X* to counteract the stimulated activation rate of K and bring K* back to basal levels . Importantly , this approach can potentially be used to measure stimulus concentrations much higher than those that would saturate pathway activity in the absence of negative feedback . In other words , by exploiting the time-dependent properties of the system , signaling pathways can increase their dynamic range . The dose-to-duration transformation described above occurs most efficiently when the activity level of the signaling component involved in the transformation is saturated . More generally , there is a repertoire of four operational regimes available to the adaptive system . These are summarized in Figure 4D and 4E . Figure 4D again shows the steady state response curves in the absence ( left ) and presence ( right ) of the negative regulator . In this figure , the graph has been expanded for illustrative purposes . The four shaded regions correspond to the different operational regimes . The first regime corresponds to low stimulus concentrations . When the stimulus strength increases within this regime , the response of the system consists of transient peaks of increasing amplitude but roughly the same duration ( Figure 4E , left ) . For each stimulus , the peak amplitude can be approximately determined by the signal-response curve in the absence of negative regulation . Increased upstream K activity increases the rate at which X is activated , and hence only a relatively weak dependence of the signal duration on the stimulus dose is observed . Regime II arises when the stimulus strength is sufficient to saturate K activity . This is the regime in which the dose-to-duration transformation occurs ( Figure 4E , center ) . Figure 4F presents the relationship between signal duration ( defined as time between half-maxima ) and stimulus concentration in Regimes I and II . In Regime III , the stimulus level is high enough so that the negative regulator is no longer able to counteract the induced activation rate of K , even when X is maximally activated . In this regime the system begins to lose its ability to adapt ( Figure 4E , right ) . If the stimulus level increases even further , the system operates in Regime IV and adaptation no longer occurs ( Figure 4E , right ) . In this regime , a sustained input produces a sustained output . Therefore , this pathway architecture is capable of acting as a switch; at low stimulus dose the response is transient , whereas at high levels the response becomes sustained . To illustrate how the transition between these regimes occurs , Figure 4G shows characteristic time series from each regime on the same graph . Physiological conditions and kinetic properties of signaling pathways may constrain some systems to operate in a subset of the theoretically possible regimes . The minimalistic systems described here are intended to illustrate some of the mechanisms available to signaling networks for producing dose-to-duration encoding . Although very simple , they are useful for understanding the behavior of more complex architectures . The addition of more pathway components would not change the underlying operating principles of dose-to-duration encoding . In fact , additional components can be used to generate more robust responses and provide more opportunities to fine-tune the input-output relations of the pathway . When operated in Regime II , the temporal profile of K* resembles a square pulse ( Figure 4C ) . This is because the signal-response characteristics of the system in the absence of the negative regulator were taken to be switch-like . Therefore , it is important to study how the dose-to-duration transformation is affected when this assumption is relaxed . We start by observing that the switch-like signal-response curves result from the small values of the Michaelis constants used in the reaction rates ( k1m , k2m , and k3m in Equation1 ) , which means that the reaction rates saturate at low substrate levels . In the opposite extreme , the activation rates operate far from saturation . In this case , the catalytic reactions can be described in terms of mass action kinetics . For the system to efficiently adapt , a relatively steep dose-response curve is still required . This can be achieved by manipulating the parameters involved in the negative regulation . For such cases , the system's response to a sustained stimulus is no longer a square pulse , but shows a more gradual decay in time ( cf . Figure 5C , middle ) . However , as we show next , the length of time required for the signal to decline below a given threshold still depends on the strength of the stimulus , and therefore the stimulus concentration can still be encoded as signal duration . The scenario discussed above is of particular interest because it applies to a situation in which the negative feedback loop acts at the level of the receptor . Figure 5A shows a schematic diagram of a model in which the ligand-bound receptor activates a negative pathway regulator X . The protein X inhibits the pathway by modifying the ligand-bound receptor ( phosphorylation in this example ) and decreasing its affinity for the ligand ( Figure 5A ) . Equations 2–4 of the Methods provide a mathematical description of this model . Figure 5B shows the steady-state receptor occupancy curves in the absence and presence of the negative regulator X . Temporal responses of the ligated receptor concentration for the four operational regimes are shown in Figure 5C . Note that while these time series do not have square pulse shapes , dose-to-duration encoding is still possible , because higher ligand levels cause active receptors to persist for longer times ( Figure 5C , middle ) . Furthermore , a square-pulse activity profile is easily generated if the pathway contains a downstream component with switch-like signal-response characteristics . As shown in Figure 5D , the pathway component B measures how long receptor occupancy remains above its activation threshold , thereby transforming the time series for receptor occupancy into a square-pulse of B activity . An important consequence of this pathway architecture is that it allows for “signaling beyond saturation” . That is , the system responds in a dose-dependent manner to ligand concentrations higher than required to saturate the receptor ( Figure 1C ) . In other words , the dissociation constant of the receptor can be dynamically modulated and exploited to expand the dynamic range of the signaling pathway . The mating response pathway of yeast mediates the organism's response to pheromone secreted into the medium by cells of the appropriate mating type . When bound with pheromone , a specific G-protein coupled receptor activates its cognate G protein causing the dissociation of the α and βγ subunits . The βγ complex then recruits the scaffold protein Ste5 to the membrane , which in turn recruits and activates a signaling cascade composed of Ste20 ( MAP4K ) , Ste11 ( MAP3K ) , Ste7 ( MAP2K ) , and the MAP kinases Fus3 and Kss1 ( Figure 6A , for a review see [3] ) . The developmental response initiated by yeast depends critically on the pheromone concentration . In the presence of very low levels or no pheromone , cells continue to grow and divide normally . At intermediate levels of pheromone , the cells become elongated and are capable of chemotropic growth towards a pheromone gradient . High levels of pheromone produce a bona fide mating response , involving cell division arrest and the emergence of mating projections [3] , [5] , [6] . We recently published an experimental study demonstrating that the scaffold protein Ste5 slows the activation rate of the MAP kinase Fus3 and that this slow activation underlies the developmental switch from chemotropic growth to mating [7] . In this section we present a mathematical analysis of the temporal profiles of MAP kinase activity measured as a part of our previous investigation . Our analysis suggests that the mating response pathway is using dose-to-duration encoding to relay information about the extracellular pheromone concentration . Figure 6B shows time course data for active ( dually-phosphorylated ) Fus3 and Kss1 as measured by immunoblotting for wild type cells in response to different pheromone concentrations ( see [7] and methods for details of the experimental methods ) . The transition from chemotropic growth to mating occurs between 3 and 10 µM , where there is a large increase in Fus3 activity [7] . Note the qualitative similarity between the experimental results and the graphs in Figure 2 ( compare pp-Fus3 to B* and pp-Kss1 to C* ) . The roughly dose-independent rate ( slope ) for Fus3 phosphorylation suggests that its activation rate is saturated . This behavior is consistent with the level of upstream kinase activity being independent of the pheromone dose , whereas the duration of this activity is dose-dependent . On the other hand Kss1 shows fast kinetics . Note that for high pheromone concentrations ( 10 µM ) , Kss1 seems to undergo two stages of phosphorylation with a second increase in phosphorylation starting around 30 min after exposure . If we disregard this second increase in Kss1 activity ( see Discussion ) , then by virtue of its fast kinetics , Kss1 phosphorylation mirrors the upstream signal dynamics . Furthermore , it appears from the data that for the doses measured , Kss1 operates in Regimes I and II ( and perhaps III ) of Figure 4 . These observations when combined with the very good correlation between the duration of Kss1 and Fus3 activity , suggest that Fus3 and Kss1 phosphorylation are driven by an upstream signal in which the pheromone dose has been converted to signal duration . To test the idea of dose-to-duration encoding , we sought to establish a single upstream input profile capable of reproducing the experimental results for both Kss1 and Fus3 . Specifically , we looked for a signal profile s ( t ) that when used as input to the equations for Fus3 activation ( Equation 7 ) and Kss1 activation ( Equation 8 ) generates the Fus3 and Kss1 activity time series shown in the left and right panels , respectively , of Figure 6B . The analysis produced the input signals and MAP kinase profiles shown in Figure 6C and 6D , respectively . The excellent agreement between the experimental data and model output provides strong evidence in support of dose-to-duration encoding by the pheromone response pathway . In principle , any of the encoding mechanisms discussed in the previous sections can produce temporal profiles similar to Figure 6C ( see Figures 4 and 5 ) , and there are several potential candidates for the negative feedback loop that mediates the dose-to-duration transformation in the mating pathway . These include transcriptional induction of either the RGS protein Sst2 , which increases the rate at which the Gα subunit hydrolyzes GTP [19] , or the protease Bar1 , which degrades pheromone [20] , [21] . Note that transcriptional induction takes 30 minutes or more . Because pathway deactivation occurs within 30 min at low pheromone concentrations , it is likely that feedback loops involving protein modifications also play a role in the dose-to-duration transformation . Furthermore , because the MAP kinases respond in a dose-dependent manner at pheromone concentrations significantly higher than the reported receptor Kd value of 5–15 nM [22] , [23] , it is plausible that dose-to-duration encoding involves feedback regulation of the receptor . This is not the only possibility and any target of feedback regulation at the level of the MAP2K Ste7 or above would work equally well ( see Discussion ) . With the above considerations in mind , we developed a mathematical model to investigate the scenario in which the negative feedback loop acts on the receptor . Figure 7A shows a schematic diagram of the system and the shape of the propagated signal at each level of the pathway . The model is described by Equations 3–5 and 7–9 of the Methods . As discussed in the previous section , because the negative feedback acts on the receptor , it is necessary to incorporate an intermediate step ( MK in Figure 7 ) to transform the propagated signal into a square-pulse . Any of the upstream kinases ( Ste20 , Ste11 or Ste7 ) are capable of performing this transformation . Figure 7B shows the predicted upstream activation profile ( compare with Figure 6C ) and the MAP kinase activation profiles produced by the models compared with the experimental results . If we again disregard the second increase in Kss1 activity at high pheromone concentrations , the correspondence between the model results and experimental data is striking , especially considering the simplicity of the model . We note that this agreement does not prove the validity of the model , but demonstrates that the mechanisms discussed above are consistent with the experimental data . The model also provides an important guide for future experimental work .
It is widely accepted that signaling pathways are capable of transmitting quantitative information about their surrounding environment . While the importance of transient versus sustained signaling has been recognized for some time [24]–[26] , most previous investigations have focused on information transfer using amplitude encoding without considering the temporal aspects of signal transduction [8] , [9] . Here we demonstrate that dose-to-duration encoding provides cells with an alternative mechanism for processing and transmitting quantitative information about their surrounding environment . The ability of signaling pathways to convert stimulus strength into signal duration results directly from the nonlinear nature of these systems and emphasizes the importance of considering the dynamic properties of signaling pathways when characterizing their behavior . Taken together , our computational and experimental results suggest that dose-to-duration encoding occurs in the pheromone response pathway of yeast and underlies the developmental switch from chemotropic growth to mating . One important advantage of dose-to-duration encoding is that it has the potential to increase the dynamic range of signaling pathways . One way this can occur is if feedback regulation allosterically modifies the receptor's affinity for the ligand . Dynamically regulating the Kd of the receptor has the interesting effect of shifting the EC50 of the cellular response to the right of the receptor occupancy curve ( Figure 1C ) . Depending on the response of downstream components , the dose-response curve for the system is not only shifted but also stretched . This highlights the important point that receptor occupancy curves are potentially time-dependent quantities and need to be interpreted with care . Interestingly , Kd values determined in vitro or in reconstituted systems usually differ from those obtained in vivo , and this discrepancy is often attributed to an abnormal conformation of the receptor in the artificial environment . The analysis presented here suggests that even when there is correspondence between the microenvironment of an in vitro experiment and the macroenvironment of a cell culture , the results of ligand binding assays might differ due to dynamic regulation of the receptor in vivo . For example , this could happen if a downstream element of the signaling pathway has been disrupted in the in vitro experiment , thereby breaking the negative feedback loop . Interactions between receptors , in particular G-protein coupled receptors ( GPCRs ) , and cytosolic proteins have been shown to affect receptor-ligand affinity [27] . Most GPCR's are known to undergo biochemical modifications , such as phosphorylation , and to interact with a number of signaling proteins , including G proteins , arrestins , kinases , RGS proteins , and to form oligomers , all of which could affect affinity for the ligand . Therefore dynamic regulation of a receptor as a mechanism for dose-to-duration encoding seems quite plausible . Dose-to-duration encoding may also provide a more robust transmission mechanism than amplitude-encoding in multilevel networks . This is because accurate transfer of information using amplitude encoding requires that the input-output characteristics of the individual components be well matched [8] . Note that dose-to-duration encoding does not have to function throughout the whole pathway . It is likely that multiple information processing strategies coexist at different levels ( or even under different conditions ) in a single pathway . In fact , the use of multiple information processing strategies may provide signaling networks with more flexibility when responding to changing environmental conditions . Another potential advantage of dose-to-duration encoding arises from the need to prevent spurious activation of pathways that share components . Recently we proposed “kinetic insulation” [28] as a strategy for achieving pathway specificity . Kinetic insulation relies solely on the temporal profiles of the propagated signals to ensure signal fidelity . It requires that at least one of the pathways responds transiently . Because signal duration is a natural strategy for pathways with transient activity to encode information , signaling systems with shared components are potential candidates for dose-to-duration encoding . Consistent with these ideas , the yeast pheromone response pathway contains several signaling proteins ( e . g . , Ste11 and Ste7 ) that are known to also participate in the hyper-osmotic shock [29] and filamentous growth [30] , [31] pathways . Our modeling results and experimental data provide compelling evidence for dose-to-duration encoding by the yeast pheromone response pathway . A key question is then what is the molecular mechanism responsible for transducing stimulus dose into signal duration ? We have demonstrated that a scenario in which feedback regulation acts at the level of the receptor is consistent with our experimental data for MAP kinase activity . Our motivation for considering such a mechanism came from data suggesting that yeast continue to respond in a dose-dependent manner to pheromone concentrations well beyond the reported value for the receptor dissociation constant . As we have shown , by dynamically altering the affinity of the receptor for pheromone , our model provides an explanation for this phenomenon of signaling beyond saturation . Modulation of the receptor affinity in yeast might occur by interactions with other receptors ( receptor dimers ) [32] , the G-protein [27] , or the RGS protein Sst2 [33] . Similarly , affinity could be altered through receptor phosphorylation or ubiquitination [34] . GTP-dependent changes in the pheromone receptor affinity attributed to the interaction with the G protein have been reported [27] . Although the physiological relevance of this effect has not been clearly established in yeast , this is an example of a phenomenon observed for many mammalian GPCR's in vitro . We note that interpreting dose-response data for the pheromone pathway is complicated by the presence of the protease Bar1 [20] , [21] . In fact , a mechanism based solely on Bar1 degradation of pheromone is in theory sufficient to achieve the dose-to-duration transformation . However , our recent experiments performed in a bar1Δ mutant show cells responding to lower pheromone doses , but with time courses of MAPK activity that are consistent with dose-to-duration encoding ( Supplemental Data [7] ) . These results argue against a mechanism involving Bar1 alone . It should be emphasized that dose-to-duration encoding does not require the negative feedback to act at the level of the receptor . For example , induction of the negative regulator Sst2 [35] , feedback phosphorylation of an upstream pathway component [36] , or receptor endocytosis could also accomplish this transformation , although they would not account for the observed shift in the EC50 . Thus , it is clear more work is necessary to unambiguously identify the mechanisms responsible for information transfer in the pheromone response pathway . However , the remarkable agreement between our modeling results and experimental data offers strong evidence in support of dose-to-duration encoding and provides a foundation for interpreting future experimental results . Interestingly , the combination of fast and slow kinetics exhibited by the two MAP kinases , Kss1 and Fus3 , has the potential to form a feed-forward adaptive system . It has been demonstrated that pheromone-induced degradation of the transcriptional activator Ste12 requires Fus3 , but not Kss1 [37] . Ste12 might also play a role in generating the second peak of Kss1 activity observed at high pheromone concentrations , a possibility that we are now investigating . In the absence of pheromone , Kss1 acts as a transcriptional repressor by forming a complex with Ste12 and the proteins Dig1 and Dig2 ( also known as Rst1/2 ) [38] . It is possible that pheromone-stimulated release of Kss1 from this complex [39] generates a second pool of Kss1 and this pool is responsible for the second peak of activity . However , at this point we cannot rule out alternative explanations including transcriptional induction , re-localization , or positive feedback . Dose-to-duration encoding is not restricted to yeast . For example the intensity of light ( number of photons ) impinging on photoreceptors in rod cells is encoded as the duration of G protein-mediated activity of the pathway [40] . It has been shown recently that the RGS protein RGS9 plays a particularly important role in determining the duration of the signal [41] . Furthermore , switches based on transient versus sustained signals , like the ones arising from transitions between the regimes of Figure 4 , have been proposed to underlie cell fate decision process in a number of systems [14] , [26] , [42] , [43] . The recent discovery that different temporal profiles of IκB kinase ( IKK ) activity in the NF-kB signaling module selectively activate different groups of target genes , further supports the notion that dose-to-duration encoding plays a significant regulatory role in determining cellular responses . In this case , stimulation of murine embryonic fibroblasts with tumor necrosis factor-α produces a short transient peak of IKK activity whereas stimulation with polysaccharides results in a slower and more sustained IKK response [44] , [45] . The fact that each profile affects the expression of different groups of genes illustrates how the temporal dynamics of a signaling pathway can play a role in determining pathway specificity . Finally , it is remarkable that the simple pathway architectures considered here can generate such a variety of responses depending on the strength of the stimulus ( Figure 4 ) . These systems not only function as amplitude and dose-to-duration encoders , but also can act as biochemical switches that transition from transient to sustained outputs potentially generating different physiological responses [26] , [42] , [46] . Typical signaling pathways involve multiple levels of regulation that in general should lead to even more complex behavior . Our results demonstrate how quantitative measurements of the temporal patterns of pathway activity when combined with mathematical modeling can be used to discover the design principles upon which signaling networks operate and decipher the code used by these systems to transmit information .
In this section we describe the mathematical models used to generate the results presented in the figures . All the differential equations were solved using Mathematica by Wolfram Research . The model depicted in Figure 3B ( left ) , in which the mechanism for pathway adaptation involves a negative feedback loop that increases the deactivation rate of K* , is described by the following two equations: ( 1 ) ( 2 ) Where [K]Total = [K*]+[K] and XTotal = [X*]+[X] . The parameter values used to generate the results shown in Figure 4 are ( in arbitrary units ) : k1 = 1 , k1m = 10−2 , k2 = 10−2 , k2m = 10−2 , k3 = 8 , k3m = 10−2 , k4 = 10−4 , k4m = 10−1 , k5 = 5 10−6 , k5m = 1 . The curves in this figure correspond to s values of: 10−2 , 2 10−2 , 3 10−2 , 4 10−2 , 6 10−2 ( region I ) , 10−1 , 3 10−1 , 5 10−1 , 1 , 1 . 5 ( region II ) , and 6 . 0 , 7 . 0 , 7 . 5 , 8 . 0 , 8 . 5 , 20 . 0 ( region III+IV ) . For the model in which the negative feedback acts on the receptor , the equations are: ( 3 ) ( 4 ) ( 5 ) where RTotal = [R]+[RL]+[RL*] . For simplicity , ligand release and receptor de-phosphorylation are taken to occur in a single step . This simplification does not affect the results provided both biochemical steps are not rate limiting . Even if this separation of times scales does not exist , we do not expect a more detailed model that separates these events to produce qualitatively different behavior . To transform the transient response in Figure 5C into a square pulse the following equation for B* was used ( 6 ) The parameters used to produce the results shown in Figure 5 are k1 = 1 , k2 = 10−2 , k3 = 80 , k4 = 1 10−4 , k4m = 10−1 , k5 = 5 10−6 , k5m = 1 , k0 = 10 , k0m = 10−1 , k6 = 10 , k6m = 10−2 , k7 = 4 , k7m = 10−2 . The curves correspond to s values of: 1 10−2 , 2 10−2 , 3 10−2 , 5 10−2 , 1 10−1 ( region I ) , 10−1 , 3 10−1 , 5 10−1 , 1 , 1 . 5 ( region II ) , and 6 , 10 , 15 , 20 , 50 , 500 ( region III+IV ) . The equations used to model the kinetics of Fus3 and Kss1 activation are ( 7 ) ( 8 ) respectively , where Fus3Total = [ppFus3]+[Fus3] , Kss1Total = [ppKss1]+[Kss1] . The parameters used to produce the results shown in Figure 6 are: k10 = 5 . 53 10−4 , k10m = 3 . 75 10−2 , k20 = 3 . 25 10−4 , k20m = 3 10−1 , k30 = 2 . 55 10−2 , k30m = 1 , k40 = 2 . 5 10−3 , k40m = 2 . The input signals consist of a square pulse of duration tpulse followed by an exponential decay ( i . e . , signal = S for time<tpulse , and signal = S e− ( time-tpulse ) /λ for time>tpulse ) . The signal parameter for each concentration were as follows: S = 0 . 2 , 0 . 25 , 0 . 75 , 0 . 75 , 0 . 75 , tpulse = 55′ , 22′ , 6′ , 4′ , 4′ , and λ = ( 50 , 50 , 250 , 300 , 300 ) ×3600 min . The full model depicted in Figure 7 is described by Equations 3–5 and 7–8 above , in which s has to be replaced by [MK*] . The following equation describes the dynamics of MK*: ( 9 ) Here MKTotal = [MK*]+[MK] . The parameters used to produce Figure 7 are ( arbitrary units ) : k1 = 2 , k2 = 3 10−2 , k3 = 1 . 9 102 , k4 = 1 10−4 , k4m = 3 10−2 , k5 = 8 . 5 10−8 , k5m = 2 . 5 10−2 , k0 = 6 . 6 101 , k0m = 5 . 1 10−2 , k10 = 4 . 1 10−4 , k10m = 4 . 4 10−4 , k20 = 5 . 9 10−4 , k20m = 4 . 6 10−1 , k30 = 2 . 8 10−2 , k30m = 2 . 6 , k40 = 1 . 15 10−3 , k40m = 3 . 8 10−1 , k6 = 3 . 2 k6m = 4 . 9 10−4 , k7 = 1 . 7 , k7m = 3 . 3 10−1 . As described in [47] the signaling modules presented above are capable of producing adaptive behavior for a wide range of parameter values . The main condition that must be met is that activation occurs on a fast time scale as compared to the feedback inhibition . The parameters for the examples used to illustrate dose-to-duration encoding were selected to comply with this requirement . The parameters for Figure 6 were tuned manually to generate a good fit to the data . However , the number of experimental points leaves significant leeway for the exact shape of the decay phase of the input signal . The parameters used to generate the curves for Figure 7 were obtained using a Monte Carlo algorithm . The values of the rate constants associated with ligand binding and dissociation in the absence of feedback regulation were fixed to reflect a Kd value of 15 nM [23] . Immunoblot data for kinases Fus3 and Kss1 were obtained from [7] . Briefly , BY4741 ( MATa leu2Δ met15Δ his3Δ ura3Δ ) cells were grown using standard practices . Cell extracts ( 20 µg/lane ) were resolved by 12% SDS-polyacrylamide gel electrophoresis and immunoblotting performed as described in [42] . Band intensity was quantified by scanning densitometry using ImageJ ( National Institutes of Health ) . | Cells must be able to detect and respond to changes in their surroundings . Often environmental cues , such as hormones or growth factors , are received by membrane receptors that in turn activate intracellular signaling pathways . These pathways then transmit information about the stimulus to the cellular components required to elicit an appropriate response . In many cases , the nature of the response depends on the dose of the stimulus . Thus , in addition to relaying qualitative information ( e . g . , the presence or absence of a stimulus ) , signaling pathways must also transmit quantitative information about the intensity of the stimulus . Here we introduce “dose-to-duration” encoding as an effective strategy for relaying such information . We demonstrate that by providing a mechanism for overcoming saturation effects , modulation of signal duration increases the range of stimulus concentrations for which dose-dependent responses are possible . This increased dynamic range produces the counterintuitive result of “signaling beyond saturation” in which dose-dependent responses are still possible after apparent saturation of the receptors . Finally , we demonstrate that dose-to-duration encoding is used in the yeast mating response pathway and presents a simple mechanism that can account for current experimental observations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"cell",
"biology/cell",
"signaling",
"physiology/cell",
"signaling",
"biochemistry/cell",
"signaling",
"and",
"trafficking",
"structures",
"pharmacology",
"biotechnology/bioengineering",
"computational",
"biology",
"computational",
"biology/signaling",
"networks",
"biophysics/cell",
"signaling",
"and",
"trafficking",
"structures",
"computational",
"biology/systems",
"biology"
]
| 2008 | Dose-to-Duration Encoding and Signaling beyond Saturation in Intracellular Signaling Networks |
Human and animal African trypanosomiasis ( HAT & AAT , respectively ) remain a significant health and economic issue across much of sub-Saharan Africa . Effective control of AAT and potential eradication of HAT requires affordable , sensitive and specific diagnostic tests that can be used in the field . Small RNAs in the blood or serum are attractive disease biomarkers due to their stability , accessibility and available technologies for detection . Using RNAseq , we have identified a trypanosome specific small RNA to be present at high levels in the serum of infected cattle . The small RNA is derived from the non-coding 7SL RNA of the peptide signal recognition particle and is detected in the serum of infected cattle at significantly higher levels than in the parasite , suggesting active processing and secretion . We show effective detection of the small RNA in the serum of infected cattle using a custom RT-qPCR assay . Strikingly , the RNA can be detected before microscopy detection of parasitaemia in the blood , and it can also be detected during remission periods of infection when no parasitaemia is detectable by microscopy . However , RNA levels drop following treatment with trypanocides , demonstrating accurate prediction of active infection . While the small RNA sequence is conserved between different species of trypanosome , nucleotide differences within the sequence allow generation of highly specific assays that can distinguish between infections with Trypanosoma brucei , Trypanosoma congolense and Trypanosoma vivax . Finally , we demonstrate effective detection of the small RNA directly from serum , without the need for pre-processing , with a single step RT-qPCR assay . Our findings identify a species-specific trypanosome small RNA that can be detected at high levels in the serum of cattle with active parasite infections . This provides the basis for the development of a cheap , non-invasive and highly effective diagnostic test for trypanosomiasis .
African trypanosomes , vector borne protozoa transmitted by tsetse flies ( Glossina species ) , cause Human African Trypanosomiasis ( HAT ) and Animal African Trypanosomiasis ( AAT ) across sub-Saharan Africa . AAT , caused by Trypanosoma congolense , Trypanosoma vivax and Trypanosoma brucei , infects approximately 70 million and kills 3 million cattle per year , and is one of the most significant infectious disease constraints upon agriculture in the region [1] . HAT is caused by two variants of T . brucei , T . b . gambiense and T . b . rhodesiense , and in recent years the impact of this disease has been significantly reduced through active case detection , with <3 , 000 cases reported in 2015 , down from ~50 , 000 in 2000 [2 , 3] . However , new and improved tools are required for both diseases; for AAT as a tool to begin to reduce the current significant infection burden , and for HAT to facilitate the delivery of the WHO aim of HAT elimination by 2030 [4 , 5] . The ability to diagnose active infections is currently still a significant challenge for both AAT and HAT . While there have been substantial efforts to develop new effective diagnostics for HAT , currently the gold standards remain microscopy ( with methods that concentrate–e . g . microhaematocrit centrifugation , quantitative buffy coat or mini anion exchange centrifuge technique ( mAECT ) [6–9]–all providing increased sensitivity ) and the card agglutination test ( CATT ) –an antibody agglutination test based upon several VSGs expressed by T . b . gambiense [10] ( for the latter assay there has also been recent adaptation to a rapid diagnostic test platform [11 , 12] ) . While some molecular tests ( e . g . loop mediated isothermal amplification–LAMP ) have demonstrated promise [13] , and for LAMP this has included the development of field-applicable kits , these have not been widely utilised in the field [14] . Tests based on antibody and DNA have their well-recognised limitations ( differentiating between exposure and infection for the former and the potential for DNA persistence following treatment , as well as contamination , for the latter ) , and a test that enabled sensitive and specific detection of active infection would be a significant advance . For AAT , diagnosis is still largely symptomatic with inherent non-specificity given lack of pathognomonic clinical signs , and occasionally microscopy may be employed [15] . Investment in development of diagnostics for AAT is increasing , with recent efforts defining antibody-based capture techniques for antigens that have been described as conserved and highly expressed [16–18] . Indeed , this approach has resulted in the first commercial diagnostic being brought to market in 2017 ( VerY Diag , CEVA ) . Therefore , available methods for both HAT and AAT have their limitations–the requirement for a test that enables detection of active infection remains–both for potential utility in the field and to improve , for example , accurate assessment of clinical efficacy of drugs and vaccines ( increasing areas of interest for AAT ) . An ideal marker for active infection is a pathogen-derived molecule that is present in high enough levels in infected animals/patients to enable sensitive detection , has properties that enable assignment to pathogen and species to a high level of confidence , and , additionally , reduces in levels quickly following removal of the pathogen ( e . g . by chemotherapy ) . Small RNAs have received much interest as potentially useful diagnostic biomarkers , particularly in human medicine and cancers [19] . This is due to higher expression of particular small RNAs ( e . g . microRNAs [miRNAs] ) in cancer cells . In these cases , diagnosis requires confirmation of higher levels of the small RNA species in comparison to non-affected cells/tissues . For application to pathogens in contrast , the test would aim to identify the binary presence or absence of a pathogen marker , a much simpler threshold to define . Trypanosomes produce multiple small RNAs ( although do not produce miRNAs ) and in the best-characterised species , T . brucei , there has been description of the small RNAome [20] . The T . brucei genome includes identification of small RNA encoding loci , including rRNA , snoRNA , tRNA and siRNA [21 , 22] ( albeit only a proportion of these have been functionally validated ) . In addition , several reports have outlined the RNA species secreted/excreted in the form of vesicles by related trypanosomatids such as Trypanosoma cruzi and Leishmania major [23 , 24] . There is less information for T . congolense and T . vivax , although both species have annotated genomes available with predicted small RNA-encoding genes [21 , 25] . In the current study we describe a trypanosome small RNA species that is present in the serum of infected animals at high levels . This small RNA is a 26-nucleotide segment of the 7SL long non-coding RNA ( ‘7SL RNA’ ) ; the latter is usually described as a cytoplasmic non-coding RNA that is part of the signal recognition particle ( SRP ) involved in protein translocation across cell membranes . The 7SL-derived small RNA ( hereafter termed ‘7SL-sRNA’ ) sequences are species-specific , enabling the design of tests that differentiate between T . congolense , T . vivax and T . brucei . The 7SL-sRNA is present at high levels in infected animals ( equivalent to levels of highly expressed bovine miRNAs ) , enabling robust detection both before detection by microscopy and during periods of infection with subpatent parasitaemia . Importantly , following post-curative treatment the levels of the 7SL-sRNA drops to undetectable levels . Therefore , we believe that the 7SL-sRNA represents a suitably sensitive and specific marker for detection of active infection in trypanosomes , with potential utility for both HAT and AAT .
Animal experiments were carried out at the Roslin Institute , University of Edinburgh under the auspices of Home Office Project License number 60/4394 . Studies were approved by the Roslin Institute Animal Welfare and Ethical Review Board ( study numbers L172 and L223 ) . Care and maintenance of animals complied with University regulations and the Animals ( Scientific Procedures ) Act ( 1986; revised 2013 ) . Protocol plans for studies carried out at Clinvet were submitted to the Institutional Animal Care and Use Committee ( IACUC ) , which issued certificates of approval . The protocol was designed to allow the use of the study animals in compliance with the Clinvet policy on the ethical use of animals , using the most recent version South African National Standard ( SANS ) 10386 ( The care and use of animals for scientific purposes ) . Approved study numbers were CV 15/192 and CV 16/306 . Samples from infected animals derive from two sources; ( i ) experimental infections carried out at the Roslin Institute ( T . brucei and T . congolense ) and ( ii ) experimental infections carried out as part of candidate drug testing by GALVmed/Clinvet ( T . congolense and T . vivax ) . T . congolense IL3000 BSF parasites were cultured in TcBSF3 medium [28] supplemented with 20% adult goat serum ( Gibco ) , 0 . 12 mM 2-mercaptoethanol and penicillin/streptomycin and incubated at 34°C , 5% CO2 . Cells were routinely passaged and maintained at a density between 5 × 104 cells/mL and 3 × 106 cells/mL , unless stated otherwise . T . brucei Lister 427 cells were cultured in HMI-11 [29] supplemented with 10% FBS , 0 . 2 mM 2-mercaptoethanol and penicillin/streptomycin , and maintained at 37°C , 5% CO2 . Cells were maintained between 2 × 104 cells/mL and 2 × 106 cells/mL . Bloodstream forms of group 1 T . b . gambiense strain ELIANE were cultured in HMI-9 supplemented with 20% serum plus , as previously described [30] . RNA extractions were conducted using the TRIzol LS reagent ( Invitrogen ) following the manufacturer’s instructions . 250 μL of starting material in the form of serum/plasma for in vivo experimental infections carried out at the Roslin Institute ( T . brucei and T . congolense ) was used; where the sample was less than 250 μL , distilled water was added to make up the volume . For in vivo samples received from GALVmed/Clinvet , RNA was extracted from 125 μL of plasma with distilled water added to make up the volume to 250 μL . In cases where in vitro culture supernatants were used for RNA extraction , 500 μL supernatant was centrifuged at 2 , 000 × g for 10 minutes to remove cells from the medium . Subsequently , 250 μL supernatant was used for downstream experiments . Libraries were prepared using the TruSeq Small RNA library preparation kit ( Illumina ) with 10 μL total RNA as starting material ( quantity of total RNA for each sample: uninfected serum– 39ng , infected sample 1 – 230ng , infected sample 2 – 270ng , parasite cell pellet– 2 . 1μg ) . Samples were enriched using 15 cycles of PCR and library products of 145–160 bp were gel purified , quantified and pooled for sequencing . The library pool was sequenced using a HiSeq 2500 with 50-base single end reads and V4 chemistry . A species-specific 7SL-derived small RNA stem loop primer-probe detection assay was optimised , using custom primer and probe mixes made by Life Technologies , based on specific sequences ( Custom TaqMan Small RNA assay , cat . number: 4398989 [assay IDs T . brucei: CTFVKNM; T . congolense: CTRWEM9; T . vivax: CTDJXGZ] ) . Reverse transcription was carried out using a commercial cDNA Reverse Transcription Kit ( Applied Biosciences , cat . number: 4368814 ) , replacing the random primers with the aforementioned TaqMan assay primer . Typically , 100 ng RNA was used per 15 μL reaction . The following thermocycling conditions were applied for the RT reaction: 16°C for 30 minutes , 42°C for 30 minutes and 85°C for 5 minutes to inactivate the reverse transcriptase . In vivo RNA samples from the GALVmed/Clinvet trial were isolated from plasma derived from heparinised blood and therefore required 2 units of Bacteroides Heparinase 1 ( New England BioLabs , cat . number: P0735 ) per RT reaction . Subsequent to the RT reaction , a qPCR was performed using a commercial kit ( TaqMan universal PCR master mix , Thermo , cat . number: 4304437 ) , according to manufacturer’s instructions . At this stage , 1 μL custom prime-probe was also added to the qPCR reaction , along with 1 . 5 μL RT reaction . The qPCR cycling profile was as follows: 50°C for 2 minutes , 95°C for 10 minutes and 40 cycles of 95°C for 15 seconds and a probe detection step of 60°C for 1 minute . When serum was used as a substrate for RT-qPCR , samples were heat treated at 65°C for 15 minutes and 6 μL of the serum used per RT reaction . Single step RT-PCR reactions were performed according to manufacturer’s guidelines ( TaqMan RNA-to-Ct 1-Step Kit , Life Technologies , cat . number: 4392653 ) . Single step RT-PCR reactions were performed using the TaqMan small RNA assay primer and primer-probe mixes mentioned previously . Raw RNA deep-sequencing data were subject to quality control using FastQC ( v0 . 11 . 5 ) [31] . Adapter sequences were then removed from the reads , and data was filtered for read length between 20 and 39 base pairs using cutadapt v1 . 7 . 1 ( parameters: “–a TGGAATTCTCGGGTGCCAAGG–m 20 –M 39” ) [32] Reads were subsequently aligned to the T . congolense IL3000 genome ( TriTrypDB , v9 . 0 ) , using Novoalign ( v3 . 02 . 12 , Novocraft Technologies ) , with the following parameters: “–l 20 –t 30 –h 60 –m–o SAM” . To quantify the alignments , the python tool HTSeq-count was used with default parameters [33] with the T . congolense IL3000 transcript reference file in . gff format downloaded from TriTrypDB ( v9 . 0 ) [21] . Data was normalised by calculating reads per million ( RPM ) : all read counts in a sample were first summed and the sum divided by 1 million to generate a “per million” scaling factor . Read counts were subsequently divided by this scaling factor to generate the RPM value for each gene . Raw and processed data is available through GEO accession number GSE122858 .
Small RNAs in blood represent attractive diagnostic biomarkers as they tend to be relatively stable , are easily accessible , and sensitive technologies exist for direct detection from serum samples . To test whether trypanosomes secrete or excrete small RNAs during in vivo infections , total RNA was extracted using Trizol LS from serum samples obtained from two cattle experimentally infected with the livestock trypanosome T . congolense ( samples taken at day 19 post-infection , parasitaemia at time of isolation approximately 5 x 106 /mL ) , serum from an uninfected control cow , and an in vitro-derived T . congolense cell pellet ( approximately 4 x 107 cells ) . The RNA was submitted for small RNA deep sequencing , selecting for RNAs between 20- and 39-bp long . The resulting reads were aligned to the T . congolense genome ( TriTrypDB v9 . 0 ) using novoalign with strict parameters ( one mismatch per read and a homopolymer filter score of 60; normalised results and alignment statistics , are available in S1 Table and S2 Table ) . A total of 15 , 645 , 557 and 16 , 770 , 619 reads were obtained for the two samples from infected cattle after filtering for read length between 20-bp and 39-bp . Of these , 4 . 2% ( 654 , 025 reads ) and 1 . 3% ( 218 , 788 reads ) were uniquely mapped to the T . congolense genome . 6 , 290 , 490 reads were obtained from the uninfected cattle sample , with only 0 . 03% ( 1 , 804 reads ) aligning uniquely to the T . congolense genome . Reads were also aligned to a bovine genome ( Bos taurus , UMD [34] ( S2 Table ) . These results indicated that 87 . 4% , 87 . 3% and 78 . 6% of reads from the first infected , second infected and uninfected cattle samples , respectively , aligned to the bovine genome ( including both uniquely aligned and multimapped reads ) . A total of 9 , 439 , 764 reads were generated from RNAseq of the T . congolense cell pellet sample , with 11 . 5% ( 1 , 084 , 005 reads ) and 58 . 9% ( 5 , 557 , 663 reads ) unique and multimapped reads aligning to the T . congolense genome , respectively . The relatively high number of unmapped reads ( 2 , 787 , 320; 29 . 5% ) in this sample with respect to the T . congolense genome is explained by the comparatively incomplete assembly and annotation of the reference T . congolense genome when compared to , for example , the genome of T . brucei . Subsequent analysis showed that the majority of mapped reads in the uninfected cattle sample that mapped to the T . congolense genome aligned to ribosomal RNA ( rRNA ) loci , the sequences of which are known to be deeply conserved in eukaryotes [35] . For this reason , rRNA alignments were omitted from downstream analyses as a data filtering step as they are therefore unlikely to be useful molecular diagnostic targets . Read counts from annotated regions of the T . congolense genome were generated using HTSeq-count , resulting in total read counts mapping uniquely to annotated features of 42 , 302 and 18 , 230 for the two infected samples , 322 , 385 for the pellet sample and 524 for the uninfected sample . Read counts were normalised for library depth ( reads per million; RPM ) . Strikingly , after eliminating reads associated with rRNA loci , the majority of reads from both infected serum samples originated from one specific 26-bp sequence ( Fig 1; normalised dataset in S1 Table ) . Indeed , there was a substantial difference between the abundance of this small RNA and the next most abundant small RNA observed in infected serum , as well as uninfected and cell pellet controls ( Fig 1A , detailed in Table 1 ) . The reduced levels of the small RNA in the cell pellet sample relative to the serum samples and subsequent analysis of culture supernatant ( see below & Fig 2 ) , suggests that the RNA species is rapidly secreted/excreted from the cell post-processing . Further analyses indicated that the sRNA uniquely mapped to a single copy locus on chromosome 8 that comprised part of the 275-bp 7SL RNA gene ( Signal Recognition Particle ( SRP ) RNA , T . congolense Gene ID: TcIL3000_8_ncRNA004; T . brucei Gene ID: Tb927 . 8 . 2861 ) ( Fig 1B ) , and is henceforth referred to as “7SL-sRNA” . The full secondary structure of the 7SL RNA is shown in Fig 1C , with 7SL-sRNA highlighted in red . Notably , a sequence corresponding to the 7SL-sRNA complementary strand was also detected in the RNAseq data , although at approximately ten-fold lower abundance , suggesting the existence of a passenger strand following processing of the 7SL-sRNA ( Fig 1B ) . There were no sequences corresponding to the host 7SL RNA detected , suggesting that generation of a small RNA from the 7SL RNA is specific to trypanosomes and not a general feature of 7SL RNA processing . The 7SL-sRNA sequence of T . congolense was aligned to the genome assemblies of several species of African trypanosome to determine whether related trypanosome species encode for the 7SL-sRNA species . Sequences corresponding to the 7SL-sRNA were clearly identifiable in the genomes of all related trypanosome species examined , suggesting expression of 7SL-sRNA may be a common feature of African trypanosomes , and indeed , related trypanosomatids . Whilst no sequence variation was observed across any of the T . brucei subspecies ( specifically T . b . brucei , T . b . gambiense , T . b . rhodesiense and T . b . evansi ) , there were several nucleotide polymorphisms relative to the T . brucei sequence in both the T . vivax and T . congolense sequences , raising the possibility that specific assays could be designed to distinguish between the three primary livestock trypanosome pathogens ( Fig 2A ) . To investigate this further , custom-designed primers were developed using existing stem-loop technology for each individual species and RT-qPCR experiments performed ( Fig 2B ) . Each primer set was applied to RNA extracted from serum samples from cattle experimentally infected with each species to test for cross-reactivity . Results show that sequence divergence of the 7SL-sRNA is sufficient to enable the design of RT-qPCR assays that differentiate between T . vivax , T . congolense and T . brucei with no detectable cross-reactivity ( Fig 2C ) . Importantly , when applied to supernatants derived from the human-infective T . b . gambiense ( ELIANE strain ) the T . brucei RT-qPCR assay resulted in positive detection of the 7SL-sRNA , highlighting the potential of the sRNA for diagnostics in human disease ( Fig 2C ) . For the 7SL-sRNA to be a suitable target for development of molecular diagnostics , there is a requirement that the sRNA is constitutively released into the bloodstream , rather than only under certain conditions such as cellular stress , as has recently been shown with , for example , the spliced leader RNA [36] . To investigate this , time courses lasting 3 days ( 72 hours ) were carried out using in vitro cultures of both T . brucei ( Lister 427 ) and T . congolense ( IL3000 ) . Cells were seeded at 5 × 104 cells/mL ( n = 2 ) , and density was periodically counted by haemocytometer and supernatant samples were taken simultaneously for RT-qPCR analysis ( Fig 3 ) . By the first time-point , the small RNA was readily detected in both T . brucei ( Fig 3A ) and T . congolense ( Fig 3B ) supernatants ( mean cell densities: T . brucei , 2 . 37 × 105 cells/mL; T . congolense , 3 . 5 × 104 cells/mL ) , as calculated relative to the zero hour time point . Furthermore , relative 7SL-sRNA levels appeared to increase correlating with cell density ( T . brucei: Pearson = 0 . 7724 , Spearman ρ = 0 . 9643; T . congolense: Pearson = 0 . 9353 , Spearman ρ = 0 . 9643; Fig 3A ) . Taken together , these data indicate that the 7SL-sRNA is constitutively released by both species of parasite , and indeed , relative abundance of the sRNA can give an indication of cell density in parasite cultures . Interestingly , when 7SL-sRNA levels were corrected for cell number and directly compared , levels of 7SL-sRNA accumulated more rapidly in T . congolense than T . brucei cultures , suggesting there may be species-specific kinetics of extracellular production ( Fig 3C ) . To further investigate the suitability of the 7SL-sRNA as a diagnostic for monitoring disease progression , serum samples were obtained from an in vivo study of six calves experimentally infected with 1 × 106 T . brucei AnTat 1 . 1 , which remained untreated for the duration of infection . The infection time courses ranged from six to 28 days depending on the severity of infection and day of euthanasia , and parasitaemia score was determined by microscopy approximately every two days ( Fig 4 ) . Total RNA was extracted from serum samples and analysed by RT-qPCR . The relative expression of 7SL-sRNA was calculated relative to the zero hour time point . Parasites were typically detected in blood by microscopy after 4 days ( Fig 4 ) . In contrast , the 7SL-sRNA was detected by day 2 , suggesting higher sensitivity compared to microscopy . Furthermore , following the first peak of parasitaemia , parasites became subpatent by microscopy , yet the 7SL-sRNA was still detectable at high levels during this time ( animals 6630 and 6632; Fig 4 ) . Interestingly , data indicated that parasitaemia in animal 6630 remained undetectable by microscopy after day 16 ( Fig 4 ) , when no further parasites were detected until infections were terminated at day 28 . However , 7SL-sRNA remained detectable , suggesting that this animal was suffering from a chronic stage of disease . Therefore , microscopy resulted in a false negative diagnosis but the RT-qPCR clearly remained sensitive , with a lower detection threshold than microscopy . However , the result could also indicate that the RNA is stable in the bloodstream and remains detectable after live parasites have been cleared . To investigate this further , we next focused our attention on animals undergoing treatment . Monitoring of disease progression is a vital aspect of treatment as well as for the development of optimised chemotherapeutics , which require their efficacy to be accurately measured during clinical trials . To this end , we used the 7SL-sRNA RT-qPCR assay on samples obtained from clinical trials performed on cattle experimentally infected with T . congolense ( Fig 5 ) or T . vivax ( Fig 6 ) . The objectives of this study were primarily to test how the assay would compare with other traditional measurements of disease progression such as microscopy , and to evaluate whether the 7SL-sRNA remains present in the bloodstream when an infection is cleared by chemotherapy . Importantly , trial animals were monitored for 85 days , allowing long-term follow-up sampling and analysis , and assessment of the utility of the 7SL-sRNA as a marker of active infection ( e . g . in the event of treatment failure ) . In all 21 cattle infected with T . congolense ( KONT 2/133 ) ( Fig 5; full data in S1 Fig ) , an initial wave of parasitaemia was observed by microscopy after ~5 days . Whilst no plasma samples were available to test between day 0 and day 8 , 7SL-sRNA was detected at the earliest post infection time-point available in all cattle . In these analyses data was normalised to a sample taken six to eight days preinfection . Upon experimental treatment of the cattle , there was a marked decrease in parasitaemia as determined by microscopy , which was mirrored by 7SL-sRNA detection assays carried out at the nearest time-points post-treatment ( Fig 5 ) . This observation was exemplified by animals 519 and 549 , where plasma sampled just one day after treatment was available for testing and there was no detectable 7SL-sRNA signal , as well as there being no detectable trypanosomes by microscopy ( Fig 5 ) . Indeed , 7SL-sRNA was rarely detected after treatment for the duration of the trial . Interestingly , in one case ( animal 549 ) , 7SL-sRNA was detected after 42 days , suggesting a relapse . Almost 15 days later , live trypanosomes were observed by microscopy , after which the infection was once again cleared ( Fig 5 ) . A similar phenomenon was observed in animals 526 and 528 , where no parasitaemia was detected by microscopy , again suggesting relapse of infection , detectable by RT-qPCR but not by microscopy . Importantly , these results indicate that 7SL-sRNA is short-lived in vivo , as successful drug treatment rapidly leads to the loss of signal , suggesting active infections are required to sustain the high abundance of the 7SL-sRNA . This further highlights the potential of 7SL-sRNA as a diagnostic for active trypanosome infections , rather than simply exposure such as is observed using antibody-based serological tests . Samples from a second clinical trial involving 21 cattle experimentally infected with T . vivax ( STIB 719 ) were also tested using the T . vivax-specific RT-qPCR assay ( Fig 6; full data in S2 Fig ) . As with the T . congolense study , parasitaemia was measured every 2–3 days , and plasma samples were obtained more sporadically ( approximately weekly ) over a period of 85 days post-infection . Treatment was administered after peak parasitaemia was observed by microscopy , typically after ~14 days . For the T . vivax trials , rescue treatment was administered if the trial compound failed . As demonstrated for T . congolense , 7SL-sRNA was detected by the first available time point , and once again mirrored parasitaemia observed by microscopy . In most cattle , T . vivax was cleared post-treatment ( exemplified by animals 496 , 511 , 515 , 520 , 522 , 532 , 544 , and 538 ) , and the 7SL-sRNA signal was absent at the next sampling timepoint ( typically 7–10 days later ) . However , in several cases , the presence of 7SL-sRNA was detected in time points where no parasites were observed by microscopy , again highlighting the increased sensitivity exhibited by the RT-qPCR compared to microscopy . Indeed , in these cases , such as animals 498 , 524 and 527 , 7SL-sRNA appeared to indicate cyclical changes in parasitaemia commonly associated with trypanosome infections ( Fig 6 ) . The above theory was further strengthened when investigating several animals that suffered from relapse of infection due to treatment failure , as confirmed by microscopy ( in particular , animals 502 , 543 and 550 ) ( Fig 6 ) . In animal 543 , 7SL-sRNA plasma levels increased after day 30 , without a corresponding increase in parasitaemia . Parasites were finally observed by microscopy on day 55 , more than 3 weeks after detection of relapse by 7SL-sRNA . Therefore , by RT-qPCR analysis of a highly abundant secretory/excretory small RNA , infection status was confirmed more accurately than by microscopy . Further analysis is required , ideally with time courses that include frequent sampling of host serum pre- and post-treatment , including at subtherapeutic doses , in order to accurately determine both the kinetics of decay of 7SL-RNA signal after successful treatment and the association between treatment failure , parasite dynamics and 7SL-RNA signal . Whilst our data suggests that the 7SL-sRNA presents a realistic target for development of a molecular diagnostic for both HAT and AAT , in reality for diagnostic assays to be field-applicable in the settings in which both diseases occur , an assay requires several attributes: high on this list are two related aspects—low cost and minimal number of processing steps . A two-step RT-qPCR involving a lengthy RNA extraction protocol is therefore not desirable . We therefore investigated whether a one-step RT-qPCR would simplify the assay and if the RNA could be detected directly from serum samples without the requirement for RNA extraction ( Fig 7 ) . Using RNA samples from T . brucei infected animal 6632 ( Fig 4 ) , we demonstrated that the one-step RT-qPCR detected 7SL-sRNA at every timepoint where the small RNA was detected by the two-step assay ( Fig 7A ) . Additionally , serum or plasma samples from both T . congolense and T . brucei-infected cattle were assayed using one-step RT-qPCR reactions , which were performed on 6 μL serum/plasma in a 15 μL reaction . For both T . brucei and T . congolense ( Fig 7B ) , 7SL-sRNA was readily detected . Finally , serum samples from a T . brucei in vivo infection time-course that included periods of patent and subpatent parasitaemia by microscopy ( animal 6630 ) were re-tested using the single step assay ( Fig 7C ) . Again the assay was able to clearly detect 7SL-sRNA , albeit with reduced sensitivity , during active infection even when infection was in remission and no parasitaemia could be detected by microscopy . These results demonstrate that assays for 7SL-sRNA are species-specific , highly sensitive , and can be detected the RNA before the onset of parasitaemia as well as during periods where there is subpatent parasitaemia by microscopy . Moreover , the 7SL-sRNA can also be detected directly from serum using a one-step RT-qPCR assay .
To meet the challenges of both elimination of HAT and management of AAT , improved diagnostic techniques are crucial . Current methods of diagnosis are suboptimal , particularly for AAT , and this is hindering progress on reducing the disease burden [3 , 4] . Traditionally diagnosis for both HAT and AAT relies upon microscopy—this depends on a minimum threshold of parasitaemia in order to accurately detect infection ( even with parasite concentration methods ) , and as our data illustrates there are often long periods of subpatent parasitaemia in trypanosome infections that is then problematic for microscopy diagnosis . In addition , the traditional reliance upon the presence of parasites in venous blood does not take into account recent reports of extravascular reservoirs of the parasite , such as the skin [37 , 38] and adipose tissue [39] ( which also seems to correlate with periods of subpatency in peripheral blood ) . Several potentially useful diagnostic approaches are sensitive and specific ( e . g . traditional PCR ) , but are expensive and not easily field-applicable , and DNA-based tests can face the issues of DNA still circulating post-treatment and the potential for easy cross-contamination . Finally , some antibody-based diagnostic tests are available , and are effective and field applicable , such as the CATT test for HAT and VerY Diag for AAT , but antibody tests have a challenge in differentiating between active infection and exposure ( additionally the CATT test only detects T . b . gambiense—there is currently no field-applicable test available for T . b . rhodesiense ) . Therefore , sensitive and specific markers of active infection are still required for both HAT and AAT . In this study we identified a species-specific small RNA excreted/secreted by all three AAT relevant species , as well as the main causative agent of HAT , T . b . gambiense , at physiologically relevant abundances both in vivo and in vitro . Crucially , the RNA , termed 7SL-sRNA , is indicative of active infection and is detected even when parasitaemia is below levels detectable by microscopy . Furthermore , successful drug treatment results in rapid loss of detectable 7SL-sRNA signal , thereby exhibiting a key characteristic of a marker that correlates with active infection . Whilst this study was able to utilise samples from multiple experimental animal trials , human blood or CSF samples that would be suitable for testing could not be identified , although we hypothesise the 7SL-sRNA would also be present at similar levels in these types of clinical samples in human patients–we tested culture supernatant from T . b . gambiense ( ELIANE ) grown in vitro , and the level of signal was the same as observed with T . b . brucei . Developing a diagnostic for HAT that is able to detect both T . b . gambiense and T . b . rhodesiense would be greatly beneficial , as currently there is no molecular test that is widely used in the field for T . b . rhodesiense diagnosis [3] . Whilst the 7SL-sRNA is detectable using laboratory PCR machines , these assays are clearly not field applicable in their current state , and further work must be carried out to adapt the assays ( e . g . to the loop mediated isothermal amplification ( LAMP ) platform; see below ) to be field applicable . However , there have also been considerable efforts to exploit differential host small RNA levels as biomarkers of disease states in human medicine , in particular miRNAs in cancer , which has led to several other potentially field-applicable diagnostic techniques that could be adapted to detect 7SL-sRNA . Several other technologies have the potential to be alternative platforms suitable for small RNA detection . In particular , the LAMP assay , which has been previously developed for all three livestock trypanosome species based on gDNA targets [13 , 40 , 41] and for which a test intended for field application was developed for HAT [14] , has been shown to be suitable for application to small RNA as an assay substrate [42 , 43] and therefore could potentially be optimised to develop a field-applicable 7SL-sRNA assay . Another recently developed method , Recombinase Polymerase Amplification ( e . g . [44] ) , exhibits potential as an extremely sensitive detection method , even surpassing the aforementioned LAMP assay . This process requires a reaction consisting of a recombinase , a single-stranded DNA-binding protein and a strand-displacing polymerase to amplify the target [45] . Importantly , this method can be carried out at low temperatures , and amplification has been shown to proceed using just body heat [46] . By coupling this assay to a lateral flow device or a dipstick using biotinylated primers , the target can be visualised by eye , thereby bypassing the need for thermocycler or real-time fluorescence detection . Whilst RPA typically requires targets consisting of >30 bp , this technology has recently been adapted to the detection of miRNAs by ligating highly specific probes to the miRNA using a PBCV-1 ligase [47] . The discovery of a small RNA secreted/excreted by African trypanosomes at high abundance also raises interesting and potentially important biological questions . There is currently a great deal of interest in the ability of pathogens to communicate with each other and to manipulate host functions through the delivery of small RNAs [48] . Communication between trypanosomes during infection is required to regulate differentiation in a population density-dependant manner [49] . Trypanosome infection also has substantial effects on host cells , for example the ablation of B cells and consequent loss of immune memory through an as yet undefined ligand [50] . A small RNA , such as 7SL-sRNA could directly regulate gene expression , as is the case with miRNAs , or act as a signalling molecule , potentially triggering or inhibiting immunoregulatory pathways in host cells . Studies have demonstrated that small RNAs are often packaged , secreted and delivered to target cells via extracellular vesicles such as exosomes . The 7SL RNA has been detected in exosomes from other trypanosomatids including Leishmania spp . [51] and Trypanosoma cruzi , a closely related pathogen that causes Chagas’ disease in South America [52] . However , in these datasets , the significance of this finding and , indeed , whether the entire 7SL RNA or just a portion of it were observed , were not discussed . Studies are currently underway to determine whether 7SL-sRNA is released from the parasite in a vesicle or exists freely . Furthermore , it is yet to be determined whether 7SL-sRNA exists as part of a larger complex . Its stability in serum would suggest the RNA is somehow protected from RNase activity . In addition to the potential biological role of 7SL-sRNA , how the small RNA is processed and released from trypanosomes is yet to be determined . The existence of a potential passenger strand could suggest a role for DICER in processing the larger 7SL RNA , as this type III endonuclease has previously been shown to process mammalian 7SL RNA [53] . Intriguingly , previous studies have noted the absence of a conserved eukaryotic SRP complex protein in a related trypanosomatid species . Instead , isolation of the 7SL RNA revealed a co-migratory tRNA-like molecule ( sRNA-85 in Leptomonas collosoma [54] , sRNA-76 in T . brucei [55] ) . The tRNA-like molecule has extensive and precise complementarity to the region of 7SL RNA that is processed to generate the small RNA . Finally , while our data is consistent with the 7SL-sRNA being actively processed and secreted/excreted , we cannot currently formally rule out that parasite cell death or membrane damage may be contributing to the 7SL-sRNA signal . In summary , we have detected a trypanosome small RNA ( 7SL-sRNA ) , derived from the non-coding 7SL RNA of the SRP , which is excreted/secreted at high levels by T . brucei , T . congolense and T . vivax during infections . Species-specific RT-qPCR assays were developed , and we have shown that there is good correlation between 7SL-sRNA levels and parasite numbers , but importantly 7SL-sRNA can be detected both before patent parasitaemia and during phases of infection when parasitaemia becomes subpatent ( both chronic infection and treatment failure ) , and critically the 7SL-sRNA signal decays rapidly after successful chemotherapy . Therefore , 7SL-sRNA represents a marker of active infection , and is a novel and viable target for the development of much needed diagnostics for both HAT and AAT , and may also provide insights into important host-pathogen interactions . | African trypanosomes cause significant disease in humans and animals across sub-Saharan Africa . For both human and animal infections diagnostics that can accurately identify an active infection are lacking–this is particularly the case in animal disease where most diagnosis is based upon clinical signs , which is not a specific or sensitive means of detecting infection . There is therefore a significant unmet need for a pathogen marker of active infection that accurately indicates whether an animal or human is currently infected . Through analysing the blood of cattle infected with trypanosomes , we identified a short sequence of RNA that was present at very high levels . This small RNA derives from the trypanosome genome , and we could identify its presence in the genome of all three species that are responsible for human and animal disease . We were able to design species-specific tests , and showed that in samples from infected animals the assays were more sensitive than the traditional microscope-based detection , importantly the signal disappeared relatively quickly after successful treatment , and when treatment failed , the assay was able to accurately identify when infection persisted . We also demonstrated that the causative agent of human trypanosomiasis secretes the marker at similar levels to that seen in the animal-infective trypanosomes . Therefore , we have discovered a marker of trypanosome infection that is present at high levels in the blood of infected animals , disappears quickly upon successful treatment , but is effective at detecting instances of unsuccessful treatment and persistent infection . This represents a potentially powerful diagnostic tool for human and animal trypanosomiasis . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
]
| [
"medicine",
"and",
"health",
"sciences",
"ruminants",
"trypanosoma",
"congolense",
"rna",
"extraction",
"vertebrates",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"mammals",
"animals",
"parasitology",
"trypanosoma",
"brucei",
"protozoans",
"extraction",
"techniques",
"veterinary",
"science",
"research",
"and",
"analysis",
"methods",
"veterinary",
"parasitology",
"trypanosoma",
"eukaryota",
"cattle",
"diagnostic",
"medicine",
"biology",
"and",
"life",
"sciences",
"trypanosoma",
"brucei",
"gambiense",
"amniotes",
"bovines",
"organisms"
]
| 2019 | Parasite specific 7SL-derived small RNA is an effective target for diagnosis of active trypanosomiasis infection |
Excitatory synapses on mammalian principal neurons are typically formed onto dendritic spines , which consist of a bulbous head separated from the parent dendrite by a thin neck . Although activation of voltage-gated channels in the spine and stimulus-evoked constriction of the spine neck can influence synaptic signals , the contribution of electrical filtering by the spine neck to basal synaptic transmission is largely unknown . Here we use spine and dendrite calcium ( Ca ) imaging combined with 2-photon laser photolysis of caged glutamate to assess the impact of electrical filtering imposed by the spine morphology on synaptic Ca transients . We find that in apical spines of CA1 hippocampal neurons , the spine neck creates a barrier to the propagation of current , which causes a voltage drop and results in spatially inhomogeneous activation of voltage-gated Ca channels ( VGCCs ) on a micron length scale . Furthermore , AMPA and NMDA-type glutamate receptors ( AMPARs and NMDARs , respectively ) that are colocalized on individual spine heads interact to produce two kinetically and mechanistically distinct phases of synaptically evoked Ca influx . Rapid depolarization of the spine triggers a brief and large Ca current whose amplitude is regulated in a graded manner by the number of open AMPARs and whose duration is terminated by the opening of small conductance Ca-activated potassium ( SK ) channels . A slower phase of Ca influx is independent of AMPAR opening and is determined by the number of open NMDARs and the post-stimulus potential in the spine . Biphasic synaptic Ca influx only occurs when AMPARs and NMDARs are coactive within an individual spine . These results demonstrate that the morphology of dendritic spines endows associated synapses with specialized modes of signaling and permits the graded and independent control of multiple phases of synaptic Ca influx .
Various theories have been proposed to explain why excitatory synapses of principal neurons are formed onto the heads of dendritic spines . These include proposals that spines increase the maximal density of contacts onto a linear dendrite [1] , [2] , that the long spine increases the repertoire of possible presynaptic partners for each dendrite [3] , and that the thin spine neck diffusionally isolates the associated postsynaptic terminal in order to allow the synapse-specific induction and expression of plasticity [4] . A further proposal that has been controversial is that the thin neck creates an electrical resistance that , in conjunction with the membrane capacitance , significantly filters synaptic potentials [4]–[7] . For example , a high spine neck resistance would allow for the generation of large synaptic potentials that are limited to the active spine head and result in differential activation of voltage-sensitive ion channels on either side of the neck [8]–[10] . Based on its morphology and diffusional properties , the electrical resistance of the neck of most dendritic spines in their basal state has been estimated to be <100 MΩ , too low to create a large voltage drop across the neck in response to typical synaptic currents as measured at the soma [1] , [2] , [11] . These estimates depend on an approximation of the resistivity of the intracellular space within the neck , which given the thinness of the structure , its high protein content , and our poor understanding of its biochemistry , may differ substantially from that of other portions of the cell . Furthermore , biophysical constraints of charge flow through a narrow juxtamembrane space may not be revealed by studies of diffusion of uncharged molecules . Lastly , current flow out of the neck into the dendrite may be attenuated even if the neck resistance is low due to opening of ion channels or transporters located selectively in the spine neck . Several recent studies have suggested that electrical filtering by the spine neck may be substantial and may be regulated by activity . First , voltage-sensitive dye and second harmonic generation imaging have suggested shunting of synaptic currents at long-neck spines [5] . Similarly , glutamate uncaging at individual spines revealed that stimulation of spines with longer necks produces smaller potentials measured at the soma [5] , [12] , [13] . However , the lack of independent quantification of the number of AMPARs on these spines renders the interpretation of the data in these studies difficult . Second , diffusional equilibration across the spine neck is retarded following paired pre- and postsynaptic stimulation [14] or prolonged postsynaptic depolarization [9] . In such post-constriction spines the spine neck resistance may approach 1 GΩ , opening the possibility that activity-dependent regulation of spine/dendrite coupling is a mechanism for dynamic control of synaptic signals . Indeed , after apparent constriction of the spine neck by depolarization to 0 mV , the contribution of voltage-gated ion channels to synaptic Ca signaling is large; however , no analysis of synapses in their basal state was performed [9] . Lastly , a separate study of spines in their basal state found that VGCCs in the spine head are activated by unitary synaptic stimuli , suggesting the existence of a large synaptic depolarization within most spines . In particular , VGCCs blocked by SNX-482 are located in the spine head and open during synaptic stimuli , both contributing to synaptic Ca influx and leading to the activation of SK channels [14] . However , SNX-482-sensitive VGCCs are not found in the dendritic shaft , and therefore their selective activation in the spine head may reflect their inhomogeneous distribution and not compartmentalization of synaptic potentials . Here we examine if the segregation of excitatory synapses onto the heads of dendritic spines endows the synapse with specialized modes of regulation and signaling that would not be available to synapses formed directly onto the dendritic shaft . Since quantifying the spine neck resistance and the voltage drop across the neck depends on difficult-to-measure cellular parameters , we avoid performing these estimates and instead focus on uncovering functional evidence that synaptic signals are influenced by the electrical and morphological properties of the spine . We find that , for spines in their basal state , the voltage drop across the spine neck during synaptic stimulation is sufficient to allow for differential activation of the voltage-sensitive ion channels in the spine and dendrite . Furthermore , depolarization reached in the spine during synaptic activity creates a large and brief component of Ca influx that is substantially accelerated compared to the typical slow kinetics of NMDAR signaling . The rapid and slow components of synaptic Ca influx arise through distinct mechanisms and are differentially affected by modulation of AMPARs and SK channels . Lastly , the biphasic nature of Ca influx arises directly from the colocalization of AMPARs and NMDARs on the spine head and cannot be recapitulated by stimulation in the adjacent dendrite .
A resistor network model of current flow in the spine and along the dendrite to the soma was used to predict the consequences of spine versus dendrite stimulation ( Figure 1B ) . In this simple model , active AMPARs are represented as current sources and capacitance has been ignored . Current through open AMPARs enters the spine head ( IH ) and flows through both the spine head membrane ( RH ) and the spine neck ( RN ) resistors . The resistance of the neck creates a drop in voltage such that the voltage in the spine ( VH ) is greater than that in the dendrite ( VD ) . Membrane and axial resistances of the dendrite ( RM and RA , respectively ) further reduce the voltage at the soma ( VSoma ) such that it is less than VD . With current injection into the dendrite ( ID ) , current will flow along the same path to the soma but in the opposite direction across the neck , thus maintaining the same relationship between VD and VSoma but resulting in VH<VD . If current is injected into the dendrite such that VSoma is equal to that evoked by current injection into the spine , then the voltage profile from the base of the spine to the soma is the same in the two conditions . Therefore , differences observed in spine head VGCC activation following these two stimuli can be attributed to differences in the voltage between the base of the spine and the spine head and would indicate that the spine neck resistance RN is sufficient to support a significant voltage drop . In order to detect if synaptic stimuli lead to differential activation of VGCCs on either side of the spine neck , we measured glutamate uncaging-evoked Ca transients in the spine head ( Δ[Ca]spine ) and neighboring dendrite ( Δ[Ca]den ) while stimulating either the spine ( Figure 1C and 1D ) or the dendrite ( Figure 1E and 1F ) . In this and all portions of the study , the uncaging laser power delivered to each spine was set in order to mimic the kinetics and amplitude of glutamate receptor activation following vesicular release using a photobleaching calibration strategy [15] . This stimulus results in ∼10–13 pA excitatory post-synaptic currents ( EPSCs ) and ∼0 . 8–1 mV potentials ( EPSPs ) , consistent with the size of potentials generated by single active synapses associated with large-head spines studied here [16] and at the large end of the distribution of responses from all synapses [17] , [18] formed onto proximal portions of apical dendrites of CA1 pyramidal neurons . When stimulating the dendrite , laser power was adjusted to evoke a similarly sized uncaging-evoked EPSP ( uEPSP ) as seen when the spine was stimulated directly . To isolate Ca influx through VGCCs , all glutamate-activated Ca sources were blocked using a combination of NMDAR , metabotropic glutamate receptor ( mGluR ) , Ca-permeable AMPAR , and kainate receptor antagonists ( CPP , MK801 , MPEP , CPCCOEt , Joro spider toxin , NASPM , UBP302 , respectively ) . SK and voltage-gated sodium channels were also blocked ( with apamin and tetrodotoxin , respectively ) to prevent nonlinear effects of these channels on synaptic signaling [15] , [19] , [20] . In these conditions , direct activation of the spine triggers a uEPSP as well as a rapidly rising and short-lived Δ[Ca]spine ( Figure 1C and 1D ) . In contrast , stimulation of the dendrite near the base of the spine elicits a similarly sized uEPSP via activation of extrasynaptic AMPARs yet it produces negligible Δ[Ca]spine ( Figure 1E and 1F ) . In this example neither stimulus produces significant Δ[Ca]den , although the magnitude of Δ[Ca]den was variable across experiments . On a spine-by-spine basis , stimulation of the dendrite reliably evoked Δ[Ca]den , but only in one case was able to evoke substantial Δ[Ca]spine compared to stimulating the spine head directly ( Figure 1G ) . In two cases stimulation of the dendrite produced a sizable local Ca increase , possibly reflecting the presence of a small spine on the dendrite oriented along the optical axis of the microscope and thus beyond our ability to visually resolve . On average , dendrite and spine stimulation evoked equal amplitude uEPSPs as measured at the soma ( 0 . 99±0 . 05 mV , n = 15 , and 0 . 91±0 . 07 mV , n = 10 , respectively ) , whereas spine stimulation evoked ∼5-fold larger Δ[Ca]spine than dendritic stimulation ( ΔG/Gsat: 15 . 7%±2 . 9% and 3 . 3%±2 . 1% , respectively ) ( Figure 2B ) . Both stimuli produced only small Δ[Ca]den ( ΔG/Gsat: 0 . 9%±0 . 3% and 3 . 9%±0 . 8% for stimulation of the spine and dendrite , respectively ) ( Figure 2B ) . Thus , dendritic stimulation is unable to activate voltage-gated Ca sources located in the spine head . To confirm that Δ[Ca]spine was mediated by VGCC activation , experiments were repeated in the additional presence of a cocktail of VGCC antagonists ( nimodipine , ω-conotoxin-MVIIC , SNX-482 , mibefradil , and nickel ) . Addition of these antagonists reduced Δ[Ca]spine ( ΔG/Gsat: 8 . 2%±0 . 9% and 2 . 9%±1 . 1% , n = 7 for spine and dendrite stimulation , respectively ) and Δ[Ca]den ( ΔG/Gsat: 1 . 0%±0 . 3% and 2 . 2%±0 . 5% ) without altering the uEPSP ( 0 . 90±0 . 05 mV and 0 . 95±0 . 10 mV ) ( Figure 2B ) . VGCCs resistant to this combination of antagonists have been reported and likely contribute the remainder of Δ[Ca]spine and Δ[Ca]den [21]–[23] . These results indicate the existence of a voltage drop across the spine neck that is sufficient to spatially compartmentalize activation of voltage-gated ion channels over a micron length scale . In order to determine if the depolarization reached in the spine during synaptic activity shapes synaptic signals , we examined if the opening of AMPARs , which provide the bulk of current influx that produces the EPSP , secondarily alters the magnitude or kinetics of Ca current into the spine ( iCa ) ( Figure 3 ) . Previous studies have reported variable effects of blocking AMPARs on the peak of synaptically evoked Ca transients [9] , [24] , [25] . However , in fluorescence imaging of Ca transients , the time course of iCa is obscured by the presence of the exogenous Ca indicator , which slows the clearance of Ca from the spine . In the absence of exogenous buffer , the low ( ∼25 ) endogenous Ca buffer capacity of the apical spines of CA1 pyramidal neurons allows spine head Ca to closely follow the kinetics of opening of Ca sources [26] . To determine the time course of iCa , we corrected for the kinetics of Ca handling by performing a deconvolution with the impulse response of Ca handling of the spine ( see methods ) [26] , [27] . The impulse response was estimated from the decay of the synaptically evoked fluorescence transient measured in the presence of NMDAR antagonists , which is generated by Ca influx that is impulse-like relative to the decay kinetics of Δ[Ca]spine . Blockade of NMDARs with CPP/MK801 had no significant effect on uEPSP amplitudes ( 1 . 04±0 . 19 mV , n = 18 , and 1 . 23±0 . 23 mV , n = 17 , control and CPP/MK801 , respectively ) but reduced the early phase ( 20–50 ms post-uncaging ) of Δ[Ca]spine ∼60% ( early ΔG/Gsat: 9 . 48%±1 . 12% and 4 . 09%±0 . 62% in control and CPP/MK801; p<0 . 05 ) while nearly eliminating the later portion ( 50–120 ms post-uncaging ) ( late ΔG/Gsat: 8 . 48%±1 . 3% and 0 . 90%±0 . 16%; p<0 . 05 ) ( Figure 3A ) . These results confirm the dominant role of Ca influx through NMDARs in generating synaptically evoked Ca transients and , in particular , in mediating the prolonged phases of synaptic Ca influx . The remaining spine Ca transient was well described by a single exponential decay with a time constant of ∼42 ms , which was used as the Ca impulse response in deconvolution analysis below . Blockade of AMPARs with NBQX largely eliminated the uEPSP ( 0 . 23±0 . 09 mV , n = 11 , p<0 . 05 versus control ) ( Figure 3C ) and decreased the initial portion of Δ[Ca]spine ( early ΔG/Gsat: 7 . 24%±1 . 5% , p<0 . 05 versus control ) without significant effect on the prolonged phase ( late ΔG/Gsat: 7 . 99%±1 . 70% ) . Deconvolution of the spine head fluorescence transients in control conditions and in the presence of NBQX reveals a large rapid phase of Ca influx that lasts ∼10 ms and that is eliminated by AMPAR blockade ( Figure 3C ) . A similar rapid and AMPAR-dependent phase of Ca influx is also evoked by briefer glutamate uncaging pulses ( 300 µs ) that elicit smaller uEPSPs ( 0 . 42±0 . 06 mV; n = 15 ) ( Figure S1 ) . These results indicate that , in addition to providing the depolarization that underlies the EPSP , AMPAR activation transiently boosts a rapidly activating Ca source in the spine that dominates the early phase of synaptic Ca transients . In a converse set of experiments , AMPAR opening was enhanced using cyclothiazide ( CTZ ) ( Figure 3D ) , which prevents AMPAR desensitization and increases the affinity of the receptor for glutamate [28] . We found that low levels of CTZ ( 2 . 5 and 5 µM ) increased the amplitude of the uEPSP in a graded manner ( Figure 3D , 3F , and 3G ) ( uEPSP: 1 . 80±0 . 23 mV , n = 17 , and 3 . 04±0 . 34 mV , n = 7 , for 2 . 5 and 5 µM , respectively ) . In addition , CTZ increased the amplitude of the early phase of Δ[Ca]spine ( early ΔG/Gsat: 11 . 2%±3 . 0% and 16 . 91%±1 . 89% , respectively ) with no significant effect on the prolonged phases ( late ΔG/Gsat: 8 . 2%±2 . 3% and 11 . 27%±1 . 33% ) . Deconvolution analysis revealed that CTZ selectively boosted the amplitude of the rapid phase of iCa , consistent with a modulation of spine potential and Ca influx during this period of depolarization . The prolonged phase of Ca influx is independent of AMPAR opening and is consistent with Ca influx through NMDARs in a spine that has returned to resting potentials . The lack of modulation of this phase is an independent confirmation that the degree of glutamate uncaging did not vary across conditions . If the rapid phase of iCa reflects Ca influx during a short-lived depolarization in the spine , its magnitude or duration should be increased by manipulations that slow the repolarization of the spine following synaptic stimuli . We have previously described that SK channels present in the spine are activated by CaV2 . 3 type VGCCs and act to negatively regulate synaptic Ca influx [15] . Here we find that application of the SK antagonist apamin boosted the amplitude of the uEPSP ( 2 . 01±0 . 43 mV , p<0 . 05 versus control ) in a manner similar to CTZ but , in contrast to CTZ , increased both the early and sustained phases of Δ[Ca]spine ( early and late ΔG/Gsat: 20 . 18%±2 . 53% and 16 . 59%±2 . 23% , respectively; p<0 . 05 for each versus control ) ( Figure 3E ) . Deconvolution revealed a larger and more prolonged rapid iCa as well as an increase in its late phase . Although SK and AMPAR modulation both proportionally regulate the peak amplitudes of iCa and the uEPSP , only apamin increases the amplitude of the prolonged phase of iCa ( Figure 3F ) . These data confirm that the rapid phase of iCa is not due to direct Ca influx through AMPARs since increasing spine depolarization without altering the number of open AMPARs reduces the driving force for Ca influx and would be predicted to decrease Ca influx through these receptors . These results are consistent with SK channels opening rapidly following synaptic stimulation and repolarizing the spine , thereby directly truncating the synaptic potential in the spine , indirectly terminating the rapid phase of synaptic Ca influx , and reducing the prolonged phase of Ca entry through NMDARs [15] , [20] . Thus , AMPARs and SK channels , which modulate the amplitude and kinetics of the synaptic potential in the spine , also regulate the kinetics of synaptically evoked Ca currents and spine Ca transients , indicating a functional role of spine depolarization in shaping synaptic signals . In order to understand if the graded modulation of synaptic Ca influx by AMPAR opening is made possible by the electrical properties of the spine neck , we examined if the opening of dendritic AMPARs located at the base of the spine is also able to enhance the rapid phase of iCa ( Figure 4 ) . For this analysis , each spine was stimulated with four different spatiotemporal patterns of glutamate uncaging designed to mimic ( 1 ) normal synaptic activation of AMPARs and NMDARs on the spine head , ( 2 ) activation of only NMDARs on the spine head , ( 3 ) activation of only AMPARs on the neighboring dendrite , and ( 4 ) near simultaneous activation of spine head NMDARs and dendritic AMPARs . As expected , stimulation of the spine in order to locally open both AMPARs and NMDARs resulted in a robust uEPSP ( 1 . 1±0 . 13 mV , n = 11 ) and Δ[Ca]spine ( early and late ΔG/Gsat: 8 . 28%±0 . 64% and 6 . 31%±0 . 58% ) ( Figure 4A ) . In order to separate AMPAR and NMDAR activation , we took advantage of the high affinity of NMDARs for glutamate [29] and the existence of extrasynaptic dendritic AMPARs [30]–[32] . Glutamate uncaging at a spot located ∼1 . 5–3 microns from the spine head for 2 ms releases glutamate that activates spine NMDARs but , because of the rapid fall of glutamate concentration from the uncaging location , is insufficient to activate AMPARs . Consistent with preferential activation of spine NMDARs , this stimulus results in a small uEPSP ( 0 . 40±0 . 05 mV , n = 11 ) and Δ[Ca]spine ( early and late ΔG/Gsat: 5 . 85%±0 . 40% and 5 . 07%±0 . 40% , respectively ) ( Figure 4A ) that is similar to that measured in the presence of AMPAR antagonists . Furthermore , in a separate set of experiments , the glutamate receptor dependence of the spine head calcium transient generated by this stimulus ( early and late ΔG/Gsat: 5 . 05%±0 . 63% and 4 . 54%±0 . 55% , respectively; n = 8 ) was analyzed and found to be blocked by CPP ( early and late ΔG/Gsat: 0 . 35%±0 . 29% and 0 . 48%±0 . 23% , respectively; n = 6 ) and insensitive to NBQX ( early and late ΔG/Gsat: 5 . 20%±0 . 90% and 4 . 17%±0 . 84% , respectively; n = 9 ) , confirming that it is mediated by NMDARs with little influence from AMPARs . Although this stimulus also generates a small uEPSP , it is referred to here as the “NMDAR only” stimulus for convenience . In order to supply AMPAR-dependent depolarization without direct stimulation of spine NMDARs , we applied a stimulus to the base of the spine as in Figure 1 that activates dendritic AMPARs and produces a uEPSP of similar amplitude ( 1 . 19±0 . 06 mV ) to direct stimulation of the spine but only minimal Δ[Ca]spine ( early and late ΔG/Gsat: 0 . 62%±0 . 13% and 0 . 87%±0 . 18% , respectively ) ( Figure 4B ) . Failure of this “AMPAR only” stimulus to induce Ca influx into the spine head is due to both a failure of the depolarization to activate spine head VGCCs and a failure of glutamate to diffuse to and significantly activate spine NMDARs ( Figure 4B ) . Lastly , we paired the AMPAR-only and NMDAR-only stimuli and compared these to results obtained when both channel types are spatially co-activated by direct stimulation of the spine . We find that the temporally paired but spatially separated activation of AMPARs and NMDARs results in a uEPSP ( 1 . 57±0 . 15 mV ) and Δ[Ca]spine ( early and late ΔG/Gsat: 5 . 93%±0 . 38% and 5 . 80%±0 . 44% , respectively ) that are predicted by the linear sums of the NMDAR-only and AMPAR-only uEPSP and Δ[Ca]spine ( Figure 4C ) . However , despite the larger amplitude of the uEPSP generated by this paired stimulus , it failed to enhance the rapid phase of Ca influx beyond that seen with NMDAR-only stimulation . Deconvolution analysis reveals that the AMPAR-only , NMDAR-only , and paired stimulus all failed to generate the spike-like rapid phase of Ca influx seen with direct stimulation of the spine . Thus , open AMPARs must be colocalized with NMDARs on the spine in order to enhance and accelerate synaptic Ca influx .
We have demonstrated that synaptic depolarization is sufficient to activate voltage-sensitive Ca sources located on the spine but that an EPSP-like dendritic depolarization is not . Previous studies have shown that synaptically evoked Ca influx is limited to the active spine but were performed in conditions in which the contribution of VGCCs to Ca influx was obscured [33] , [34] . These studies had been designed to measure NMDAR-mediated Ca influx and either held the neuron at 0 mV to intentionally prevent stimulus-evoked VGCC activation , or imaged using a high affinity Ca buffer , which enhances the contributions of long-lived Ca sources such as the NMDAR relative to those of short-lived Ca sources such as VGCCs . Our study demonstrates that under conditions in which Ca influx through spine VGCCs can be measured , their activation only occurs when they are located on the active spine . This indicates that VGCCs and other voltage-gated channels in one spine are unlikely to be opened by synaptic activity at neighboring spines or subthreshold depolarizations in the dendrite . Since specializations of Ca handling prevent diffusion of Ca from active to inactive spines [26] and electrical filtering prevents the spread of VGCC-activating potentials , signaling cascades triggered by synaptic activation of VGCCs are limited to the active spine . Thus , synaptic cross-talk among neighboring spines is likely to be predominately biochemical in nature and downstream of Ca [35] , [36] . It is important to note that these conclusions do not apply to stimuli that trigger local or back-propagating dendritic action potentials , which reliably invade active and inactive spines [17] , [21] , [34] , [37] , [38] . In addition , we find that the existence of a large depolarization within the active spine transiently couples AMPAR activation and synaptic Ca sources , generating a fast , spike-like phase of Ca influx . This rapid phase is temporally and mechanistically distinct from the prolonged phase of NMDAR-dependent Ca influx . Since the activation of biochemical events downstream of Ca depend on both the kinetics and amplitude of the Ca transients , independent modulation of these parameters may fine-tune the activation of Ca-dependent signaling pathways . Furthermore , differential regulation of each phase may permit subtle and nuanced regulation of synaptic signaling cascades , such as those underlying spike-timing dependent potentiation and depression . The current study was performed using combined 2PLSM and 2PLP , which allowed delivery of controlled stimuli to visually identified portions of the dendritic arbor [17] , [35] , [39] , [40] . This approach also allows the examination of postsynaptic signaling in conditions , such as blockade of voltage-gated Na and Ca channels , which prevent action potential-evoked release of neurotransmitter from the presynaptic terminal . When using glutamate uncaging , an important parameter that must be carefully considered is the strength at which individual postsynaptic terminals are stimulated and how the amount of glutamate released compares to that contained in a single vesicle . Here we calibrate uncaging laser power delivered to each spine using a photo-bleaching strategy [15] . This approach delivers an amount of glutamate that , on average and under conditions of excellent space-clamp , produces a uncaged-evoked EPSCs of ∼10–13 pA in amplitude , similar to that previously reported for spontaneous miniature EPSCs in hippocampal pyramidal neurons [18] , [41] , [42] . In current clamped neurons , this same stimulus results in ∼0 . 8–1 mV uEPSPs ( see Figure 2 ) [15] . This amplitude is similar to that of unitary EPSPs in 2- to 4-wk-old rat tissue evoked by stimulation of individual synapses formed onto large dendritic spines similar to those studied here [16] . In contrast , a study of adult ( 6–10 wk ) rat hippocampal CA1 pyramidal neurons found that single vesicle EPSPs recorded from proximal dendrites and measured in Sr2+ or following hypertonic sucrose puffs averaged 0 . 2 mV in amplitude and ranged from 0 . 1 to 1 . 5 mV while EPSCs ranged from a few to tens of picoamps [18] . The large amplitude of uEPSPs in our hands despite the relatively small size of uncaged-evoked EPSCs may result from the smaller size and hence larger input resistance of juvenile mouse CA1 pyramidal neurons . In addition , in this study we selected large-head spines , which have higher AMPA receptor content compared to the average of all synapses found on the apical dendrite and are thus expected to generate larger potentials [43] , [44] . In order to ensure that our results are not due to overly strong stimulation , we analyzed the kinetics and NBQX sensitivity of spine head Ca transients evoked by smaller depolarizations ( ∼0 . 4 mV ) evoked by uncaging for 300 µs . This smaller stimulus also produced biphasic calcium influx with a rapid phase that was blocked by antagonists of AMPARs ( Figure S1 ) . Taken together , these points suggest that while the uncaging stimulus is at the upper end of the physiologically relevant range , biphasic Ca influx occurs over a spectrum of relevant stimulus strengths . However , one cannot fully discard the possibility that the spatiotemporal pattern of glutamate release following uncaging enhances effects that contribute less robustly during release of glutamate from the presynaptic terminal . Our results predict that changes in the numbers of AMPARs at the synapse over the course of development or following the induction of many forms of plasticity will not only determine the amplitude of the synaptic potential but will also directly alter the synaptic Ca transient . Finally , the morphological features of the spine , specifically the dimensions of the spine neck , can restrict the movement of molecules [11] , [14] , [45] and attenuate electrical signals exchanged between the spine head and dendrite [5] . Thus , identification of molecules involved in regulating the length and , in particular , the diameter of the spine neck will be important future avenues of study .
Animals were handled according to protocols that have been approved by the Harvard Standing Committee on Animal Care and are in accordance with National Institutes of Health guidelines . Transverse hippocampal slices were prepared from C57/Blk6 mice from P15–P18 as described previously [15] . Animals were anesthetized by inhalation of isoflurane . The cerebral hemispheres were quickly removed and placed into cold choline-based artificial cerebrospinal fluid ( choline-ACSF ) containing 25 mM NaHCO3 , 1 . 25 mM NaH2PO4 , 2 . 5 mM KCl , 7 mM MgCl2 , 25 mM glucose , 1 mM CaCl2 , 110 mM choline chloride , 11 . 60 mM ascorbic acid , and 3 . 10 mM pyruvic acid , and equilibrated with 95%O2/5%CO2 . Tissue was blocked and transferred into a slicing chamber containing choline-ACSF . Transverse hippocampal slices ( 300 µm ) were then cut with a Leica VT1000S ( Leica Instruments , Nussloch , Germany ) and transferred into a holding chamber containing ACSF consisting of ( in mM ) 127 NaCl , 2 . 5 KCl , 25 NaHCO3 , 1 . 25 NaH2PO4 , 2 . 0 CaCl2 , 1 . 0 MgCl2 , and 25 glucose , equilibrated with 95%O2/5% CO2 . Slices were incubated at 32°C for 30–45 min and then left at room temperature until recordings were performed . Whole-cell recordings were made from CA1 pyramidal neurons visualized under infrared differential interference contrast microscopy ( IR-DIC ) . Patch pipettes ( open pipette resistance 2 . 5–4 . 5 MΩ ) were filled with an internal solution containing ( in mM ) 140 KMeSO4 , 8 NaCl , 1 MgCl2 , 10 HEPES , 5 MgATP , and 0 . 4 Na2GTP ( pH 7 . 3 ) . In the first set of experiments ( Figures 1 and 2 ) , 150 µM Fluo-5F ( Molecular Probes , KD∼1 . 1 µM ) and 5 µM Alexa Fluor-594 were included in the internal solution; in the remaining experiments ( Figures 3 and 4 ) , 300 µM Fluo-5F and 10 µM Alexa Fluor-594 were used . Recordings were made with an Axoclamp 200B or Multiclamp 700B amplifier ( Axon Instruments , Union City , CA , USA ) . Data were filtered at 5 kHz and sampled at 10 kHZ . Cells were held at −65 mV in voltage-clamp mode , and DC current was injected to hold cells at approximately −65 mV in current-clamp mode . Cells were rejected if holding currents exceed −50 pA . Series and input resistance were measured throughout the experiment , and recordings were discarded if series resistance exceeded 20 MΩ . All recordings were done at 32°C and within 7 h of slice preparation . D-serine was included in the ACSF in all recordings to reduce NMDAR desensitization . Pharmacological agents were used at the following final concentrations as indicated in the text ( in µM ) : 10 D-serine ( Sigma-Aldrich , St . Louis , MO , USA ) , 2 . 5 or 5 CTZ ( Tocris Biosciences , Ellisville , MO , USA ) , 0 . 1 apamin ( Calbiochem , La Jolla , CA , USA ) , 20 CPP ( Tocris ) , 40 MK-801 ( Tocris ) , 10 NBQX ( Tocris ) , 25 UBP302 ( Tocris ) , 10 NASPM ( Tocris ) , 0 . 5 Joro spider toxin ( Sigma-Aldrich ) , 1 MPEP ( Tocris ) , 100 CPCCOEt ( Tocris ) , 1 tetrodotoxin ( Tocris ) , 20 nimodipine ( Sigma-Aldrich ) , 1 ω-conotoxin-MVIIC ( Peptides International , Louisville , KY , USA ) , 0 . 3 SNX-482 ( Peptides International ) , 10 mibefradil ( Sigma-Aldrich ) , and 50 nickel ( Sigma-Aldrich ) . Ca-permeable AMPARs are expressed in synapses of CA1 pyramidal neurons only in a narrow time window following LTP induction [46] or homeostatic plasticity [47] , [48] . Nevertheless , for the analysis in Figure 1 we included antagonists that we previously have shown block Ca influx through these channels in spines of other neuron classes [39] , [49] . Custom built 2-photon laser scanning microscopes based on BX51Wl microscopes ( Olympus ) were used as described previously [39] . Ti-sapphire lasers ( Mira/Verdi , Coherent ) tuned to 840 and 725 nm were used for imaging and glutamate uncaging , respectively . In all uncaging experiments , 3 . 75 mM MNI-glutamate ( Tocris ) was included in a small volume ( ∼9 ml ) of re-circulating ASCF . Unless otherwise stated in the text , uncaging laser pulse duration was 0 . 5 ms and power delivered to each spine was standardized to bleach ∼50% of the red fluorescence in the spine head as described [15] . This corresponded to uncaging laser power as measured at the focal plane of the back aperture of the objective of ∼35 mW , similar to that used in other 2P-uncaging studies [36] . Image and electrophysiology acquisition was controlled by custom software written in MATLAB ( Mathworks ) . Cells were filled with two fluorescent dyes: a Ca-insensitive fluorophore ( Alexa Fluor-594 ) , which fluoresces in the red ( collected via 600–660 interference filter ) , and a Ca-sensitive fluorophore ( Fluo-5F ) , which fluoresces in the green ( collected via 500–550 interference filter ) . Red fluorescence was used to identify spines and dendrites . For a given stimulus , changes in fluorescence were quantified as: ΔG/R ( t ) = ( Fgreen ( t ) −Frest , green ) / ( Fred−Idark , red ) . Fgreen ( t ) is the green fluorescence signal as a function of time , Frest , green is the green fluorescence before stimulation . Idark , red is the dark current in the red channel . ΔG/R was measured in saturating Ca ( Gsat/R ) for each dye combination and batch of intracellular solution by imaging a sealed pipette with equal volumes intracellular solution and 1 M CaCl2 . ΔG/R measurements from the spine were divided by Gsat/R , yielding the reported measures of ΔG/Gsat . This value is independent of the collection efficiencies of red and green photons and should be directly comparable across laboratories . In order to estimate the time course of synaptic Ca influx , synaptic Ca transients were deconvolved using the impulse response kernel derived as follows . Spines were stimulated with 2-photon uncaging in the presence of CPP/MK801 to block NMDARs . The resulting spine ΔG/Gsat transients are well fit by a single exponential with a time constant of 42 ms ( see Figure 3B ) . Since this represents the response to an impulse-like synaptic stimulus that is much briefer than the kinetics of Ca clearance , it is a good approximation of the impulse response and can be used as a deconvolution kernel . In order to perform the deconvolution , ΔG/Gsat transients were smoothed using a 3- or 5-point box smoothing algorithm . The Fourier transforms of both the smoothed transients and the kernel were computed , and the transform of the fluorescence transient was divided by that of the kernel . The inverse Fourier transform of the resulting trace was computed , yielding a trace that is proportional to Ca current and that reports the time course of Ca influx . This deconvolution approach assumes linearity of Ca handling once it has entered the spine and uses the time course of fluorescence transients to extract the time course and the relative amplitude of Ca current . This approach does not require linear activation of Ca channels and is robust in the face of nonlinear interactions between Ca sources . In the presence of Fluo-5F , Ca handling within the spine is linear and stimulus-evoked Ca transients are well described by the convolution of the time course of Ca influx and an exponential impulse response [26] , [27] . Off-line data analysis was performed using custom software written in Igor Pro ( Wavemetrics ) and MATLAB ( Mathworks ) . The early and late phases of the Ca transient were calculated by averaging 20–50 ms and 50–120 ms post-uncaging , respectively . The early and late phases of iCa were calculated by using the peak and the average from 10–110 ms post-uncaging , respectively . All data are expressed as the mean±SEM . In the figures , average traces are shown as the mean ( line ) ±the SEM ( shaded regions ) . A two-tailed t-test was used to determine significance of differences in uEPSP and ΔG/Gsat across conditions . p<0 . 05 was considered significant . | The vast majority of excitatory synapses in the mammalian central nervous system are made onto dendritic spines , small ( < 1 fL ) membranous structures stippled along the dendrite . The head of each spine is separated from its parent dendrite by a thin neck – a morphological feature that intuitively suggests it might function to limit the transmission of electrical and biochemical signals . Unfortunately , the extremely small size of spines has made direct measurements of their electrical properties difficult and , therefore , the functional implications of electrical compartmentalization have remained elusive . In this study , we use spatiotemporally controlled stimulation to generate calcium signals within the spine head and/or neighboring dendrite . By comparing these measurements we demonstrate that spines create specialized electrical signaling compartments , which has at least two functional consequences . First , synaptic stimulation , but not similar dendritic depolarization , can trigger the activation of voltage-gated calcium channels within the spine . Second , voltage changes in the spine head arising from compartmentalization shape the time course of synaptically evoked calcium influx such that it is biphasic . Thus , the electrical compartmentalization provided by spines allows for multiple modes of calcium signaling at excitatory synapses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"neuroscience/neuronal",
"signaling",
"mechanisms"
]
| 2009 | Biphasic Synaptic Ca Influx Arising from Compartmentalized Electrical Signals in Dendritic Spines |
The common pathogen Epstein-Barr virus ( EBV ) transforms normal human B cells and can cause cancer . Latent membrane protein 2A ( LMP2A ) of EBV supports activation and proliferation of infected B cells and is expressed in many types of EBV-associated cancer . It is not clear how latent EBV infection and cancer escape elimination by host immunity , and it is unknown whether LMP2A can influence the interaction of EBV-infected cells with the immune system . We infected primary B cells with EBV deleted for LMP2A , and established lymphoblastoid cell lines ( LCLs ) . We found that CD8+ T cell clones showed higher reactivity against LMP2A-deficient LCLs compared to LCLs infected with complete EBV . We identified several potential mediators of this immunomodulatory effect . In the absence of LMP2A , expression of some EBV latent antigens was elevated , and cell surface expression of MHC class I was marginally increased . LMP2A-deficient LCLs produced lower amounts of IL-10 , although this did not directly affect CD8+ T cell recognition . Deletion of LMP2A led to several changes in the cell surface immunophenotype of LCLs . Specifically , the agonistic NKG2D ligands MICA and ULBP4 were increased . Blocking experiments showed that NKG2D activation contributed to LCL recognition by CD8+ T cell clones . Our results demonstrate that LMP2A reduces the reactivity of CD8+ T cells against EBV-infected cells , and we identify several relevant mechanisms .
Epstein-Barr virus ( EBV ) , which belongs to the human herpesvirus family , is a persistent virus carried by more than 90% of the adult population worldwide . EBV has a preferential B cell tropism , and latently infected B cells constitute the viral reservoir in healthy carriers [1] . Acute infection can lead to infectious mononucleosis ( IM ) , a self-limiting lymphoproliferative disease characterized by expansion of EBV-infected B cells and virus-specific CD8+ T cells [2] . EBV is an oncovirus , and can contribute to the development of various cancers , such as Burkitt lymphoma , nasopharyngeal carcinoma and Hodgkin lymphoma [3 , 4] . In healthy carriers , EBV infection is under control of a diverse repertoire of antigen-specific T cells , and an important role is played by CD8+ T cells that recognize viral protein-derived peptides presented by MHC class I molecules [2] . In contrast , immunosuppressed patients who lack EBV-specific T cell responses , such as patients after transplantation , are prone to developing EBV-associated lymphoproliferative disease . This condition can be treated or prevented by transfer of EBV-specific T cells [5–7] . In immunocompetent EBV carriers , a majority of EBV-infected B cells in peripheral blood carry EBV without expressing any viral protein , a state that is called "true latency" or "latency 0" [4 , 8] . Thus , such latently infected B cells are invisible to EBV-specific T cells . In contrast , during lytic EBV replication many viral proteins are expressed [9 , 10] . In this situation , the virus would be particularly vulnerable to immune control . Thus , EBV has evolved a number of proteins expressed in the lytic cycle that interfere with the display of viral antigens to CD8+ T cells . These proteins include BNLF2a , which inhibits the transporter of antigen processing [11] , BILF1 , which induces MHC class I internalization and degradation [12] , and BGLF5 , which inhibits cellular protein biosynthesis [13] . In proliferating infected B cells , EBV installs another program of gene expression , the "growth" or "latency III" program . This type of latency is found in in vitro EBV-induced lymphoblastoid cell lines ( LCLs ) , in post-transplant lymphoproliferative diseases [14] , as well as in EBV-infected B cells in lymphoid organs during primary and persistent EBV infection , where this program is thought to lead to amplification of EBV load through proliferation of infected cells [4 , 8] . Several immunogenic EBV antigens , the latent membrane proteins ( LMP1 , LMP2A , LMP2B ) and the Epstein-Barr nuclear antigens ( EBNA1 , -2 , -3A , -3B , -3C , -LP ) , are expressed in latency III EBV-infected B cells [9 , 10] . How do B cells expressing the EBV "growth program" manage to escape from recognition and elimination by virus-specific T cells ? Previous studies on immunoevasion in EBV latency have focused on the nuclear protein EBNA1 or the latent membrane protein LMP1 . EBNA1 interferes with its own presentation to CD8+ T cells through its glycine-alanine repeat domain [15 , 16] , which reduces processing by the proteasome [17] and interferes in cis with EBNA1 translation [18–20] . As a result , presentation of EBNA1 epitopes on MHC class I to T cells is reduced . Likewise , LMP1 interferes in cis with presentation of its own epitopes [21] . Although several other viral proteins are expressed in the EBV growth program , it has remained unknown whether presentation to T cells of epitopes from these proteins may be suppressed by viral mechanisms . The EBV latent protein LMP2A is a regular constituent of the EBV growth program , and is also expressed in a variety of EBV-associated cancers [9 , 10] . LMP2A has various functions in infected cells . Reminiscent of the accessory subunits of the B-cell receptor , the N-terminal cytoplasmic domain of LMP2A activates protein tyrosine kinases and induces downstream pathways of B cell activation [22 , 23] . Accordingly , LMP2A can stand in for deficient B-cell receptor signaling in mouse or human models , ensuring B cell survival [24 , 25] . In EBV-infected B cells , however , LMP2A counteracts lytic EBV reactivation triggered by cross-linking of the B-cell receptor [26–28] . No consensus has been reached yet on the importance of LMP2A in B cell proliferation and transformation [25 , 29–35] . Given these complexities , we hypothesized that LMP2A may have other functions that are not cell-intrinsic or directly related to virus replication , but related to immune control . This possibility was already suggested by the observation that LMP2A modulates signalling of type I/II interferon receptors in epithelial cells [36] , that the presence of LMP2A alters the expression of several immune-related genes [37] , and that LMP2A increases expression of the cytokine interleukin-10 ( IL-10 ) [38] , which may exert immunomodulatory functions . In this study , we investigated the influence of LMP2A in recognition of infected cells by immune effector cells . We show that LMP2A reduces recognition of infected B cells by EBV-specific CD8+ T cells , and we describe several mechanisms that may contribute to this effect .
We established EBV-transformed B cell lines ( lymphoblastoid cell lines , LCLs ) with an EBV deleted for LMP2A [25] . This virus is deleted for the promoter and the first exon of LMP2A on a background of EBV strain B95 . 8 . Expression of LMP2B is still possible in this mutant . In line with previous findings [25 , 29] , we found that the LMP2A-deficient virus ( ΔLMP2A ) had reduced efficiency of B cell transformation . To facilitate the establishment of LMP2A-deficient LCLs , we infected primary B cells with mutant EBV on a layer of murine fibroblasts overexpressing human CD40 ligand ( CD40L ) . Infection with recombinant EBV 2089 [39] that contains the complete B95 . 8 EBV genome ( here denoted "wild-type" , WT ) , which is parental to the ΔLMP2A construct , was carried out in parallel under the same conditions . Outgrowing B cell cultures were expanded and maintained in the absence of CD40L stimulators . Under these conditions , WT and ΔLMP2A LCLs could be established with similar efficiency , and expanded in parallel using the same procedures . A closer analysis of established WT and ΔLMP2A LCLs showed that the rate of apoptosis was the same , but proliferation was somewhat slower in ΔLMP2A LCLs ( S1 Fig ) . Thus , LMP2A increased the efficiency of EBV transformation in vitro , but was not essential for the proliferation of established LCLs . We analyzed the reactivity of EBV-specific CD8+ T cells to ΔLMP2A and WT LCLs ( Fig 1 ) . We found that CD8+ T cell clones specific for epitopes from all latent antigens tested ( EBNA1 , EBNA3A , LMP2 ) showed a higher IFN-γ release in response to ΔLMP2A LCLs than to WT LCLs ( Fig 1A and 1C ) . CD8+ T cells specific for the LMP2 epitope CLG recognized ΔLMP2A LCLs , because the CLG peptide is derived from a transmembrane region that is shared between LMP2A and LMP2B . CD8+ T cell clones specific for lytic-cycle antigens ( BRLF1 , BZLF1 ) showed weak recognition of both types of LCLs , and therefore differences in recognition could not be detected ( Fig 1B ) . Thus , LMP2A interferes with CD8+ T cell recognition of EBV latent antigens . To confirm that these differences in T cell recognition were caused by LMP2A and not some other unrecognized deviations between the two EBV constructs , we tested the effect of LMP2A on CD8+ T cell recognition in isolation , in the absence of an EBV genome ( Fig 1D ) . Co-transfection of LMP2A reduced CD8+ T cell recognition of 293T kidney cells transfected with the HCMV antigen pp65 . This experiment showed that the effect of LMP2A on T cell recognition was not limited to the context of the EBV genome . In the early stages of infection , there are differences in EBV gene transcription in B cells carrying LMP2A-negative EBV as opposed to LMP2A-positive EBV [35] . Thus , we investigated whether the observed differences in T cell recognition of ΔLMP2A and WT LCLs were related to differential expression of EBV antigens . Average transcript levels of several EBV latent antigens ( EBNA1 , EBNA3A , LMP2 ) appeared to be increased in ΔLMP2A LCLs ( Fig 2A ) . However , this difference reached p < 0 . 05 only for EBNA1 . No difference between WT and ΔLMP2A LCLs was seen for median expression of the lytic-cycle genes BZLF1 and gp350 . Thus , LMP2A may downmodulate the expression of some latent antigens in EBV-infected B cells , in particular EBNA1 . This may contribute to the reduced presentation of these antigens to CD8+ T cells by WT LCLs . LMP1 is an EBV protein that may alter CD8+ T cell recognition of infected cells , in particular by inducing MHC I pathway components through NF-κB , but also by inducing immunomodulatory genes [21 , 40] . We found that expression of LMP1 at the protein level was somewhat reduced in ΔLMP2A LCLs ( Fig 2B ) . This argued against a possible role of LMP1 in contributing to increased recognition of ΔLMP2A LCLs by upregulating MHC I presentation . Next , we investigated whether LMP2A modulated the reactivity of CD8+ T cells to EBV-infected B cells by mechanisms other than altering the availability of EBV antigens . We loaded WT and ΔLMP2A LCLs exogenously with peptides CRV and VLE , derived from the human cytomegalovirus ( HCMV ) protein IE-1 , and we analyzed LCL recognition by HCMV-specific CD8+ T cell clones ( Fig 3 ) . Peptide-loaded ΔLMP2A LCLs were more strongly recognized by these CD8+ T cells than peptide-loaded WT LCLs , resulting in higher IFN-γ release . We also investigated direct killing by cytotoxic CD8+ T cells , but did not observe differences in killing of WT and ΔLMP2A LCLs loaded with HCMV peptides ( S2 Fig ) . The reasons for differential regulation of IFN-γ secretion and direct cytotoxicity in this setting remain to be elucidated . Because the intracellular antigen processing machinery was bypassed in these peptide-loading experiments , LMP2A appears to act on CD8+ T cells through mechanisms other than regulation of EBV antigens or of intracellular processing pathways . Therefore , we studied the effect of LMP2A on cell surface-residing or secreted factors relevant for CD8+ T cell recognition . It was recently shown that LMP2A increases IL-10 production in infected B cells [38] . The possibility of a similar effect in our system was intriguing , because cellular IL-10 and its viral homolog reduce the antiviral activity of different types of immune effector cells [41–43] . In accordance with Incrocci and colleagues [38] , we found that WT LCLs released higher amounts of IL-10 than LCLs lacking LMP2A ( Fig 4A ) . These levels of secreted IL-10 were not mirrored by transcription levels for human IL-10 ( Fig 4B ) , which suggested an effect of LMP2A on post-transcriptional regulation of IL-10 [44] . In contrast to cellular IL-10 , transcription of viral IL-10 was very low in each type of LCL ( Fig 4B ) , in accordance with its description as a lytic-cycle gene [45] . To determine whether differences in IL-10 release could directly influence T cell reactivity to LCLs , we used specific antibodies to block IL-10 receptor on CD8+ T cells ( Fig 4C and 4E ) , or to neutralize IL-10 in the supernatant ( Fig 4D and 4F ) . In each case , recognition of WT or ΔLMP2A LCLs was not altered . Thus , modulation of IL-10 secretion by LMP2A did not directly affect the ability of CD8+ T cells to recognize infected B cells . This experiment did not rule out indirect effects of secreted IL-10 , which may act back on the LCLs over time in culture and modulate their immunogenicity . We continued by analyzing ΔLMP2A and WT LCLs for cell surface molecules involved in the interaction between CD8+ T cells and LCLs . First , we determined the levels of total MHC-I and individual MHC-I allotypes ( Fig 5 ) . MHC-I was marginally increased in LCLs deleted for LMP2A as compared with WT LCLs ( p = 0 . 0046 ) . A similar tendency was observed for some of the individual MHC-I allotypes , but did not reach p < 0 . 05 . Next , we examined whether expression of selected costimulatory and immunomodulatory molecules on the surface of LCLs was altered in the absence of LMP2A ( Fig 6 ) . We found strong differences in expression for some of these molecules . The coinhibitory B7 family molecule PD-L1 ( B7-H1 ) was ( somewhat unexpectedly ) induced in ΔLMP2A LCLs , whereas the costimulatory B7 molecule CD86 was equally expressed on ΔLMP2A and WT LCLs . CD11a , the α chain of the integrin LFA-1 that plays important roles in the immunological synapse , was strongly downregulated in the absence of LMP2A , whereas ICAM-1 ( CD54 ) , its counterpart , was expressed equally in the presence or absence of LMP2A . So far , these alterations were not obviously connected with the increased susceptibility of ΔLMP2A cells to CD8+ T cell recognition . Recent reports suggested that EBV infection induces ligands of the coactivatory receptor NKG2D , a molecule expressed on T and NK cells [46–49] . However , a comprehensive analysis of NKG2D ligands on LCLs has not previously been performed . Our analyses by flow cytometry showed that EBV infection induced the expression of three NKG2D ligands ( MICA , MICB and ULBP4 ) on LCLs ( Fig 6 ) . These molecules were not expressed on primary B cells . Markedly higher levels of MICA and ULBP4 were detected on ΔLMP2A LCLs as compared to WT LCLs , whereas MICB levels did not differ ( Fig 6 ) . We could not detect expression of the other five NKG2D ligands ( ULBP1 , 2 , 3 , 5 , 6 ) on the surface of WT or ΔLMP2A LCLs with available monoclonal antibodies , but this does not rule out that these molecules may as well be modulated by LMP2A . Our results suggested a possible contribution of NKG2D ligands to differential recognition of LCLs by CD8+ T cells . We tested the functional relevance of differential NKG2D ligand expression for CD8+ T cell recognition . An analysis of NKG2D levels on several CD8+ T cell clones showed that all were positive for NKG2D ( Fig 7A ) . Differences in NKG2D expression levels were not correlated with antigen specificity . When we blocked NKG2D on EBV-specific CD8+ T cells with a specific antibody , IFN-γ release after contact with LCLs was reduced ( Fig 7B–7D ) . A reduction in the reactivity of CD8+ T cells to both WT and mutant LCLs was observed after blocking , but reduction was even slightly stronger for ΔLMP2A LCLs than for WT LCLs ( Fig 7C and 7D ) . Likewise , blocking NKG2D on HCMV-specific CD8+ T cell clones led to reduced recognition of peptide-loaded LCLs ( Fig 7E ) . Thus , NKG2D ligands on LCLs contribute to their recognition by CD8+ T cells irrespective of antigen specificity . LMP2A reduces CD8+ T cell recognition of EBV-infected B cells by reducing the expression of NKG2D ligands . Since expression of PD-L1 , a ligand of the immunomodulatory receptor PD-1 on T cells , was increased on ΔLMP2A LCLs ( Fig 6 ) , the question emerged whether PD-L1 may counteract T cell recognition of ΔLMP2A LCLs . In this case , even greater differences in T cell recognition of ΔLMP2A LCLs as opposed to WT LCLs might be revealed by masking the effects of PD-L1 . Since it was reported that stimulation of PD-L1 on LCLs induces their apoptosis in a T-cell-independent manner [50] , we used a PD-1-blocking antibody , EH12 . 2H7 , that was described to interfere with T-cell-inhibitory interactions of PD-L1 and PD-1 [51] . Thus , we tested blocking antibodies to NKG2D and PD-1 in T cell recognition assays . Interestingly , blockade of PD-1 did not increase T cell recognition of ΔLMP2A LCLs , but reduced it , although less so than blockade of NKG2D ( Fig 8 ) . Addition of PD-1 antibody to NKG2D antibody did not further modify recognition of ΔLMP2A LCLs , although recognition of WT LCLs was additively reduced by the two antibodies . We conclude that the increased amounts of PD-L1 on ΔLMP2A LCLs did not counteract T cell recognition and resulting IFN-γ production .
In this report , we show that LMP2A interferes with CD8+ T cell recognition of latently infected B cells , and identify several mechanisms that may contribute to this interference . First , we found that LMP2A decreased mRNA expression levels of EBV latent antigens targeted by CD8+ T cells , in particular EBNA1 . Second , LMP2A downregulated MHC class I , although to a limited extent . Third , two ligands of the coactivatory receptor NKG2D were strongly upregulated in LMP2A-deficient LCLs , and blocking of NKG2D reduced T cell recognition of infected cells . We conclude that LMP2A hampers CD8+ T cell recognition of infected cells through different mechanisms including regulation of NKG2D ligands . A basis for the present work was the efficient generation of ΔLMP2A LCLs . The importance of LMP2A for human B cell transformation by EBV has been controversial: some studies did not identify a role of LMP2A [30–34] , but others reported that LMP2A increases B cell proliferation and transformation [25 , 29 , 35] . In our experience , LMP2A is important for establishment of EBV latent infection in vitro , as we found it difficult to establish ΔLMP2A LCLs under standard conditions . However , when we supplemented a strong CD40 stimulus for the first days after infection [52 , 53] , ΔLMP2A LCLs and WT LCLs could be generated with similar yield , and ΔLMP2A LCLs could then be maintained autonomously . This finding confirmed that LMP2A is not essential for maintenance and proliferation of established LCLs , as long as the B cell receptor is expressed [25] . Another important function of LMP2A is its role in "stabilizing latency" , i . e . prevention of lytic-cycle induction . A complex picture has emerged , in which LMP2A interferes with lytic induction after exogenous B-cell receptor stimulation [26–28 , 30 , 54] , but induces basal levels of lytic gene expression in the absence of such a stimulus [55] . In accordance with earlier results [35] , we found that baseline lytic gene expression was low both in WT and ΔLMP2A , and therefore was without consequence for recognition by T cells [2 , 56] . Thus , the "latency-stabilizing" function of LMP2A does not appear to be relevant for T cell recognition of B cells that express the growth program in the absence of exogenous triggers . It remains to be investigated whether the immunomodulatory functions of LMP2A extend to cells in lytic cycle . Two EBV latent proteins , EBNA1 and LMP1 , were previously shown to modulate antigen presentation to CD8+ T cells . EBNA1 does not affect presentation of other antigens , but specifically interferes in cis with its own presentation by blocking its own proteasomal processing [17] and modulating its translation [19 , 20] , both through its glycine-alanine-rich domain . LMP1 is a strong inducer of MHC I presentation through activation of the NF-κB pathway [21 , 57] , but contains a structural element that acts in cis to protect epitopes derived from its own sequence from efficient presentation [21] . LMP1 is also unusual in that peptides derived from secreted LMP1 were shown to interfere with T cell activation [58] . However , this requires amounts of LMP1 that are much higher than those secreted by EBV-infected cells [58] . To our knowledge , it has remained untested whether EBV latent antigens more generally affect recognition by and activation of CD8+ T cells . Suggestions regarding a role of LMP2A in T immune modulation emerged from comprehensive microarray-based analyses of LMP2A-mediated changes to the transcriptome of mouse and human B cells [37] . Interestingly , transcription of genes in the inflammation/immunity category , including interferon-regulating factors , was repressed by LMP2A in human LCLs , whereas no such genes were induced [37] . Among genes of direct relevance for the B-cell—T-cell interface , CD86 was induced and LFA-1 was repressed by LMP2A in BJAB cells , but no differential expression of these genes was found in LCLs with or without LMP2A [37] . These findings highlight that the effects of LMP2A depend on the cellular context , and that T-cell-modulatory functions of LMP2A in more restricted modes of EBV latency may hypothetically be even stronger than in the LCL system studied here . For example , the ability of LMP2A to interfere with signaling through interferon receptors [36] may further contribute to LMP2A-mediated evasion from T cell recognition . Our data demonstrated that LMP2A markedly reduced the reactivity of EBV-specific CD8+ T cells against LCLs . This was true for all latent EBV antigens investigated ( LMP2 , EBNA1 , EBNA3A ) . The epitopes we analyzed are processed by different pathways for their presentation: CLG and FLY are TAP-independent epitopes , with FLY being immunoproteasome-dependent [59 , 60] , while RPP and HPV are TAP-dependent [61] . A reduction of CD8+ T cell reactivity was also observed on LCLs loaded with exogenous peptides from a different pathogen , which makes it clear that LMP2A does not exclusively affect intracellular mechanisms of antigen provision and presentation . Reduced CD8+ T cell reactivity in the presence of LMP2A was observed in the context of all HLA allotypes that were studied: HLA A*0201 , B*0702 , and B*3501 for intracellularly processed EBV epitopes , HLA A*0201 and C*0702 for exogenously loaded HCMV epitopes . Thus , our data indicate that LMP2A affects CD8+ T cell reactivity regardless of the identity of the peptide presented , the mechanism of processing , or the presenting HLA molecule . However , our results also suggested an antigen-specific aspect to the immunomodulatory effects of LMP2A , because we found a trend toward elevated expression of several latent genes in the absence of LMP2A . This is in line with the idea that LMP2A may mediate global B-cell transcription factor regulation to reduce expression of EBV latency proteins [62 , 63] . This was not true for LMP1 , however , whose protein expression in the absence of LMP2A was reduced . Our findings are in line with similar tendencies in EBV latent gene expression in the first seven days after B cell infection with EBV ΔLMP2A [35] . It is intriguing that EBNA1 was the EBV latent antigen whose mRNA expression was most clearly reduced by LMP2A , since both antigens are part of the restricted EBV gene expression spectrum in latency II EBV malignancies such as nasopharyngeal carcinoma and Hodgkin lymphoma [64 , 65] . If LMP2A interferes with presentation of EBNA1-derived and other peptides also in latency II type cancers , this will have important implications for their immune surveillance . Among immunomodulatory cytokines , IL-10 was a particularly interesting candidate in our context , because it is constitutively produced at high levels by EBV-transformed B cells [66 , 67] , and a recent report showed that LMP2A increased IL-10 production in Burkitt lymphoma cell lines [38] . Moreover , EBV encodes a viral homologue of human IL-10 [68] . Both human and viral IL-10 were suggested early on to interfere with cellular immune responses to EBV [41 , 69] , but it may be difficult to distinguish an immunomodulatory function of cellular or viral IL-10 from their function as growth factors for EBV-infected B cells [66 , 70 , 71] . vIL-10 contributes to downregulation of the transporter of antigen processing 1 ( TAP1 ) and MHC-I in the early phase of B cell infection [43] , but recognition of early-infected B cells by EBV-specific CD8+ T cells was not increased in the absence of vIL-10 [42] . Our data showed that LCLs lacking LMP2A released lower amounts of IL-10 compared to WT LCLs , but reactivity of CD8+ T cell clones was not altered by neutralization of IL-10 or blocking of the IL-10 receptor . However , a more indirect role of IL-10 remains possible . Therefore , LCL-secreted IL-10 may act back on the LCLs over time , and thus downregulate MHC-I or other relevant molecules [43 , 72] in WT LCLs more strongly than in ΔLMP2A LCLs . PD-1 is an immunomodulatory receptor found on activated T cells , on exhausted virus-specific T cells in chronic infection , but also on functional EBV-specific effector memory T cells in latent infection [73] . Blocking of the interaction between PD-1 and its ligand , PD-L1 , may restore antiviral T cell function [74] . Somewhat counter-intuitively , we found PD-L1 to be downregulated in LCLs in the presence of LMP2A . When we blocked PD-1 on CD8+ T cells , we did not observe increased reactivity to LCLs , but rather a reduction in reactivity . The possibility remains that PD-L1 on EBV-infected B cells exerts a more long-term influence on shaping specific CD8+ T cell repertoires and functions that was not tested in our experiments . NKG2D is an agonistic receptor on T and NK cells and recognizes a number of ligands that are upregulated on target cells in conditions such as malignant transformation , viral infection or heat shock [75] . Increased expression of some NKG2D ligands after EBV infection was described [46–48 , 76] , but a comprehensive analysis of NKG2D ligands on LCLs has been lacking . Our analysis of the eight known NKG2D ligands showed that EBV infection induced the expression of three of them ( MICA , MICB , and ULBP4 ) , and that induction of MICA and ULBP4 was even more increased in the absence of LMP2A . In addition , we demonstrated that blocking of NKG2D on CD8+ T cells distinctly affected the recognition of LCLs by these effector cells . A recent study has shown that in patients with genetic deficiencies in the magnesium transporter MAGT1 , who are particularly susceptible to EBV infection and EBV+ lymphomas , NKG2D plays an important role in the control of EBV infection by NK and CD8+ T cells [46] . A role for NKG2D in control of EBV-associated cancer has been further illustrated in a mouse model of LMP1-induced cancer that could be therapeutically targeted through NKG2D [76] . Targeting of the NKG2D ligand MICB by an EBV-encoded miRNA may decrease susceptibility of EBV-infected B cells to lysis by NK cells [77] . Thus , NKG2D ligands represent important coagonists for EBV-specific adaptive and innate immunity , and it appears an efficient strategy for the virus to blunt cellular immune responses by decreasing surface expression of NKG2D ligands through the action of LMP2A . Taken together , we describe here a functional immunomodulatory effect for the EBV protein LMP2A , and show that LMP2A mediates partial escape of infected B cells from recognition by CD8+ T cells . Several immunoevasive mechanisms are driven by LMP2A in EBV-infected B cells . Thus , it will be urgent to determine the role played by LMP2A in evasion from T and NK cell recognition in other modes of EBV infection , and in different types of EBV-associated lymphoproliferative and malignant diseases .
Mononuclear cells were isolated from standard blood donations by anonymous healthy adult donors purchased in the form of buffy coats from the Institute for Transfusion Medicine , University of Ulm , Germany , or from voluntary healthy adult blood donors providing written informed consent . The institutional review board ( Ethikkommission , Klinikum der Universität München , Grosshadern , Munich , Germany ) approved this procedure . All work was conducted according to the principles expressed in the Helsinki Declaration . Virus-producing cell lines for recombinant EBV 2089 ( EBV WT ) , derived from EBV strain B95 . 8 [39] , and its ΔLMP2A-deleted derivative EBV 2525 [25] were provided by Wolfgang Hammerschmidt [25 , 39] . In 293HEK-derived EBV producer cell lines , which stably carry the EBV genome in an episomal form , the viral lytic cycle was induced by transient transfection of expression plasmids coding for transcription factor BZLF1 and glycoprotein gp110/BALF4 [78] . EBV-containing supernatant was harvested three days later , centrifuged to reduce cellular debris , filtered ( 0 . 8 μM ) , and stored at 4°C . Titer of infectious virus was determined by infecting Raji cells for three days and quantifying GFP-positive cells as described [78] . Infection of B cells was performed at 0 . 1 virus units per cell . Standard cell culture medium was RPMI 1640 with 10% foetal bovine serum , 100 U/ml penicillin , 100 μg/ml streptomycin ( all from Invitrogen ) , and 100nM sodium selenite ( ICN ) . Stimulator cell line LL8 was generated by stable transfection of L929 mouse fibroblasts with an expression plasmid for human CD40 ligand carrying a G418-selectable marker , followed by two rounds of single-cell cloning under selection . We found this stimulator cell line to be functionally analogous to the one described earlier [53 , 79] . Lymphoblastoid cell lines were established from primary B cells purified from freshly isolated PBMCs . Untouched B cells were negatively isolated using Human B Cell Isolation Kit II ( Miltenyi Biotec , Bergisch Gladbach , Germany ) . Enrichment of B cells was verified by flow cytometry ( anti-CD19 clone HIB19; anti-CD3 clone HIT3a , Biolegend ) , and was in the range of 95–98% . B cells were plated at 1×105 cells/well in 24-well plates on an adherent cell layer of irradiated ( 180 Gy ) CD40 ligand-expressing LL8 cells in standard medium supplemented with 1 mg/mL cyclosporine A ( Novartis ) . B cells were infected with 0 . 1 virus units per cell . Half of the culture supernatant was exchanged at day 1 post infection . Outgrowing cultures were transferred after 1–2 weeks to a fresh plate , and further cultivated autonomously in the absence of LL8 cells . Presence of mutant EBV and absence of endogenous EBV wild type was confirmed by PCR every few weeks with primers L2BRC ( 5'-GCTTCCTCGTGCTTTACGGTATC-3' ) and L2BRD ( 5'-AAGAACTTTGACCTGTTGTCCCTG-3' ) for amplification of a product bridging the LMP2A deletion , primers L2INA ( 5'-CATTGCGGGTGGATAGCCTC-3' ) and L2BRD for amplification of the deleted sequence . Proliferating EBV-infected LCLs were analyzed and used in T cell assays between 1 . 5 and 4 months after infection . DNA transfection experiments were performed with 293T human embryonic kidney cells with a plasmid expressing HCMV pp65 fused to a destabilized green fluorescent protein ( GFP ) under the HCMV immediate-early promoter ( provided by Manuel Albanese and Wolfgang Hammerschmidt ) ( pp65-GFP ) . A plasmid expressing full-length LMP2A under the control of the SV40 early promoter ( pSV-LMP2A ) and the corresponding control ( pSV ) were used for co-transfection . The pp65-GFP plasmid and pSV plasmids were used at a 1:10 ratio . One day before transfection , 293T cells were plated in 24-well plates ( 1×105 cells/well ) . Transfection was performed with a mix of 1 . 17 μl of Metafectene Pro ( Biontex ) and 390 ng of plasmid DNA in a volume of 100 μl of OptiMEM ( Gibco ) for each well , according to the manufacturer's protocol . Two days after transfection , cells were used for T cell assays and assessment of transfection efficiency by FACS . About 40–50% of cells were GFP-positive . To identify EBV-positive donors among anonymously purchased buffy coats , serum-containing cell-free supernatant was tested for IgG specific for EBV EBNA1 and VCA by a rapid immunofiltration assay ( RDT EBV IgG Assay , Bio-Rad ) . EBV-specific T cells were directly isolated from PBMCs of EBV-seropositive , HLA-typed donors after stimulation with a matched peptide and IFN-γ secretion assay ( Miltenyi Biotec ) . For single T cell cloning , isolated IFN-γ-secreting cells were seeded into round-bottom 96-well plates at 0 . 7 or 2 . 5 cells per well in 200 μl of medium supplemented with 1000 U/mL rIL-2 , 1×105/mL irradiated ( 50 Gy ) HLA-matched LCLs , and 1 . 5×106/mL of a mixture of irradiated ( 50 Gy ) allogeneic PBMCs from at least three different donors . Outgrowing T cell clones were expanded in round-bottom 96-well plates by restimulating every 2 weeks under the same conditions . The specificity of the T cell clones was determined by IFN-γ ELISA with individual antigenic peptides ( see below ) , and by staining with HLA-peptide pentamers ( Proimmune , Oxford , UK ) . The T cell clones were specific for the following epitope peptides , abbreviated by their first three amino acids in one-letter code: CLG ( CLGGLLTMV , LMP2 , A*0201 ) [80] , YVL ( YVLDHLIVV , BRLF1 A*0201 ) [81] , FLY ( FLYALALLL , LMP2 , A*0201 ) [60 , 82] , RPP ( RPPIFIRRL , EBNA3A , B*0702 ) [83] , RAK ( RAKFKQLL , BZLF1 , B*0801 ) [84] , HPV ( HPVGEADYFEY , EBNA1 , B*3501 ) [85] . HCMV-specific CD8+ T cells clones specific for NLV ( NLVPMVATV , pp65 , A*0201 ) [86] , CRV ( CRVLCCYVL , IE-1 , C*0702 ) [87] and VLE ( VLEETSVML , IE-1 , A*0201 ) [88] were obtained as described [87] . Flow cytometric analysis was performed with a BD FACS Calibur or a BD LSR Fortessa machine . Analysis of WT and mutant LCLs lacking LMP2A established from the same donor was always conducted in parallel and for at least one WT line and one ΔLMP2A line . Generally , 1–1 . 5×105 cells were stained in a V-bottom 96-well plate at 4°C for 20 minutes in a volume of 20 μl , washed twice in 200 μl of buffer ( PBS + 2% FCS ) , resuspended in buffer , and analyzed immediately . When unlabeled antibodies were included in the staining that required counterstaining with labeled anti-Ig antibodies , two to three rounds of staining and washing were performed as necessary . Antibodies anti-MICA ( clone 159227 , unlabeled ) , anti-MICB ( clone 236511 , unlabeled ) , and anti-ULBP1 ( clone 170818 , unlabeled ) were from R&D Systems and were counterstained with anti-IgG-APC ( clone Poly4053 ) from Biolegend . Anti-ULBP2/-5/-6 ( clone 165903 , PE-labeled ) and anti-ULBP3 ( clone 166510 , PE-labeled ) , were from R&D Systems; anti-ULBP4 ( 6E6 , unlabeled ) was from Santa Cruz; anti-CD11a ( clone G43-25B , PE-labeled ) was from BD Bioscience; anti-CD48 ( clone 5F4 , PE-labeled ) , anti-PD-L1 ( clone 29E . 2A3 , APC-labeled ) , anti-CD86 ( clone IT2 . 2 , APC-labeled ) , anti-ICAM-1 ( clone HCD54 , APC-labeled ) , anti-HLA-ABC ( clone W6/32 , APC-labeled ) , and anti-HLA-A2 ( clone BB7 . 2 , PE-labeled ) were from Biolegend; and anti-HLA-B7 ( clone BB7 . 1 ) was from Millipore . Antibody anti-HLA-A3 ( clone 4i87 , IgM , USB ) was counterstained with anti-IgM-PE ( clone RMM-1; BioLegend ) . A hybridoma producing the HLA-C/HLA-E-specific antibody DT9 ( IgG2b ) was kindly provided by Véronique Braud , Nice , France [89] , and counterstained with anti-mouse IgG-APC ( clone Poly4053 ) from Biolegend . HLA-Bw6 was stained with a PE-labeled human antibody ( REA143 , 130-099-843 ) from Miltenyi Biotec . Isotype controls were IgG2A ( clone 133304 ) and IgG2B ( clone 133303 ) from R&D Systems; IgG1 ( clone MOPC-21 ) , IgG2A ( clone MOPC-173 ) , and IgG2B ( clone MG2b-57 ) from Biolegend . T cells were stained with antibodies anti-NKG2D ( 1D11 , APC-labelled ) , anti-CD8 ( RPA-T8 , Pacific Blue- or FITC-labelled ) , anti-CD3 ( HIT3a , PE-Cy5-labelled ) from BioLegend . Combined analysis of proliferation and apoptosis of LCLs was performed using CellTrace Violet ( Life Technologies ) , AnnexinV-Cy5 conjugate ( ApoScreen , Southern Biotech ) , and propidium iodide ( PI , Life Technologies ) . One million cells was stained with 1 μl of CellTrace Violet in 1 ml PBS , washed , cultivated in 3 ml of full medium in a 12 well plate at 1x106 cells/well , and incubated for 4 days . Cells were harvested , counted , and 2 . 5×105 cells were stained in 200 μl buffer with 2 μl of AnnexinV-Cy5 and propidium iodide at 1 μg/ml , before proceeding to flow cytometric analysis . For IFN-γ ELISA , effector cells ( 2 . 5×104 cells/well ) and target cells ( 5×104 cells/well ) were co-cultivated in 200 μl/well in a V-bottom 96-well plate at 37°C and 5% CO2 . For IL-10 ELISA , LCLs were plated at 5×105 cells/ml in a 12-well or V-bottom 96-well plate and incubated at 37°C and 5% CO2 . Supernatants were harvested after 16–18 hours . ELISA was performed according to the manufacturer's instructions ( Mabtech , Nacka , Sweden ) . Blocking by specific purified antibodies was performed where indicated . Antibody was added to the effector cells ( anti-NKG2D , anti-IL10R ) or to the target cells ( anti-IL10 ) at a pre-established concentration and incubated for 1 hour at 37°C prior to the addition of the target or effector cells , respectively . We used antibodies to IL-10 and IL-10R that were previously shown to neutralize activity at the same or lower concentrations [90–92] . Antibodies used for blocking , and matched isotype controls , were all low-endotoxin , azide-free ( LEAF ) and purchased from Biolegend: anti-NKG2D ( clone 1D11 , used at 50 μg/ml ) with isotype ( mouse IgG1 , clone MOPC-21 ) , anti-IL10R ( clone 3F9 , used at 20 μg/ml ) with isotype ( rat IgG2a , clone RTK2758 ) , anti-IL10 ( clone JES3-9D7 , used at 20 μg/ml ) with isotype ( rat IgG1 , clone RTK2071 ) , anti-PD-1 ( clone EH12 . 2H7 , used at 10 μg/ml ) with isotype ( mouse IgG1 , clone MOPC-21 ) . Investigation of the recognition by CD8+ T cell clones of WT and ΔLMP2A LCLs established from the same donor was always performed in parallel and for at least one WT line and one ΔLMP2A line . Statistical analysis was performed with GraphPad Prism software . The cytotoxic reactivity of CD8+ T cell clones against target cells was measured by calcein-release assay . Target cells ( 4×105 ) were labeled with 1 μg/ml in 500 μl medium . After incubation for 1 hour at 37°C cells were washed 3 times with sterile PBS , and 5×103 target cells/well were plated in a V-bottom 96-well plate ( final volume 200 μl/well ) . For each target cell type , spontaneous release ( no effector cells , 0% lysis ) and maximal release ( addition of 0 . 5% of triton-X 100 , 100% lysis ) was determined . Effector cells were co-incubated with target cells for 3 hours at a 2:1 ratio . Afterwards , 100 μl of supernatant were collected and transferred to a fresh flat-bottom 96-well plate and fluorescence intensity at 485/535 nm was measured in an Infinite F200 PRO fluorometer ( Tecan ) . RPMI without phenol red was used to reduce background fluorescence . Total RNA was extracted from LCLs with the RNeasy Mini Kit , and cDNA synthesis was performed with the QuantiTect kit , both from Qiagen , Hilden , Germany . Quantitative PCR was performed on a LightCycler 480 ( Roche , Basel , Switzerland ) using the SYBR Green LC480 Mix . Primers were as follows: human IL-10 ( forward: 5'-GCAGGTGAAGAATGCCTTTA-3' , reverse: 5'-CCCTGATGTCTCAGTTTCGT-3' ) , BZLF1 unspliced ( forward: 5'-GCACATCTGCTTCAACAGGA-3' , reverse: 5'-CCAAACATAAATGCCCCATC–3' ) , EBNA1 ( forward: 5'-CGCAAGGAATATCAGGGATG-3' , reverse: 5'-TCTCTCCTAGGCCATTTCCA-3' ) , gp350 ( forward: 5'- TTGTGAAATTTCGCCATCCT-3' , reverse: 5'-CAAAACCCCGTGTACCTG-3' ) . Primers specific for BCRF1 ( vIL-10 ) , EBNA3A , LMP2AB , and GUSB were described before [42] . Specific mRNA expression was standardized to the housekeeping gene β-glucuronidase ( GUSB ) [91 , 93] . WT and ΔLMP2A LCLs simultaneously established from the same donor were always analyzed in parallel . Cells were incubated for 15 min on ice with lysis buffer ( 50 mM Tris/HCl pH 7 . 4 , 150 mM NaCl , 1% NP40 , 0 . 5% DOC , 0 . 1% SDS ) together with protease inhibitor ( completeMini , Roche ) . Protein concentration was determined with the Bio-Rad Protein Assay . Proteins were separated on an 8% SDS-PAGE gel and transferred to a nitrocellulose membrane by semi-dry blotting . Blots were probed with antibodies specific for LMP1 ( 1G6 , provided by Elisabeth Kremmer , 1:25 dilution ) [94] and GAPDH ( 1A7 , 1:10 dilution ) . Blots were further probed with secondary antibodies conjugated to horseradish peroxidase , and immunoreactive proteins were detected by incubation with chemoluminescence substrate ( 0 . 1M Tris/HCl , pH 8 . 8 , 200 mM p-Coumaric Acid in DMSO , 1 . 25 mM Luminol in DMSO ) and exposure of CEA RP NEW films ( Agfa HealthCare , Belgium ) . | Epstein-Barr virus ( EBV ) is carried by most humans . It can cause several types of cancer . In healthy infected people , EBV persists for life in a "latent" state in white blood cells called B cells . For infected persons to remain healthy , it is crucial that they harbor CD8-positive "killer" T cells that recognize and destroy precancerous EBV-infected cells . However , this protection is imperfect , because the virus is not eliminated from the body , and the danger of EBV-associated cancer remains . How does the virus counteract CD8+ T cell control ? Here we study the effects of latent membrane protein 2A ( LMP2A ) , which is an important viral molecule because it is present in several types of EBV-associated cancers , and in latently infected cells in healthy people . We show that LMP2A counteracts the recognition of EBV-infected B cells by antiviral killer cells . We found a number of mechanisms that are relevant to this effect . Notably , LMP2A disturbs expression of molecules on B cells that interact with NKG2D , a molecule on the surface of CD8+ T cells that aids their activation . In this way , LMP2A weakens important immune responses against EBV . Similar mechanisms may operate in different types of LMP2A-expressing cancers caused by EBV . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Latent Membrane Protein LMP2A Impairs Recognition of EBV-Infected Cells by CD8+ T Cells |
Under normal conditions the immune system has limited access to the brain; however , during toxoplasmic encephalitis ( TE ) , large numbers of T cells and APCs accumulate within this site . A combination of real time imaging , transgenic reporter mice , and recombinant parasites allowed a comprehensive analysis of CD11c+ cells during TE . These studies reveal that the CNS CD11c+ cells consist of a mixture of microglia and dendritic cells ( DCs ) with distinct behavior associated with their ability to interact with parasites or effector T cells . The CNS DCs upregulated several chemokine receptors during TE , but none of these individual receptors tested was required for migration of DCs into the brain . However , this process was pertussis toxin sensitive and dependent on the integrin LFA-1 , suggesting that the synergistic effect of signaling through multiple chemokine receptors , possibly leading to changes in the affinity of LFA-1 , is involved in the recruitment/retention of DCs to the CNS and thus provides new insights into how the immune system accesses this unique site .
Resistance to the parasite Toxoplasma gondii is characterized by an acute phase of infection associated with activation of innate immune cells such as neutrophils , dendritic cells ( DCs ) , and macrophages that produce IL-12 , which promotes NK and T cell production of IFN-γ [1] , [2] These events lead to the early control of parasite replication and the transition into the chronic phase of infection where the parasite persists as tissue cysts , most notably within the CNS [3]–[5] . While the immune privileged nature of the brain may contribute to the ability of T . gondii to persist , several studies have highlighted that the inflamed brain ( during CNS infections and during EAE ) is an immunologically active site where activated T cells are constantly surveying their environment [6]–[9] . CD4+ and CD8+ T cells contribute to the local control of T . gondii in the CNS but much less is known about the contribution of various APC populations within the brain to T cell mediated resistance to this infection [10]–[12] . DCs are regarded as the most potent APCs responsible for initiation of immune responses and have a critical role in promoting resistance to T . gondii [13]–[16] . Highly activated DCs are also present within the brain during TE [17] , [18] and similar accumulation of DCs within the CNS has been reported in other models of neuro-inflammation [19] , [20] . While the DCs that are found in the CNS under normal conditions ( primarily in the meninges and choroid plexus ) have been proposed to be tolerogenic [21]–[23] , the DCs present during neuro-inflammation have been ascribed numerous functions . During EAE , DCs isolated at the onset of disease were found to prime inflammatory T cells , while the DCs isolated at the peak of disease were relatively poor APCs and supported the development of regulatory T cells [19] , [24] , [25] . Other studies indicate that CNS DCs can function as efficient APCs and may serve to amplify intra-cerebral T cell responses [17] , [26] . Regardless , during TE , the behavior and function of DCs within the CNS remains unclear . Multiple mechanisms have been proposed to account for the accumulation of DCs within the CNS during neuroinflammation . Some studies indicate that microglia can differentiate into DCs during inflammation [18] , [27] , while others have proposed that DCs in the brain arise from peripheral haematopoietic cells [19] , [28] . During TE , the precise origin of the CNS DCs is unclear: whether they are a local population that is constantly renewed or they are recruited from the peripheral circulation has not been addressed . Furthermore , while recent studies have highlighted the molecular signals required for the migration of DCs within secondary lymphoid compartments [29] , much less is known about the migratory patterns of DCs into peripheral sites such as the inflamed brain and the signals required for their retention within these sites . To better understand the phenotype and function of the CNS resident DCs during TE , a combination of imaging techniques , transgenic reporter mice ( CD11cYFP and OT1GFP ) and recombinant parasites expressing ovalbumin and fluorescent tags ( td Tomato ) were used . These studies revealed that during TE , the CD11c+ cells form an extensive network consisting of microglia and DCs . In vitro antigen presentation studies and real time imaging suggest that the CNS resident CD11c+ cells can function as efficient APCs . These studies also showed that CD11c+ cells isolated from the brain expressed higher levels of the chemokine receptors CCR7 , CCR5 , CXCR3 , CX3CR1 compared to their splenic counterparts , suggesting a possible role for these receptors in the migration and retention of these cells within the brain . However , the loss of any individual chemokine receptor ( tested ) did not affect the ability of DCs to migrate into the brain , in an adoptive transfer model . Nevertheless , this process was pertussis toxin sensitive ( indicating a role for Gαi coupled chemokine receptors ) and LFA-1 dependent . These studies point towards a model in which the synergistic effect of signaling through multiple chemokine receptors , potentially leads to a change in the affinity of the integrin LFA-1 on the surface of DCs and controls their migration and retention within the CNS during TE .
TE is associated with increased numbers of CD11c+ cells within the brain [17] , however little is known about the subset composition , phenotype and behavior of these cells in vivo . Therefore , the APC populations within the brain during TE were characterized by flow-cytometry . In uninfected mice there were minimal numbers of immune cells present in the brain , and microglia ( CD45int CD11b+ cells ) were the dominant immune population ( Figure 1A top ) . During TE , in addition to the microglia , there is an increase in the number of APCs ( CD3−CD19−NK1 . 1− ) of haematopoietic origin ( CD45hi ) ( Figure 1A bottom ) . These include macrophages ( CD11c−CD11b+ ) and DCs ( CD11c+ ) . The myeloid DCs ( CD11b+CD11c+ ) form the largest subset ( about 80–90% of the CD11c+ fraction ) , followed by the lymphoid ( CD8+CD11c+ ) and plasmacytoid ( PDCA+ B220+ ) subsets ( each constitute <5% of the CD11c+ population ) . All of these populations were activated based on their expression of high levels of markers such as MHC I , MHC II , CD80 and CD86 ( Figure S1 ) . To directly visualize CD11c+ cells within the brain during TE , CD11cYFP mice were used . In these reporters , few YFP+ cells were identified within the brain parenchyma in naïve mice ( Figure 1B ) . During TE , consistent with the flow-cytometric data , there was an increase in the number of CD11cYFP cells ( Figure 1B ) , which formed an extensive network within the brain parenchyma . Analysis of YFP+ cells revealed that they consisted of DCs ( CD11chiCD45hi ) and microglia ( CD11bhiCD45int ) , consistent with previous reports that microglia can up-regulate CD11c during inflammation [30]–[32] ( Figure 1C ) . In situ imaging by multiphoton microscopy using the CD11cYFP mice revealed two distinct morphological phenotypes: amoeboid and dendriform cells ( Figure 1D ) . In the naive brain the CD11c+ cells exhibited an amoeboid structure whereas during infection there was an increased proportion of dendriform CD11c+ cells ( Figure 1E ) . In comparison to the cells seen in the naïve brain , during TE there was marked increase in the average surface area of the CD11c+ cells , ( Figure 1F right ) and an increased proportion of CD11c+ cells that displayed a vacuolated phenotype ( Figure 1F left ) . Using fluorescent parasites ( PruOVA Tomato ) , infected cells could be identified ( Figure 1G ) , however a large proportion of the vacuolated cells were not associated with parasites . The presence of vacuoles in the absence of identifiable parasites has been noted previously both with DCs in the lymph node and astrocytes in the brain during T . gondii infection [8] , [14] and is indicative of an activated phenotype . The proportion of CD11c+ cells ( DCs and microglia ) that were infected with parasites was less than 10% ( Figure 1H ) . CD11c+ cells were also observed associated with parasite cysts within the CNS ( Figure 1I , Video S1 ) , with about 60–70% of the cysts intimately associated with CD11c+ cells ( Figure 1J , Video S2 ) . Together , these studies show that TE results in a significant increase in CD11c+ cells composed of a mixed population of activated microglia and dendritic cells , which exhibit distinct morphologies and behaviors . Since CNS resident DCs can have immuno-stimulatory or tolerogenic properties , studies were performed to assess whether the CD11c+ cells present in the brain during TE can present antigen to T cells . Naïve OT1GFP T cells were adoptively transferred into CD11cYFP mice followed by infection with Pru or Pruova parasites , to determine whether antigen specific T cells interacted with DCs in vivo . In CD11cYFP mice infected with parental strain of the parasite ( Pru ) there was minimal infiltration of OT1GFP cells into the brain ( Figure 2A ) , but during infection with PruOVA , significant numbers of T cells were present within the brain ( Figure 2A , Video S3 ) . The OT1GFP cells interacted with the CD11cYFP cells and made sustained contacts ( highlighted with red circles ) or short-lived contacts ( highlighted with blue circles ) as seen in the representative time series ( Figure 2B ) . Further analysis revealed that the T cells showed a range in their motilities ( average velocity ∼10 µm/min ) with highly motile cells , cells that paused briefly and cells that were rounded up and stationary for much of the imaging period ( 10–15 minutes ) . In contrast , the CD11c+ cells showed minimal motility ( average <2 µm/min ) ( Figure 2C left ) . The duration of the interactions between the T cells and DCs was calculated and the frequency of T cells exhibiting different contact durations was determined ( Figure 2D ) . Majority of the T-DC interactions observed were between uninfected DCs and T cells , which is consistent with the data that only a small percentage of CD11c+ cells are infected . ( Figure 1H ) . Analysis of the interaction times revealed two populations: a larger population ( P1 ) of motile T cells that interacted with CD11c+ cells for shorter durations ( less than 4 minutes ) and a smaller subset ( P2 ) of more rounded T cells that were engaged in prolonged interactions with DCs ( Figure 2D ) . The prolonged interactions were however not restricted to the infected DCs and could thus be indicative of antigen presentation or target cell lysis . Since CNS resident DCs have been proposed to generate tolerogenic T cell responses [21] , [33] , [34] , the ability of DCs present in the CNS during TE to prime naïve antigen specific T cells was evaluated . CD11c+ cells were isolated from either the brain or spleens of infected mice , incubated with ovalbumin and co-cultured with naïve CFSE labeled OT1 T cells . The OT1 T cells that were primed by the CD11c+ cells from the brain proliferated ( as indicated by dilution of CFSE ) , albeit at a slightly lower rate compared to the T cells that were primed by splenic DCs ( Figure 2E ) . The CD8+ T cells primed under the two different conditions ( splenic or brain CD11c+ cells ) exhibited a similar pattern of cytokine secretion and produced IFN-γ but not IL-10 ( Figure 2F ) . Control experiments were also performed to compare DCs isolated from the spleens of naive and infected mice and these studies did not reveal any significant differences in their ability to stimulate OT1 T cells ( data not shown ) . Together , these data indicate that the CD11c+ cells found during chronic toxoplasmosis are capable of processing and presenting antigen to CD8+ T cells and do not skew naïve T cells to a tolerogenic phenotype . The source of the CD11c+ cells found within the CNS during TE is not clear and several alternate mechanisms have been proposed for their recruitment and retention during neuro-inflammation . In order to identify whether CD11c+ cells present in the brain during TE were derived from haematopoietic precursors , bone marrow cells ( depleted of mature T cells ) from CD11cYFP transgenic mice were used to reconstitute lethally irradiated WT recipient mice . In such CD11cYFP->WT chimeras very few CD11cYFP cells ( bone marrow derived ) could be found within the brain in the naïve mice ( Figure 3A ) . However , during chronic TE , significant accumulation of CD11cYFP cells was observed within the brain ( Figure 3A ) . Radiation chimeras were also generated using congenic mice , which differed in the CD45 allotype ( CD45 . 1->CD45 . 2 ) to distinguish donor and recipient cells . These studies revealed that greater than 95% of the DCs and approximately 70-75% of the microglia found within the brain during TE were of donor origin ( Figure 3B ) indicating that the microglia and DCs present within the brain during chronic TE are largely bone marrow derived . In order to understand the factors that influence the recruitment of DCs to the brain during TE , the levels of chemokine receptor mRNA and surface expression on CD11c+ cells within the brain during infection were compared to those in the periphery . The number of CD11c+ cells that could be isolated and purified from an uninfected brain was minimal and precluded analysis . Nevertheless , there were significant differences in the levels of chemokine receptor mRNA expression between CD11c+ cells isolated from the brains of infected mice compared to splenic populations . While chemokine receptors such as XCR1 and CXCR5 were not altered , others such as CCR1 were lower on the CNS resident DCs compared to splenic DCs ( Figure 3C ) . However , CD11c+ cells from the brain expressed increased levels of transcripts for CCR7 , CXCR3 , CCR5 and CX3CR1 ( Figure 3D left ) and flow cytometric analysis showed that both microglia and DCs upregulated these chemokine receptors . In contrast , CXCR4 was expressed only on the DCs and not on microglia during infection . These studies identify multiple chemokine receptors that represent candidates to mediate DC migration and retention within the brain . To understand the factors that control the recruitment and retention of DCs during TE , an adoptive transfer system was developed to quantify the migration of DCs into this site . Initial experiments using either immature or mature bone marrow DCs showed limited migration of this DC subset into the CNS ( data not shown ) . As an alternate approach , CD11c+ cells that were expanded in vivo with FLT3L expressing tumor cells and subsequently purified using CD11c+ MACS beads , were used as a source of DCs [35] . Morphologically these cells resemble DCs ( Figure S2A ) and flow cytometric analysis of these cells showed that they are CD11c+ , lineage negative ( CD3−CD19−NK1 . 1− ) , MHC class II+ and the proportion of various DC subsets is comparable to those found in naïve spleens ( CD11b+CD11c+ , CD11b−CD8+ ) as seen previously [35] , [36] . The in vivo expanded DCs showed a chemokine receptor pattern typical of mature splenic DCs and express CCR7 , CXCR3 , CCR5 and CX3CR1 ( Figure S2A ) . To determine if these DCs could traffic into the brain during chronic TE , they were transferred iv into naïve or infected mice and were tracked by 1 ) using congenic markers ( CD45 . 1 ) 2 ) loading DCs with NIR−polymersomes or 3 ) labeling DCs with dyes such as TRITC ( Figure S2B ) . In the first example , using the congenic marker CD45 . 1 , transferred DCs could be tracked within the brain of infected recipients by flow cytometry 18–24 hours post transfer ( Figure 4A ) . Analysis of various markers including CD11c , CD11b , MHCII and Ly6C indicated that the DCs that migrated to the brain were representative of the transferred population and not significantly different from those found in the other compartments such as the blood , lymph nodes and spleen ( Figure 4A ) . In a different approach to track the transferred population , DCs that phagocytosed polymersomes ( containing near infrared dyes ) were transferred into naïve or infected mice and 18–24 hours later near infrared imaging of whole organs was performed ( Figure 4B ) . In naïve mice , fewer DCs migrated to the brain and these cells were primarily within the spleen and lymph nodes . In contrast , in infected mice significant numbers of polymersome loaded DCs could be detected within the brain ( Figure 4B ) . DCs labeled with TRITC could similarly be tracked within the brain using multi photon microscopy and the transferred DCs could be co-imaged with OT1GFP cells ( Figure 4C , Video S4 ) and endogenous CD11cYFP cells ( Figure 4D , Video S5 ) . Morphologically the transferred DCs exhibited lower average surface area and had a slightly higher velocity compared to endogenous DCs ( Figure 4E , Video S5 ) . However , these cells exhibited a surface phenotype typical of DCs ( CD11c+ , CD11b+ , MHC class II+ , Ly6C− ) as seen in Figure 4A . In other experiments transferred monocytes were similarly shown to traffic to the brain but exhibited a distinct phenotype ( CD11c− , Ly6C+ , CD11b+ , CD115+ ) compared to the DCs ( Figure S3 ) . These approaches validate a model in which the entry of mature DCs from the circulation to the brain during chronic infection can be quantified and these populations can be visualized . To determine whether the migration of DCs into the brain was dependent on chemokine receptor mediated signaling , the effect of pertussis toxin ( PTX ) , an inhibitor of Gαi coupled chemokine receptors , was tested . In these studies , DCs were pretreated with PTX in vitro for three hours , washed and transferred into naive or T . gondii infected mice ( day 21 post infection ) , and analyzed 18–24 hours post-transfer . Flow cytometric analysis revealed that significantly lower numbers of PTX treated DCs accumulated within the brains of infected mice compared to control DCs ( Figure 5A ) . The numbers of DCs in the spleen was not altered by the PTX treatment , however their migration into the lymph nodes was also PTX sensitive . The peripheral blood showed an increase in the number of DCs in the PTX treated group , suggesting that the cells that are not recruited in the tissue compartments remain in the blood . Similar studies using polymersome loaded DCs and whole animal imaging revealed that the accumulation of PTX treated DCs was much lower than untreated DCs within the infected brain ( Figure 5B ) . These data demonstrate that the accumulation of adoptively transferred DCs in the brain during TE is a pertussis toxin sensitive process . Since the migration of transferred DCs into the brain was blocked by PTX , the role of individual chemokine receptors in DC migration was tested . Since CCR7 was expressed at high levels on brain DCs , the expression of the chemokine CCL21 ( ligand for CCR7 ) was elevated during infection ( Figure 6A ) , and CCR7 on the brain DCs was functional in in vitro chemotaxis studies ( Figure 6B ) , this receptor was first targeted . To directly test the role of CCR7 in the trafficking of DCs into the brain , the transfer system described above was modified to allow co-transfer of equal numbers of CFSE labeled WT and TRITC labeled CCR7−/− DCs . Real time imaging studies indicated that both WT and CCR7−/− DCs could be found in equivalent frequencies within the brain parenchyma ( Figure 6C top ) and there was no significant difference in their average migration velocity ( Figure 6C bottom ) . To track these cell populations by flow cytometry CD45 . 2+ WT and CD45 . 2+ TRITC+ CCR7−/− were transferred into infected CD45 . 1+ mice thus enabling identification of the WT and CCR7−/− DCs within the same recipients ( Figure 6D ) . However , the absence of CCR7 did not significantly affect the accumulation of DCs in the brain ( Figure 6E ) . The role of other chemokine receptors that were up regulated on the CNS DCs including CCR5 , CXCR3 , CX3CR1 , CCR2 and CCR6 were examined using the respective chemokine receptor deficient mice in this competitive transfer model . In these studies the absence of any one of these receptors did not affect the ability of DCs to migrate into the infected brain ( Figure 6E ) , suggesting the cooperative action of multiple chemokines in this process or the role of an as yet untested chemokine receptor/s . Signaling downstream of the Gαi coupled receptors has been shown to increase the affinity of many integrins particularly LFA-1 and VLA-4 , for their respective ligands [37]–[40] . Since our studies showed an effect of pertussis toxin treatment on migration of DCs into the brain , the role of integrins in this process was examined . Previous studies have highlighted the importance of VLA-4 in the migration of T cells into the inflamed CNS [9] , [41] however the role of integrins in the migration of DCs to this site is less clear . Flowcytometric analysis of BMNCs revealed that several integrins including LFA-1 ( αLβ2 ) and VLA-4 ( α4β1 ) were expressed by the microglia and DCs , but there was minimal expression of CD103 ( αEβ7 ) by the CD11c+ cells from the brain ( Figure 7A ) . Of the various integrins tested LFA-1 was expressed at the highest level by the DCs . While DCs in the brain or spleen did not change their expression of LFA-1 during infection , the microglia showed an upregulation of LFA-1 . To test whether high levels of LFA-1 seen on the surface of the DCs had a functional role in their migration to the CNS , the high affinity form of LFA-1 was blocked using monoclonal antibodies in the adoptive transfer model with DCs . DCs were pretreated in culture for 3 hours with anti-CD11a antibody ( or control antibody ) , or alternatively , the mice were treated with the blocking antibodies prior to DC transfer . Both pre-treatment of DCs with anti CD11a antibody and in vivo antibody administration , resulted in an inhibition in the ability of the DCs to migrate to the brain ( Figure 7B: Group 1 versus Group 2 and 3 ) . Mice that received anti CD11a treated DCs and were additionally treated in vivo with the blocking antibody displayed the most pronounced inhibition ( Group 1 versus Group 4 ) . Blocking VLA-4 had minimal effect on the migration of DCs into the brain using this adoptive transfer system ( Figure S4 ) . Together these data show that the migration of adoptively transferred DCs into the CNS during TE is an LFA-1 dependent process .
Dendritic cells have been identified within the CNS during various parasitic , bacterial , viral infections and autoimmune conditions such as EAE [6] , [18] , [19] , [42] , [43] . However , there is no clear consensus about the function of CNS resident DCs , their origin , behavior and migratory patterns . Thus , several studies have highlighted that CNS DCs are inhibitory for adaptive immune responses [21] , [22] , [33] , while others have indicated that they are capable of inducing inflammatory T cell responses [18] , [19] , [44] . The present studies indicate that the CD11c+ cells isolated from the brain during TE are capable of processing and presenting antigen to naïve T cells in vitro to generate functional T cell responses . The in vivo imaging studies show that while the majority of the T cells interacted transiently with the CD11c+ cells , a smaller subset of T cells engaged in prolonged interactions , could be identified which is consistent with the T cell-APC contacts observed recently by Robey and colleagues in the brain [45] . Our studies also revealed that only a small proportion of the DCs that were interacting with the T cells were actively infected and the prolonged interactions were not restricted to the infected cells , which suggests the involvement of both direct and cross presentation pathways in the activation of the CD8+ T cells . These data are in support of a model in which DCs and microglia provide local tonic signals to maintain the effector functions of the intracerebral T cells . In addition , the imaging studies have revealed the close association between parasite cysts and CD11c+ cells . Neurons are considered to be the major cell type that houses cysts [46] , and based on immunohistology the regions around the cyst were thought to be largely quiescent . However , the current studies reveal that the microglia and DCs are intimately associated with the parasite cysts with their dendrites and cell bodies tightly wrapped around the cysts , suggesting possible innate sensing of this stage of the parasite . Several alternate models have been proposed for the origin of DCs within the brain . Our studies using radiation bone marrow chimeras shows that the majority of the CD11c+ cells ( >95% ) that expand in the CNS during TE are bone marrow derived . This could represent constant recruitment of DCs from the periphery during infection or a local expansion of bone marrow precursor cells that initially seeded the CNS . Recent studies by Merad and colleagues has shown that the microglia are derived from primitive myeloid progenitors and not from post natal haematopoietic precursors [47] , [48] . However several other studies indicate that damage and inflammation within the CNS can result in greater proportions of haematopoietic precursor derived microglia [49] , [50] . In the present studies , the majority of the microglia found during TE are bone marrow derived . Interestingly the CD45int microglia are derived from the CD45hi precursors , however whether the down-regulation of CD45 occurs within an intermediate precursor population or on the cells that are resident within the brain for prolonged periods , is not completely clear . The adoptive transfer studies suggest that migration of DCs from peripheral circulation is one of the mechanisms that contribute to the accumulation of the CD11c+ cells within the brain during TE . Migration of monocytes and their conversion to DCs is another proposed mechanism that contributes to the DC populations within the inflamed CNS and our preliminary studies using purified monocytes indicate that they can migrate to the brain during TE ( Figure S3 ) . While conversion of the monocytes to DCs was not observed immediately ( 18-24 hours ) after transfer , whether this happens over time , remains to be tested . Several chemokines and their receptors , including CCR7 and its ligands ( CCL19 and CCL21 ) and the CCL3/CCR5 receptor ligand system have been implicated in the migration of DCs into the CNS , primarily based on EAE studies and clinical data from MS patients [51]-[53] . During chronic TE , several chemokines ( CCL19 , CCL21 , CCL3 , CCL5 , CXCL9 , CXCL10 ) are up regulated within the CNS [54] , [55] . In our studies the adoptively transferred DCs migrated to the brain through a pertussis toxin sensitive mechanism , suggesting that the signaling though the Gαi coupled chemokine receptors was involved in this process . Since the lack of any individual chemokine receptor did not affect the ability of these DCs to migrate to the brain , we conclude that this process is possibly mediated through the synergistic effect of signaling events downstream of multiple receptors [56] . A downstream effect of signaling through Gαi coupled receptors results in changes in the conformation of integrins from low to high affinity forms and has been shown to control the migration of immune cells into lymph nodes [37]–[40] . Recently the integrin VLA-4 has been shown to be involved in the recruitment of immature DCs into the CNS during EAE [34] . In our model , the migration of DCs into the brain was found to be dependent on LFA-1 , but not VLA-4 . Thus , our current studies are consistent with a model in which signaling events downstream of multiple Gαi coupled chemokine receptors results in changes in the affinity of LFA-1 on the surface of the DCs and influences their recruitment into the inflamed CNS [37]–[40] , [56] . Until recently the role of integrins within the CNS , was largely focused on the effects of VLA-4 in the migration of T cells and this forms the basis for the clinical development of a monoclonal antibody ( Natalizumab ) against α4 integrin to treat MS [57] , [58] . However , a subset of patients treated with Natalizumab developed progressive multifocal leukoencepahalopathy , due to reactivation of latent JC polyoma virus infection [59] . Interestingly another monoclonal antibody ( Efalizumab ) directed against LFA-1 , which was developed for the treatment of Psoriasis resulted in a similar complication [60] . These clinical studies indicate the importance of LFA-1 in addition to VLA-4 in the migration of immune effectors into the CNS . The identification of LFA-1 as an important molecule required for the recruitment of APCs into the CNS bears significance for the design of therapeutic targets to control CNS inflammation .
WT C57BL/6 mice and CD45 . 1 congenic mice were purchased from Jackson laboratories ( Bar Harbor , ME ) . OT1GFP mice were obtained by crossing DPE-GFP transgenic mice to OT1 TCR transgenic mice , as described previously [14] . CD11CYFP transgenic mice were obtained from Michel C Nussenzweig [61] . CCR2−/− , , CX3CR1−/− , CCR7−/− , CCR5−/− , CCR6−/− were purchased from Jackson laboratories . The CXCR3−/− mice were obtained from Andrew Luster ( MGH , Boston MA ) . The various transgenic mice colonies were maintained in a specific pathogen-free ( SPF ) facility in the Department of Pathobiology at the University of Pennsylvania in conformance with institutional guidelines for animal care . All animal studies were carried out in compliance with the guidelines of the Institutional Animal Care and Use Committee ( IACUC ) of the University of Pennsylvania and in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The animal protocol was approved by the Institutional Animal Care and Use Committee ( IACUC ) of the University of Pennsylvania , Philadelphia PA ( Multiple project assurance # A3079-01 ) . The Prugniaud strain of Toxoplasma gondii was maintained as tachyzoites by serial passage through HFF monolayers . Transgenic parasites engineered to express cytoplasmic tdTomato and to secrete ovalbumin protein into the parasitophorous vacuole ( described previously ( 46 ) ) were maintained by antibiotic selection in the cultures . In general , mice were infected ip with the various strains of T . gondii at a dose of 104 parasites per mouse . To block LFA-1 or VLA-4 in vitro , the purified DCs were treated with 30 µg/ml of anti CD11a antibody ( Clone M-17/4 , BioXcell , West Lebanon , NH ) or anti VLA-4 ( Clone PS/2 , BioXcell , West Lebanon , NH ) for 3 hours prior to transfer . For in vivo treatments chronically infected mice were treated with two doses of the anti LFA-1 or anti VLA-4 ( 150 µg each ) antibody , one dose prior to transfer of DCs ( −6 hours ) and one dose at the time of DC transfer . The control antibody used in these cases was Rat IgG . The mice were sacrificed 18–24 hours after transfer of DCs . Purified splenic DCs were loaded with Tat-functionalized near-infrared emissive polymersomes , as described previously [62] . The various organs ( brain , spleen and lymph nodes ) were isolated from the animals 18–24 hours later and fluorescence imaging was performed using an Odyssey infrared imaging system ( LI-COR Biosciences , Lincoln , NE ) . Mice were sacrificed by CO2 asphyxiation and the brain was removed with minimal damage to the tissue . A horizontal slice of the brain ( between 1 . 5–2 mm thick ) was cut as described previously [8] and placed in a temperature controlled imaging chamber and the tissue was held in place using a nylon mesh and an organ holder ( Warner Instruments , Hamden CT ) . The chamber was constantly perfused with warm ( 37°C ) oxygenated ( 95% O2/5%CO2 ) media ( RPMI+10% FBS ) and the temperature was maintained at 37°C using heating elements and a temperature control probe . Ex vivo imaging was done using a Leica SP5 2 photon microscope equipped with a picosecond laser ( Coherent Chameleon; 720 nm–980 nm ) and tunable internal detectors that allow simultaneous detection of emissions of different wavelengths and second harmonic signals ( SHG ) . EGFP , YFP and tdTomato were excited using laser light of 920 nm . Typically , z stacks of a series of x-y planes with a total thickness of 30 µm and step size of 6 µm were captured . The LAS-AF software ( Leica ) was used for image acquisition . For 4D data sets , images were captured every 25–30 seconds and the imaging depths are as specified in the different images . Volocity software ( PerkinElmer , Waltham MA ) was used to convert the three dimensional image stacks into time series . Single cell tracking was done by a combination of manual tracking and automated tracking and mean migratory velocities were calculated as described previously . For measurement of T cell–DC interactions , the duration of contacts of all the T cells observed in a given filed of view were calculated for the entire imaging period ( typically lasting 12–15 minutes ) . Cellular contacts were determined manually as a lack of space between the interacting cells and the data obtained from an average of 3–5 imaging regions from 3 individual mice per group were used to obtain the frequency of T cells with different durations of contacts with DCs . RNA was isolated from CD11c+ cells purified using CD11c MACs micro-beads ( Miltenyi Biotech , Auburn , CA ) from the brain mononuclear cells or splenocytes isolated from chronically infected mice . TRIzol ( Invitrogen ) and DNase treated total RNA was reverse transcribed using Superscript II ( Invitrogen , Carlsbad CA ) using standard protocols . Quantitative PCR was performed with customized primer sets for CCR7 , CCR5 , CX3CR1 , CXCR4 , XCR1 , CXCR5 and HPRT ( QIAGEN , Valencia CA ) using SYBR green PCR assay system and an ABI 7500 instrument ( Applied Biosystems , Carlsbad , CA ) . The values for the various chemokine receptors were normalized to HPRT and displayed as fold induction over naïve splenic DC controls . Brain mononuclear cells were isolated using a standard protocol described previously [8] . Briefly , mice were perfused with cold PBS , the brain was removed , diced , passed through an 18-gauge needle and digested with Collagenase/Dispase and DNAse for 45 minutes at 37°C . The cell suspension was then washed and fractionated on a 30%–60% percoll gradient ( Pharmacia ) for 20 minutes . The cells in the interface consisted of mononuclear cells , which were washed prior to experiments . Peripheral blood was obtained from mice by cardiac puncture , and the RBCs were separated using a Lympholyte-M gradient ( Cedarlane , ON , Canada ) . Lymphocytes were isolated from spleens and cervical lymph nodes by mechanical homogenization followed by lysis of RBCs ( for spleens ) using lysis buffer ( 0 . 846% NH4Cl ) . Freshly isolated cells were stained with the antibodies purchased from eBioscience ( San Diego , CA ) or BD Biosciences ( San Jose , CA ) . For intracellular staining the cells were stimulated ex-vivo for 5 hours in complete media with Brefeldin A , either in the presence or absence of SIINFEKL peptide ( 1 µg/ml ) . Following surface staining , the cells were fixed with 4% PFA for 10 minutes at room temperature . The cells were permeabilized using 0 . 3% saponin in staining buffer . The stained samples were run on a FACSCanto ( BD , San Jose , CA ) and results were analyzed using FlowJo software ( TreeStar Inc . , Ashland , OR ) . Statistical significance of differences between groups of mice was tested using the student's t test or ANOVA/ Kruskal-Wallis test for multiple groups . In all the cases , p<0 . 05 was considered significant . | Toxoplasmic encephalitis ( TE ) , caused by the protozoan parasite Toxoplasma gondii , can be potentially life threatening especially in immuno-compromised individuals . Immune cells including dendritic cells have been shown to accumulate in the brain during chronic toxoplasmosis; however , little is known about their function , their behavior in vivo , and the mechanisms by which they migrate into the brain . In the present studies , we utilize a combination of real time imaging , transgenic reporter mice , and recombinant parasites to reveal the distinct behavior and morphologies of dendritic cells within the brain and their ability to interact with parasites and effector T cells during TE . The CNS DCs were also found to exhibit a unique chemokine receptor expression pattern during infection , and the migration of DCs into the brain was mediated through a pertussis toxin ( which blocks signaling downstream of several chemokine receptors ) sensitive process and dependent on the integrin LFA-1 . There is currently a poor understanding of the events that lead to DC recruitment to the CNS during inflammation in general , and our studies provide new insights into the mechanisms by which antigen-presenting cells gain access to the brain during infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"biology"
]
| 2011 | Analysis of Behavior and Trafficking of Dendritic Cells within the Brain during Toxoplasmic Encephalitis |
Loss or gain of DNA methylation can affect gene expression and is sometimes transmitted across generations . Such epigenetic alterations are thus a possible source of heritable phenotypic variation in the absence of DNA sequence change . However , attempts to assess the prevalence of stable epigenetic variation in natural and experimental populations and to quantify its impact on complex traits have been hampered by the confounding effects of DNA sequence polymorphisms . To overcome this problem as much as possible , two parents with little DNA sequence differences , but contrasting DNA methylation profiles , were used to derive a panel of epigenetic Recombinant Inbred Lines ( epiRILs ) in the reference plant Arabidopsis thaliana . The epiRILs showed variation and high heritability for flowering time and plant height ( ∼30% ) , as well as stable inheritance of multiple parental DNA methylation variants ( epialleles ) over at least eight generations . These findings provide a first rationale to identify epiallelic variants that contribute to heritable variation in complex traits using linkage or association studies . More generally , the demonstration that numerous epialleles across the genome can be stable over many generations in the absence of selection or extensive DNA sequence variation highlights the need to integrate epigenetic information into population genetics studies .
Continuous trait variation in natural and experimental populations is usually attributed to the actions and interactions of numerous DNA sequence polymorphisms and environmental factors [1] . These so-called complex traits encompass many of the prevalent diseases in humans ( e . g . diabetes , cancer ) as well as many agriculturally and evolutionarily important traits ( e . g . yield , drought resistance , or flowering time in plants ) . The heritable basis of complex traits is classically thought to rest solely on the transmission from parents to offspring of multiple DNA sequence variants that are stable and causative [1] . However , accumulating evidence suggests that this view may be too restrictive , insofar as chromatin variation ( such as differential DNA methylation ) can also be propagated across generations with phenotypic consequences , independent of DNA sequence changes [2]–[7] . Indeed , examples of spontaneous , single-locus DNA methylation variants ( epialleles ) have been reported to influence a range of characters , such as flower shape or fruit pigmentation in plants [8] , [9] and tail shape or coat color in the mouse [10] , [11] . By extension , these observations raise the possibility that the genome-wide segregation of multiple epialleles could provide a so far unexplored basis of variation for many commonly studied complex traits [12] . In the flowering plant Arabidopsis thaliana , recent large-scale DNA methylation profiling has revealed a substantial degree of differences between natural accessions [13] , [14] . As these accessions also differ in their DNA sequences , experimental populations derived from them , such as backcrosses , F2-intercrosses or Recombinant Inbred Lines ( RILs ) could potentially segregate two independent sources of heritable phenotypic variation , which are difficult to disentangle from each other [12] . As a consequence of this confounding issue , there has been little effort to date to quantify the impact of epigenetic factors on complex traits and to assess their role in the creation and maintenance of phenotypic diversity in experimental or natural settings [2] , [6] . To overcome this problem as much as possible , we established a population of epigenetic Recombinant Inbred Lines ( epiRILs ) in Arabidopsis . This population was derived from two near-isogenic parental lines , one wild type ( wt ) and the other mutant for the DDM1 gene . DDM1 encodes an ATPase chromatin remodeler that is primarily involved in the maintenance of DNA methylation and silencing of repeat elements [15]–[18] . Thus , ddm1 mutant plants exhibit a ∼70% reduction of DNA methylation overall , as well as a widespread over-accumulation of transcripts corresponding to transposable elements ( TEs ) [15] , [16] , [18] . Despite this , few TEs appear to show increased transposition in ddm1 [19] , [20] , perhaps as a result of many TEs still being targeted by the RNAi-dependent DNA methylation machinery in this mutant background [21] . Consistent with these molecular properties , ddm1 plants exhibit only mild phenotypic alterations , except after repeated selfing , in which case the severity and the number of aberrant phenotypes tend to increase [22] . Genetic analysis has shown that many of these phenotypes segregate independently of the ddm1 mutation and are conditioned by recessive or dominant alleles of single loci . Furthermore , molecular characterization of five of these alleles indicated that they arose through TE-mediated gene disruption in one case [19] and through late onset epigenetic alteration of gene expression in the other cases , often in the context of genes that are tightly associated with TE sequences [19] , [22]–[27] . Based on these observations , and given a constant environment , variation in complex traits between epiRILs is expected to result from the stable inheritance of multiple epigenetic differences ( epialleles ) induced by ddm1 and/or from a small number of DNA sequence differences that might also be present between epiRILs , notably as a result of ddm1-induced mobilization of some TEs . Here , we describe the phenotypic analysis of the epiRIL population , which revealed a high degree of heritability for flowering time and plant height . We also show that the epiRILs differ by numerous parental epialleles across the genome , which demonstrates that DNA methylation differences can be stably inherited over at least eight generations in the absence of extensive DNA sequence polymorphisms and with no selection . These findings provide a first indication of the potential impact of epigenetic variation on complex traits .
The epiRIL population was initiated using two closely related parents of the same accession ( Columbia , Col ) , one homozygous for the wild type DDM1 allele ( Col-wt ) , and the other for the ddm1-2 mutant allele ( Col-ddm1 , 4th generation ) . Therefore , these two parents should differ extensively in their DNA methylation profiles [18] , but only marginally in their DNA sequence , namely at the DDM1 locus itself and at a few other sites , such as those affected by ddm1-induced mobilization of transposable elements ( see Materials and Methods and below ) . A single F1 plant was backcrossed as female parent to the Col-wt parental line . From the backcross progeny , we selected over 500 individuals of DDM1/DDM1 genotype , from which a final population of 505 Col-wt epigenetic Recombinant Inbred Lines ( Col-wt epiRILs ) were derived through six rounds of propagation by single seed descent and no selection bias ( Figure 1; Materials and Methods ) . The Col-wt epiRILs should therefore have highly similar genomes , but markedly distinct epigenomes , if the many DNA methylation variants induced by ddm1 are stably inherited . Phenotypic analysis of the Col-wt epiRILs was performed for two quantitative traits , flowering time and plant height at maturity ( Table S1 ) . As illustrated in Figure 2 and Figure 3 , larger phenotypic variation was observed among the Col-wt epiRILs , than among the Col-wt or Col-ddm1 parental lines ( see also Tables S2 , S3 , S4 ) . Increased phenotypic variation of this kind is indicative of a component of segregational variance that typically arises in the construction of Recombinant Inbred Lines obtained from parents that differ by numerous DNA sequence polymorphisms [1] , except that in the present design the two parents are expected to be nearly isogenic . To decompose the sources of phenotypic variation observed among the Col-wt epiRILs , a series of linear mixed models were fitted ( Materials and Methods and Table S5 ) . As in classical quantitative genetics analysis , the estimated between-line variance ( line-effect ) gives a direct estimate of broad-sense heritability , i . e . the fraction of phenotypic variance that is not due to environmental effects ( ; ref [1] ) . Large and significant heritability values were obtained for flowering time ( 0 . 26 , p<0 . 0001; Figure 4A; Table S5 ) and plant height ( 0 . 32 , p<0 . 0001; Figure 4B; Table S5 ) . The fact that the means ( ‘genetic’ values ) of the Col-wt epiRILs for the two traits appear to follow a continuous distribution ( Figure 4C and 4D ) suggests that both traits are subject to a “polygenic” rather than a single locus inheritance model . Moreover , the line means of flowering time and plant height are only weakly correlated with each other ( Figure 4E ) . This observation points towards a distinct heritable basis for these two traits , and indicates that the two heritability estimates obtained here are not redundant . The excess of variance and the high heritability values observed in the population of Col-wt epiRILs could be caused by ( i ) segregation of multiple parental epialleles , ( ii ) segregation of parental differences in DNA sequence created by ddm1-induced mobilization of transposable elements , and ( iii ) mutation or epimutation accumulation in the Col-wt epiRILs as a result of selfing over multiple generations . We explored the latter possibility by first comparing the heritability estimates obtained for the Col-wt epiRIL population with those from a panel of 24 Col-wt control lines ( N = 144 ) that were derived from the Col-wt parent and propagated along with the Col-wt epiRILs through six rounds of single-seed descent ( see Materials and Methods and Text S1 ) . The heritability estimates obtained in these control lines were negligible for flowering time ( ) and plant height ( ) and significantly lower compared to those of the Col-wt epiRILs ( ; , respectively; Text S1 ) . Furthermore , sublines that were derived at the F7 generation ( Figure 1 ) of the Col-wt epiRIL design made only a small contribution to the total phenotypic variance ( Figure 4A and 4B ) , suggesting that epimutation accumulation or increased mutation rate ( notably through continuing transposon mobilization , [28]; see below ) contribute minimally . Although the subline effect for flowering time does explain about 6% of the variance , this estimate is not specific to a source of new ( epi ) mutational variance but rather reflects a compound estimate that also includes gene x environment interactions as well as maternal effects . Hence , based on these subline estimates , novel DNA sequence or methylation variants that could have arisen during the selfing of the Col-wt epiRILs appear to have little phenotypic consequences . This conclusion is also supported by the very small number of lines that were lost during the construction of the Col-wt epiRILs ( 4 out of 509; Materials and Methods ) , and by the limited number of outliers ( ±3SD ) for the two complex traits considered ( Figure 2 , bottom panels ) . This contrasts with the progressive phenotypic degeneracy that has been observed upon repeated selfing of ddm1 mutant plants [22] . Evidence for a stable heritable basis of both flowering time and plant height also comes from the observation that the phenotypic means in the Col-wt epiRILs are in each case closer to the Col-wt than to the Col-ddm1 parental mean ( Figure 2 ) . This is of course entirely consistent with the backcross scheme used to derive the Col-wt epiRILs ( Figure 1 ) . Taken together , these findings provide evidence that heritable variation for flowering time and plant height in the Col-wt epiRIL population is due to the stable inheritance and segregation of parental epialleles and/or DNA insertion variants , rather than the accumulation of new mutations or epimutations . The inheritance of parental epialleles was tested by analyzing the methylation state of several loci in a number of Col-wt epiRILs . Genomic DNA was digested with the enzyme McrBC , which only cuts methylated DNA , and specific sequences were amplified by real-time PCR ( Text S1 ) . Eleven sequences were chosen that are methylated in the Col-wt and hypomethylated in the Col-ddm1 parental lines ( Figure 5 , Figure S1 ) , including the FWA gene , for which hypomethylation and ectopic expression have been associated with a large delay in flowering [24] , [29] . Additionally , three control sequences were chosen that are not methylated in either of the two parents ( Figure 5 , Figure S1 ) . Twenty-two Col-wt epiRILs were sampled at the F9 ( BC1-S7 ) generation from both ends ( but excluding outliers , see Figure 2 ) of the flowering time distribution ( Figure 5 ) . Results were consistent with the three non-methylated parental sequences being stably inherited in their non-methylated state in the Col-wt epiRILs , and with five of the eleven differentially methylated parental sequences segregating in a Mendelian or near-Mendelian manner ( 72 . 8% [∼16/22] met . /met . , 0 . 4% [∼0/22] met . /hypomet . , 26 . 8% [∼6/22] hypomet . /hypomet . at BC1-S7; Figure 5 and Text S1 ) . In contrast , the other six sequences that were differentially methylated in the parental lines , including FWA , did not segregate in the Col-wt epiRILs . Rather , these sequences were found in the fully methylated state in all 22 lines , except for the At4g0376 sequence , which was unmethylated in one line ( Figure 5 ) . These results confirm and extend those of our previous analysis which indicated that while some hypomethylated epialleles induced by ddm1 are stably inherited over at least eight generations , others efficiently regain wt DNA methylation within two to five generations following restoration of DDM1 function , as a result of being targeted by the RNAi-dependent DNA methylation machinery [21] . Taken together , these findings provide evidence that the stable inheritance and segregation of parental epialleles is likely involved in the heritable variation for flowering time and plant height in the Col-wt epiRIL population . Furthermore , the efficient DNA remethylation of a subset of ddm1-induced epialleles could partly explain the closer proximity of the Col-wt epiRIL phenotypic means to those of Col-wt parental line ( Figure 3 ) . Previous studies have shown that ddm1-induced hypomethylation and ectopic expression of FWA can be stably inherited over many generations independently of the ddm1 mutation and cause severe delay in flowering time [23] , [24] . However , our observation that FWA had wt DNA methylation levels in all 22 Col-wt epiRILs analyzed , which included 12 late flowering lines ( Figure 5 ) , suggested instead efficient RNAi-mediated DNA remethylation of this locus , and therefore at best a marginal contribution of FWA to the continuous variation for flowering time in the Col-wt epiRIL population . To explore this further , FWA methylation and expression were measured for an additional set of four early and four late flowering lines that fall within three standard deviations from the mean ( 38±10 days , Figure 2 ) , as well as for the three late flowering outlier lines ( >48 days , Figure 2 ) that are present in Col-wt epiRIL population . While FWA methylation and expression were indistinguishable from wt in all of the non-outlier lines , hypomethylation was observed in the three late flowering outlier lines and was associated with high-level expression in seedlings , where the gene is normally not expressed ( Figure 6 ) . Moreover , FWA hypomethylation and transcript accumulation in these outlier lines were much more pronounced than in the Col-ddm1 parental line and were similar to those of a previously described , ddm1-induced late flowering line ( Figure 6; [23] , [24] ) . Thus , while the FWA allele of the Col-ddm1 parent was efficiently remethylated and resilenced upon restoration of DDM1 function , further hypomethylation and reactivation occurred instead in rare cases , leading to overtly late flowering Col-wt epiRILs . These results confirm that epiallelic variation at FWA has a major effect on flowering time , but indicate also that it is rare in the Col-wt epiRIL population , concerning phenotypic outliers that were removed from the quantitative genetics analysis . We conclude therefore that epiallelic variation at FWA contributes little to the continuous variation in flowering time observed in the Col-wt epiRIL population . Apart from epialleles , DNA sequence variants caused by ddm1-induced transposon mobilization could also segregate among the epiRILs . To test this possibility , we carried out Southern blot analysis of the insertion profile of CACTA and MULE transposons , which are the two TE families for which ddm1-induced mobility has been documented [19] , [20] . Little transposition was detected for any of the three MULE copies in either the three Col-ddm1 individuals or the eight Col-wt epiRILs that were analyzed ( Figure 7A ) . In contrast , several transposition events could be detected for CACTA in the individuals of the Col-ddm1 parental line as well in the Col-wt epiRILs . More specifically , excision events were observed for three of the five CACTA copies that are present in wt Columbia , as indicated by the disappearance of the corresponding hybridizing fragments ( Figure 7B , white asterisks ) . In addition , new insertions were detected , in the form of new hybridizing fragments ( Figure 7B , black asterisks ) . The observation of continuing CACTA mobilization in the Col-wt epiRILs is consistent with previous results indicating that CACTA copies remain transpositionnally active following restoration of wild type DDM1 function through backcrosses [28] . This highly mobile transposon family may therefore contribute to the heritable variation observed among the Col-wt epiRILs , although no obvious association between specific CACTA insertion differences and flowering time variation could be detected based on our limited sampling ( Figure 7B ) .
Using a population of “epigenetic” Recombinant Inbred Lines ( epiRILs ) in the flowering plant Arabidopsis thaliana , we have demonstrated that multiple DNA methylation changes induced across the genome can be stably inherited over at least eight generations in the absence of selection , and that these changes were associated with substantial heritable variation in two complex traits . Furthermore , we show that epiallelic variation at the FWA locus has a major effect on flowering time but is rare in our epiRIL population , indicating that other loci are involved in the continuous variation for that trait in this population . In practical terms , our findings pave the way for the identification of causative epigenetic quantitative trait loci ( phQTLepi; [12] ) in the Col-wt epiRIL population using whole genome DNA methylation profiling and classical linkage mapping methods , without the confounding effect of widespread DNA sequence polymorphisms [12] . By combining bisulphite methodology to interrogate the methylation status of individual cytosines with next generation sequencing [30] , [31] , it may now be possible to identify simultaneously the epigenetic variants segregating in the Col-wt epiRIL population and the inevitable rare DNA sequence variants also present in this population , notably as a result of ddm1induced transposable element mobilization ( Figure 7 ) . Alternatively , epigenotyping and genotyping could be carried out independently , using immunoprecipitation of methylated DNA ( MeDIP ) followed by hybridization to whole genome tiling arrays and next generation sequencing , respectively . The heritability values ( around 30% ) obtained in our study are similar to those considered in classical breeding programs for the improvement of agronomic traits . If QTL mapping of the Col-wt epiRILs were to confirm that heritability is largely due to variations in DNA methylation states , the view that DNA sequence variation is the sole basis of the heritability of complex traits may need to be revised substantially . In addition , QTL mapping will provide valuable insights into how epigenetic variation can modulate the rate of DNA sequence change in a population , notably through TE mobilization . In the context of evolutionary biology , the existence of an additional mechanism for the creation of heritable variation in complex traits could explain the faster than expected adaptation to environmental change that is often observed in natural populations [32] . There is indeed mounting evidence that epigenetic alterations ( epimutations ) can arise at high frequency , in response to environmental challenges or ‘genomic shocks’ [5] , [33] , [34] . Furthermore , our findings provide clear evidence that many epigenetic variants can be stably inherited over numerous generations in the absence of selection ( [21]; this study ) . Such stability could thus provide populations with sufficient time to explore the adaptive landscape [35] , and for neutral mutations to accumulate over the new epialleles , in a process that could ultimately lead to genetic assimilation [36] . On the other hand , the observation that about one half of DNA hypomethylation variants induced by ddm1 systematically regain wt DNA methylation over two to five generations ( [21]; Figure 5 ) illustrates the potentially transient nature of many epialleles . However , analysis of FWA indicates that even in the case of these so-called remethylatable alleles , stable transmission of hypomethylated ( and reactivated ) states can occur at low frequency ( Figure 6 ) . Indeed , our findings are consistent with previous observations of sporadic occurrence of stable , phenotypic FWA hypomethylated epialleles ( fwa ) in ddm1 mutant lines [23] . Furthermore , comparison of FWA methylation and expression levels between the Col-ddm1 parental line and fwa as well as Col-wt epiRIL late flowering outliers suggests that stable transmission of hypomethylated/reactivated FWA can only occur when specific thresholds of hypomethylation/reactivation are reached ( Figure 6A ) . Finally , although no naturally hypomethylated FWA epiallele has been recovered in a survey of 96 Arabidopsis accessions [13] , it is tempting to speculate , on the basis of our observations at this locus , that the varying stability of epialleles could underlie the variable penetrance of disease-causing alleles that segregate in pedigrees , as well as the variable onset of many heritable diseases in response to developmental or environmental cues [37] . In summary , our study provides important new evidence that epigenetic variation can contribute significantly to complex traits , and lays the foundation for identifying causative loci . The conditions that promote the occurrence of epialleles and their transgenerational stability in natural settings will need to be further elucidated in order for epigenetics to be fruitfully incorporated into the quantitative genetic analysis of experimental and natural populations [12] .
The recessive ddm1-2 mutation was isolated in a screen for marked decrease in DNA methylation of centromeric repeats in EMS-mutagenized seeds of the Columbia ( Col ) accession [16] . The Col-wt and Col-ddm1 parental lines were both derived from a ddm1/DDM1 plant stock that had been maintained in the heterozygous state by repeated backcrossing to a wild type Columbia line over six generations to remove EMS-induced mutations unlinked to ddm1 ( a kind gift from Eric Richards , Washington University , Saint Louis , MO , USA ) . Homozygous DDM1/DDM1 and ddm1/ddm1 progeny were subsequently selfed for four generations . In ddm1/ddm1 plants , this generated genome-wide DNA hypomethylation as well as mobilization of some transposable elements ( [16] , [18]–[20]; Figure 7 ) . A single plant of each genotype ( Col-wt and Col-ddm1 ) was then used for the initial Col-wt epiRIL cross ( Figure 1 ) . Unlike in classical RIL construction , the two parents were thus near isogenic , being derived from siblings that underwent four generations of selfing , but differed extensively in their levels and patterns of DNA methylation . The Col-ddm1 parent that was used to initiate the Col-wt epiRIL cross looked normal and did not display any of the developmental epimutant phenotypes that have been reported in advanced ddm1 lines , such as superman [27] , fwa [23] , [24] , ball [22] , [25] , or bonsai [26] . A single F1 individual was backcrossed to the Col-wt parental line ( Figure 1 ) . The BC1 progeny was screened by PCR-based genotyping ( Text S1 ) : of the 1140 BC1 individuals genotyped , 577 were ddm1/DDM1 , 521 were DDM1/DDM1 , and 42 were ddm1/ddm1 . This last genotype was indicative of low-level contamination of the backcross progeny with seeds produced by self-pollination of the female F1 parent . Indeed , subtracting 42 and 84 potential self-pollination contaminants from the DDM1/DDM1 and ddm1/DDM1genotypic classes , respectively , gives a corrected total of 479 DDM1/DDM1 and 493 ddm1/DDM1 individuals , close to the 1∶1 ratio expected for the backcross . Only the DDM1/DDM1 individuals were considered for the construction of the Col-wt epiRILs ( Figure 1 ) , and our calculations show that this amount of contamination ( 42 out of 521 or 8% of DDM1/DDM1 BC1 individuals ) has a negligible effect on the expected epigenotype frequencies in subsequent generations ( Text S1 ) . In total , 509 out of the 521 DDM1/DDM1 BC1 individuals were selfed and one seedling per line was randomly retained from four seeds sown . This process was repeated at each of the following generations ( single seed descent ( SSD ) approach ) and ensured that seedlings could be recovered in most instances with no selection bias . Under the assumption of epiallelic stability , each of the DDM1/DDM1 BC1 founders should have inherited from the female F1 parent , on average , 50% of the transmissible DNA methylation alterations that were present in the ddm1/ddm1 grandparent ( Figure 1 ) . This should lead , after repeated selfing , to the inheritance of an average of 25% of these alterations in each Col-wt epiRIL , except of course for the 8% of Col-wt epiRILs expected to derive from self-pollination of the female F1 parent , which should have each inherited instead 50% of these alterations on average . Four Col-wt epiRILs were lost during propagation and each of the remaining 505 Col-wt epiRILs was subdivided into three sublines at the F6 generation ( Figure 1 ) to obtain 3×505 BC1-S5 ( F7 ) plants . These were again selfed , and two BC1-S6 ( F8 ) individuals per subline were retained for the phenotypic and quantitative genetics analyses . Since the ddm1 mutation is recessive , it follows that the sublines obtained at BC1-S6 had been free of the conditioning ddm1 mutant allele effect for a total of 8 generations . We also established 24 Col-wt control lines , starting from 24 full-sib individuals of the Col-wt parental line ( hence of the same genetic background as the Col-wt epiRILs ) . These control lines were propagated by repeated SSD , and subdivided into three sublines before phenotypic analyses , using the same method as described above with the Col-wt epiRILs ( Figure 1 ) . The Col-wt epiRILs ( N = 3030 ) , the Col-wt control lines ( N = 144 ) , the Col-wt ( N = 200 ) and Col-ddm1 ( N = 200 ) parental populations were grown simultaneously in two replicate climate-controlled greenhouses under long day conditions ( day: 16 h - 20°C/22°C , night: 8 h - 16°C/18°C ) with complement of artificial light ( 105 µE/m2/s ) when necessary . For the Col-wt epiRILs , one of the two BC1-S6 plants for each subline was grown in each greenhouse ( i . e . 3×505 Col-wt epiRIL plants in each greenhouse ) . Within each greenhouse , the Col-wt epiRIL plants were randomized over 28 tables ( 3×1 m2 ) . In addition , two or three plants from each parental line were systematically placed on each table . Finally , the positions of Col-wt epiRILs and parental lines were randomized within tables . Plants were grown in individual pots ( 7×7×7 cm3 ) filled with a 90∶10 mix of peat and volcanic sand , and topped with a thin layer of granulated cork . About 15 seeds were sown per pot and seedlings were thinned out to retain a single plant that appeared representative of the whole family . Plants were supplemented twice with a nutritive solution during the reproductive phase . Of the planned design , >99% of plants were available for trait measurements . Flowering time ( i . e . number of days between sowing and opening of the first flower ) was recorded during plant growth . When plants ceased flowering , they were harvested and stored in herbaria . Plant height was then measured on the dried plants . Phenotypic means and variances were calculated for the Col-wt and Col-ddm1 parental lines , the Col-wt epiRILs and the Col-wt control lines . The corresponding 95% confidence intervals were obtained empirically from 3000 non-parametric bootstrap draws . For the Col-wt epiRIL and Col-wt control populations , in which individual plants were phenotypically more similar than plants taken at random , a stratified bootstrap approach was implemented where each line was taken as an independent stratum . In this way , the boostrap estimates are consistent with the stochastic structure of the data and should therefore be unbiased [38] , [39] . This resulted in slightly more conservative confidence intervals compared to analytical estimates . This re-sampling strategy was further employed to test for differences in means and variances of the traits between selected sample pairs ( i . e . Col-wt epiRIL vs . Col-wt , Col-wt vs . Col-ddm1 , Col-wt epiRIL vs . Col-ddm1 , etc ) , yielding a bootstrapped t statistic ( tB ) and F statistic ( FB ) and their corresponding p-values ( pB ) , see Tables S2 , S3 , S4 . To test for mean differences we considered the null hypothesis against its alternative . Differences in variances were assessed by testing the null hypothesis against the alternative , where the subscripts distinguish the two different samples in the comparison . To decompose the different sources of phenotypic variation in the Col-wt epiRILs , a linear mixed model was fitted . This model took the following form: , where P is the vector of Col-wt epiRIL phenotypic values , I represents the design matrix for the fixed-effects intercepts β for each of the two greenhouses , E is a vector of micro-environmental values ( Text S1 ) with fixed effect α , L2 is the design matrix for the random Line-effect vector b2 , L2 , 3 is the design matrix for the random nested Subline-effect vector b2 , 3 , and ε is the residual error matrix . From the resulting estimates , the variance associated with the Line-effect should be directly interpreted as the portion of total phenotypic variance that is due to epigenetic differences between the lines [40] , whereas the Subline-effect estimates the variance due to new DNA sequence mutations or epimutations that may have accumulated independently in the different sublines , gene×environment interactions and maternal effects . All data points exceeding three standard deviations were excluded from the analyses . The p-values associated with each of these effects were obtained from hypothesis testing using the likelihood ratio test , where is the likelihood of the full model and is the likelihood of the reduced model ( the full model without the variable of interest ) . The is distributed as a chi-square random variable with the number of degrees of freedom equal to the difference in the number of parameters between the full and the reduced model . The 95% confidence intervals surrounding the parameter estimates were computed from 5000 parametric bootstrap samples . All analyses were performed in R [41] . DNA and RNA were extracted from seedlings and young rosette leaves , respectively , using DNeasy and RNeasy Qiagen kits , respectively . McrBC ( New England Biolabs ) digestion was performed on 200 ng of genomic DNA . Quantitative PCR was performed using an ABI 7900 machine and Eurogentec SYBR green I MasterMix Plus on equal amounts of digested and undigested DNA samples . Results were expressed as percentage of loss of molecules after McrBC digestion . Reverse transcription was performed on 1 ug of total RNA using oligodT and Superscript II ( Invitrogen ) . Quantitative PCR was performed as above . Results were expressed as percentage of expression relative to the mean value obtained for three genes ( At2g36060; At4g29130; At5g13440 ) that show invariant expression over hundreds of publicly available microarray experiments . Southern blot analysis of TE mobilization was performed as previously described , using 1 µg of genomic DNA [19] , [20] . | DNA methylation is defined as an epigenetic modification because it can be inherited across cell division . Since variations in DNA methylation can affect gene expression and be inherited across generations , they can provide a source of heritable phenotypic variation that is not caused by changes in the DNA sequence . However , the extent to which this type of phenotypic variation occurs in natural or experimental populations is unknown , partly because of the difficulty in teasing apart the effect of DNA methylation variants ( epialleles ) from that of the DNA sequence variants also present in these populations . To overcome this problem , we have derived a population of epigenetic recombinant inbred lines in the plant Arabidopsis thaliana , using parents with few DNA sequence differences but contrasting DNA methylation profiles . This population showed variation and a high degree of heritability for two complex traits , flowering time and plant height . Multiple parental DNA methylation differences were also found to be stably inherited over eight generations in this population . These findings reveal the potential impact of heritable DNA methylation variation on complex traits and demonstrate the importance of integrating epigenetic information in population genetics studies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"genetics",
"and",
"genomics/genomics",
"plant",
"biology/plant",
"genetics",
"and",
"gene",
"expression",
"genetics",
"and",
"genomics/complex",
"traits",
"plant",
"biology/agricultural",
"biotechnology",
"evolutionary",
"biology/genomics",
"plant",
"biology",
"molecular",
"biology/dna",
"methylation",
"genetics",
"and",
"genomics/epigenetics",
"molecular",
"biology",
"genetics",
"and",
"genomics",
"genetics",
"and",
"genomics/plant",
"genetics",
"and",
"gene",
"expression",
"genetics",
"and",
"genomics/population",
"genetics"
]
| 2009 | Assessing the Impact of Transgenerational Epigenetic Variation on Complex Traits |
Although the genome contains all the information necessary for maintenance and perpetuation of life , it is the proteome that repairs , duplicates and expresses the genome and actually performs most cellular functions . Here we reveal strong phenotypes of physiological oxidative proteome damage at the functional and genomic levels . Genome-wide mutations rates and biosynthetic capacity were monitored in real time , in single Escherichia coli cells with identical levels of reactive oxygen species and oxidative DNA damage , but with different levels of irreversible oxidative proteome damage ( carbonylation ) . Increased protein carbonylation correlates with a mutator phenotype , whereas reducing it below wild type level produces an anti-mutator phenotype identifying proteome damage as the leading cause of spontaneous mutations . Proteome oxidation elevates also UV-light induced mutagenesis and impairs cellular biosynthesis . In conclusion , protein damage reduces the efficacy and precision of vital cellular processes resulting in high mutation rates and functional degeneracy akin to cellular aging .
Proteome activity sustains life , whereas genome assures perpetuation of life by ongoing renewal of the proteome , granted the capacity of the proteome to repair , replicate and express the genome . Dedicated proteins determine mutation rates via the precision of the DNA replication machinery and the efficacy and precision of DNA repair systems such as DNA base pair mismatch and damage repair . Since errors in protein biosynthesis are 105 times more frequent than mutations [1] , it would seem reasonable to expect that these errors should , when affecting key proteins , have a cascading effect by allowing additional errors in both DNA replication , causing mutations , and protein biosynthesis , causing further errors . Leslie Orgel has proposed just such a vicious circle of biosynthetic errors as a primary cause of aging [2] . High fidelity performance of key cellular proteins is achieved through selective kinetic proofreading steps in the course of DNA , RNA and protein biosynthesis [3] , [4] and by the molecular repair , error correction and maintenance ( e . g . , selective turnover ) systems . Therefore , the quality of the proteome is expected to affect the quality of the genome as well as the catalytic activities , the precision of protein interactions and the control of gene expression . Here we investigate the effects of physiological oxidative damage , inflicted specifically to proteins , on cellular biosynthetic systems at both the genome and proteome levels . We test the prediction that proteome damage should affect cell fate - mutagenesis and survival - more than does the inflicted reparable genome damage . Studies of induced mutagenesis typically measure DNA damage inflicted by the mutagenic agent , ignoring the fact that DNA damaging treatments also produce oxidative damage to proteins and other cellular components . Induced mutations arise by the processing of residual ( unrepaired ) DNA damage , therefore the efficacy of relevant repair and replication proteins should determine also the frequency of induced mutations . We have measured major oxidative damage to proteins ( irreversible protein carbonylation , PC ) and DNA ( reparable 8-oxoguanine ) and found a remarkable correlation between PC and both spontaneous and UVC light-induced mutagenesis , as well as reduced DNA repair activity . Our results lend support to Orgel's error catastrophe hypothesis by showing that protein damage can lead to , or even directly produce , DNA mutations . However , unanticipated by Orgel is our finding that errors in protein biosynthesis and folding predispose proteins to irreversible oxidative damage that ultimately alters or destroys their function .
Biological effects of oxidative stress are difficult to interpret because oxidative processes mediated by reactive oxygen and nitrogen species ( ROS and RNS ) damage all classes of biological molecules . To study specifically the biological consequences of irreversible oxidative damage to proteins ( protein carbonylation , PC ) , we produced changes in intracellular PC levels , at constant levels of ROS and oxidative damage to DNA . To increase or decrease the susceptibility of proteins to carbonylation , we made use of the observations that conditions increasing errors in protein biosynthesis and folding lead to increased PC [5]–[7] . We have focused on ribosome-associated chaperone , the trigger factor ( Tig , a functional homologue of the eukaryotic RAC/NAC ) [8] , [9] as well as chaperonin GroEL/ES and chaperone DnaK/DnaJ complexes ( homologues of eukaryotic Hsp60/Hsp10 and Hsp70/Hsp40 complexes , respectively ) . Tig acts during the nascent polypeptide synthesis , to prevent premature folding , i . e . , misfolding , of proteins exiting the ribosomal tunnel . After their release from the Tig , some proteins fold without any further assistance but , depending on their size and domain complexity , populations of “client” proteins are delivered either to the GroEL/ES or to the DnaK/DnaJ complex [8] , [10] . Using ribosomal fidelity mutants ( rpsL141 and rpsD14 ) and deletion or overexpression of three classes of chaperones ( Tig , DnaK/DnaJ and GroES/EL ) , we show that the increase or decrease of errors in protein synthesis and folding elevates or reduces PC at constant ROS ( Figure 1A and 2B ) . Since the accuracy in protein biosynthesis and folding affects also the saturating levels of PC upon UVC irradiation ( details below ) , it appears that the deviation from the native structure determines protein “target size” ( a subpopulation of proteins sensitive to carbonylation ) for oxidative damage in a cell . These results relate to the early proposal of Kurland [11]: the decreased cellular fitness via decreased translational accuracy may be a consequence of the impact of mis-sense errors on the structures of proteins as well as on the growth of cells . To quantify the effect of proteome oxidation on overall cellular biosynthetic capacity , we used a validated genetic method - the number of λ phages produced by individual E . coli cells ( single burst size ) - that has revealed a strong correlation between the radiation-induced PC and the single burst size [12] . Bacteriophage λ requires and diverts host cell transcription , translation and DNA replication machineries for its own reproduction . In Figure 1A and 1B , we show the cellular PC before infection ( absorbance range 0 . 16 to 0 . 51 ) and the single burst size ( range 3 to 90 ) after infection by a single viral particle . There is a robust negative linear correlation ( R2 = 0 . 75 ) between the λ single burst size and physiological host cell PC ( Figure 1C ) . The level of ROS was identical in all studied strains ( Figure 2B ) , and the infecting viral genome was undamaged , suggesting that the large differences in λ burst size are due to large differences in the host cell PC and/or some other parallel oxidative protein damage . Moreover , the antioxidant treatment with 1 mM water-soluble vitamin E ( trolox ) - present only before infection - greatly improved λ production in all tested strains in proportion to the antioxidant effect on PC ( Figure 1 , Table S1 ) , suggesting that PC , rather than misfolding alone , precludes biosynthesis in E . coli . Double mutant ΔtigΔlon , lacking the ribosome-associated chaperone and the main protease accumulates large amounts of carbonylated proteins , is unable to form colonies and produces only 3–5 phages per cell ( Figure 1B ) . However , it forms colonies on plates with trolox and produces about 30 phages per cell when grown in the presence of trolox prior to infection ( compare 17 and 18 in Figure 1B ) . Support for the idea that proteome alterations should lead to genome alterations ( mutations ) came from the observation that two mutator ( high mutation rate phenotype ) loci mutA and mutC encode altered tRNAs - key elements in protein biosynthesis [13] . Here we address the question: can we detect genetic consequences of proteome damage ? To determine the global genomic mutation rates in growing E . coli ( strains listed in Figure 1D and Table S2 ) , we used a unique method that detects every genomic mutation ( each seen as a fluorescent MutL-CFP focus ) emerging in the last DNA replication round - in single cells and in real time - providing directly the numbers for genome-wide mutation rates [14] . Functional MutL protein tagged with a fluorescent protein ( CFP ) forms persistent foci only on unrepaired mismatches that are the emerging new mutations [14] . Only the mismatch-recognizing protein MutS is required for the formation of MutL foci at unrepaired mismatches [14] . In mutH mutants , deficient in the late stage of mismatch repair , all replication errors remain in the newly synthesized DNA forming MutL-CFP foci such that the mutation rate corresponds to the frequency of DNA replication errors . We found that in both wild type and in mutH cells , high-fidelity ribosomal mutation ( rpsL141 ) and over-expression of Tig , GroEL/ES and DnaK chaperones decrease the mutation rate to create the most potent anti-mutator effect ever observed ( in average , 10-fold in mutH and 3-fold in wild type , Figure 2A ) . The frequency of cells with a MutL-CFP focus is reduced from 0 . 3% in the wild type ( corresponding to the genetic estimates [15] ) to 0 . 09% or below . The 3-fold anti-mutator effect ( p<0 . 0001 ) in the wild type is mirrored by an average of 9-fold mutator effect in strains with low-fidelity ribosome ( rpsD14 ) or deletions of tig and dnaK ( groEL/ES deletion is lethal ) exhibiting around 3% of cells bearing a MutL-CFP focus ( Figure 2A ) . The ROS levels and the levels of 8-oxoguanine , the most mutagenic oxidative DNA damage , remain equal in all strains ( Figure 2B ) . This massive change in mutation rates is not due to changing amounts of the mismatch-binding MutS protein ( Figure S1 ) and a standard genetic method measuring rifampicin resistant colonies [14] ( Figure S2A ) shows a robust correlation with the microscopic method [14] , Figure S2B , Figure S3 ) . Damage to DNA replication , mismatch repair and other proteins assuring low mutation rates ( e . g . , dNTP pool biosynthesis and sanitation systems ) is expected to cause elevated mutagenesis . Indeed , spontaneous mutation rates correlate well with constitutive PC ( Figure 2C ) . Doubling PC correlates with over 100-fold increase in mutation rate . Eliminating mismatch repair by a mutH gene deletion increases the frequency of cells bearing MutL-CFP foci to ∼30% ( Figure 2A ) – a 100-fold mutator effect . However , over-expressing Tig , GroEL/ES and DnaK chaperones in mutH mutant reduces emerging mutations to only 2–5% of the cells , presumably by the improved DNA replication fidelity . Deletions of tig and dnaK in the mutH mutant increase further the mutation rate: ∼50% of cell population bears at least one new mutation as the result of reduced replication fidelity . A key question imposes itself: are the error-bearing and misfolded proteins immediately nonfunctional or malfunctional , and their oxidation is just an epiphenomenon or a tag for proteolysis , or can the misfolded proteins still function and their activity be rescued ( e . g . , by chaperones ) unless oxidized while in the misfolded state ? Only in the latter case do we expect antioxidants to rescue protein function and only if PC ( and/or some other parallel oxidative damage ) is the cause of mutations do we expect a decrease in mutation rates by antioxidant treatment . We found a striking parallelism between the decrease by 1 mM trolox of both constitutive PC and mutation rates ( p<0 . 0001 ) in the three strains ( Figure 3 ) . The level of suppression of intracellular ROS production ( about 40% , Figure 2B ) by trolox corresponds to the decrease in both PC and mutation rates ( Table S1 ) . The results in Figures 2 and 3 identify , for the first time , the principal cause of spontaneous mutations in bacterial cells . A decrease in spontaneous mutation rates is possible only by diminishing the dominant source of mutations , if there is a single one . The anti-mutator effect of reduced PC ( at constant ROS and DNA damage levels ) by two means ( chaperone overproduction and trolox ) identifies oxidative proteome damage as the principal determinant of spontaneous mutation rates . Clearly , perturbations in the native structure , probably exposing sensitive amino acid ( Lys , Arg , Pro , Thr ) side chains [16] , predispose proteins to oxidation . Since only a small number of carbonylated proteins can be detected on 2D western blots in E . coli [5] , [6] and human cells [17] , it seems that most native proteins have an evolved oxidation-resistant structure . Consistent observation of significantly higher PC levels in the mutH than in its wild type ancestor ( Figure 1A ) can be due to either some ongoing reversible effect of the mutH mutation or to the irreversible accumulation of mutations during propagation of mutH cells ( at its mutator rate of 0 . 3 mutations per cell generation ) that include oxidation-prone protein variants . Introduction of the mutH+ copy of the gene on a plasmid ( pBAD ) restored the wild type mutation rate but did not decrease the PC level ( not shown ) suggesting that the accumulated genomic mutations ( polymorphism ) are the likely cause of elevated PC . The detection of PC increase due to mutational protein polymorphism in mutH above the 100 times higher background of translation errors can be explained by the “monoclonality” of mutant proteins in the cell , while each random translation error affects only a single protein molecule . Given the small fraction of the E . coli proteome that is susceptible to carbonylation , a strong predisposition to carbonylation of few mutant high copy number proteins could increase the global PC level . Only those polymorphic protein changes that are neutral to cell fitness , or became neutral because of the buffering effect of chaperones [18] , will persist in a growing liquid culture . Indeed , the excess PC ( over the wild type level ) in the mutH mutator strain is subject to larger effects of chaperone deletions and over-expressions than in the wild type strain ( Figure 1A ) . Hence , by measuring PC , we can detect the consequences of random protein imperfections occurring during their biosynthesis and folding , as well as of recurrent specific protein imperfections encoded by mutant genes . UV light is probably the most ubiquitous extracellular source of mutagenic activity in nature generating high amounts of ROS ( Insert to Figure S4A ) . Highly mutagenic and cytotoxic UVC light is a model mutagen that we chose to test in strains with different chaperone activities and protein oxidation levels . Deletions of tig or dnaK increase the maximal PC levels ( Figure S4A ) whereas over-expression of either Tig , DnaK or GroEL/ES chaperone complex results in lower levels of PC at saturation than in wild type cells ( Figure S4B ) . If PC were also the key determinant of induced mutation rates , then the plateaus of PC at saturation by UVC light in different strains ( Figure S4 ) should coincide with plateaus of UVC-induced mutation frequencies . Since all damage-bearing mutational intermediates are not well recognized by mismatch repair proteins [14] , we exploited the high potency of UVC light in inducing mutations towards rifampicin resistance ( Figure 4A ) . UVC-induced mutation frequencies reach saturation at UVC doses that saturate also PC ( between 250 and 300 J/m2 , Figure S4 ) . E . coli with low fidelity ribosomal mutation rpsD14 and those deficient in Tig or DnaK reach a plateau at one order of magnitude higher mutation frequency ( 10−3 ) relative to the wild type ( 10−4 ) . Strains bearing high ribosomal fidelity rpsL141 mutation or over-expressing Tig , DnaK or GroEL/ES saturate at mutation frequencies of 10−6–10−5 . Chaperone deletions increase the susceptibility to UVC-induced mutations at all UVC doses tested , while chaperone over-expression causes a decrease in UVC-induced mutagenesis . Similarly to correlations of PC ( Figure 4B , zero UV ) with spontaneous mutation rates ( Figure 4A , zero UV ) in all studied strains , UVC-induced mutation frequencies extend from spontaneous mutation frequencies as parallel curves along the UVC dose range ( Figure 4A ) . It appears that the spontaneous oxidative proteome damage sensitizes cells to UVC-mutagenesis proportionally to the level of PC before irradiation ( Figure 4A ) . Plotting spontaneous and UVC-induced mutation frequencies versus PC in Figure 4C shows a positive correlation extending from spontaneous into UVC-induced mutation frequencies that saturate at the same UVC dose as PC in each strain . UVC-induced mutagenesis depends upon SOS response that induces synthesis of proteins required for the fixation of induced mutations , e . g . , the RecA and DNA polymerase V [19] . The correlation between UVC-induced mutagenesis and PC does not appear to involve the SOS system because ( i ) the non-inducible SOS repressor mutation lexA1 precludes UVC mutagenesis but does not alter significantly the UVC-induced PC ( Figure S5 ) and ( ii ) SOS induction ( measured by the expression of the sulA gene ) in DnaK deletion or over-expression strains cannot account for the observed differences in mutagenesis ( Table S3 ) . This means that SOS system does not control induced mutagenesis by controlling protein damage and that protein damage does not control SOS induction at UVC doses tested . However , the correlation between PC and UVC mutagenesis can be accounted for by the effect of PC on residual , unrepaired , DNA damage - the substrate for UV induced mutagenesis ( Figure 4 and 5 ) . In other words , UVC-induced mutagenesis correlates with PC because PC reduces repair efficiency of mutagenic DNA damage ( below ) . The spectra of sequenced mutations show that unrepaired DNA damage is the ultimate mutagenic intermediate in induced mutagenesis [20] . To evaluate the amount of UVC-induced oxidative DNA damage , we have measured 8-oxoguanine - one of the most mutagenic DNA lesions - in the eight strains , at several UVC doses , immediately after irradiation and after a dose-dependent post-irradiation recovery period . In contrast to both PC and mutation frequency patterns ( above ) , ( i ) 8-oxoguanine level in DNA measured immediately after irradiation increases linearly with UVC dose and does not saturate in the dose range applied ( Figure 5A ) and ( ii ) the UVC-induction of 8-oxoguanine in DNA is identical for all strains regardless of their chaperone activity and translational fidelity . Clearly , the induced mutation frequencies do not correlate with the DNA damage inflicted by UVC light . After the recovery period , 8-oxoguanine level decreases in all strains as the consequence of DNA repair ( Figure 5B ) . However , the extent of decrease varies: the rpsL141 mutant and strains over-expressing the chaperones display a larger decrease in the 8-oxoguanine levels , i . e . , more efficient repair , whereas rpsD14 mutant and strains with chaperone deficiencies show less repair of 8-oxoguanine than the wild type ( Figure 5B ) . To establish the relationship between the acute protein damage ( before and after exposure to UVC ) and the efficacy of the post-irradiation DNA repair , we have correlated the amount of unrepaired 8-oxoguanine to the level of PC present immediately after irradiation ( Figure 5C ) . Low initial levels of PC ( rpsL141 and chaperone over-expressions ) incurred by radiation correlate with efficient removal of the 8-oxoguanine during the post-irradiation recovery . Moreover , high constitutive levels of PC ( rpsD14 and chaperone deletions ) correlate with reduced post-irradiation repair of the 8-oxoguanine . Since the unrepaired damage matters for survival and mutagenesis , we also sought for correlations between the residual 8-oxoguanine level and the residual PC level after the recovery period . Figure 5D displays the relationship between these two parameters: lower levels of the residual ( post-incubation ) PC correlate with more efficient repair of the UVC-induced DNA damage leaving less residual 8-oxoguanine in the DNA of the ribosomal rpsL141 mutant and in the strains over-expressing chaperones than in wild type . High levels of residual PC ( in rpsD14 mutant and the strains with chaperone deficiencies ) correlate with high levels of 8-oxoguanine remaining in the genome reflecting reduced DNA repair resulting in elevated UVC-induced mutation frequencies . We conclude that whereas induced mutation frequencies do not correlate with the initial DNA damage inflicted by UVC , they do correlate with the residual , unrepaired , DNA damage resulting from PC related reduction in DNA repair . Hence the observed positive correlation between the cumulative ( spontaneous and UVC induced ) PC and induced mutagenesis . We have discovered potent phenotypic effects of the damaged proteome - including mutator and anti-mutator phenotypes - by manipulating exclusively the oxidative damage to proteins . Our finding that spontaneous mutation rates and UVC light-induced mutation frequencies are determined by the level of oxidative proteome damage - rather than by incurred DNA damage - is apparently paradigm breaking . However , only unrepaired copy errors ( measured here in vivo at single cell level ) and DNA damages are mutagenic , and we show that oxidative protein damage severely limits the efficacy of DNA repair proteome . Since reducing the physiological oxidative proteome damage by chaperone overproduction , or by antioxidant treatment , reduces spontaneous mutation rates below the wild type level , we identify oxidative proteome damage as the principal cause of spontaneous mutations in dividing bacteria . The correlation between mutation rates and protein oxidation , in the absence of mismatch repair , suggests that the precision of the DNA replication machinery is also governed by its oxidative damage . Furthermore , the frequency of induced mutations is determined by the cumulative ( spontaneous and UVC light-induced ) protein damage that appears as a bottleneck in the repair of UVC-inflicted DNA damage . The accumulated oxidative protein damage , whether caused by radiation-induced ROS [12] or by increased susceptibility to ROS [5]–[7 , this work] , is shown here to decrease the efficacy and precision of vital biosynthetic processes ( Figure 1 ) including DNA replication and repair ( Figure 5 ) , resulting in high mutation rates ( Figures 2–4 ) , cellular morbidity ( Figure 1 ) and bacterial [12] , [21] , [22] and animal [23] cell death . Therefore the present work may be relevant to the elucidation of the still obscure basic mechanisms of natural and drug-induced killing of bacteria [24] and tumor cells [25] . Progressive loss of many vital cellular functions and increased mutation rates are the characteristics of aging human cells that accumulate PC exponentially with person's age [26] similarly to the increase in the rates of age-related diseases and death . This raises an intriguing conceptual question: Since the phenotypes of oxidative proteome damage mimic aging , is aging the phenotype of increasing proteome damage ? Whereas PC is the best available biomarker of aging [26] , our work suggests that PC and other protein damage might be its most likely cause . It is conceivable that the described functional and mutagenic consequences of PC apply also to human cells . In that case , the oxidative proteome damage becomes a lead candidate for a common , preventable , underlying cause of aging and age-related diseases .
All strains used are listed in the Table S2 . They were derived from the sequenced wild-type E . coli MG1655 by P1 transduction and/or transformation . For all experiments , bacteria were grown in LB broth at 37°C to the exponential phase ( OD600 = 0 . 2–0 . 3 ) . It should be emphasized at this point that the double deletion mutants of E . coli were characterized by compromised growth and abundance of protein aggregates that prevented us from reliable determination of the mutation rates by the microscopic method we used . Bacteria were grown in LB broth at 37°C to the exponential phase ( OD600 = 0 . 2–0 . 3 ) , washed in 0 . 01 M MgSO4 , and concentrated five times . All radiation experiments were performed on ice at the dose rate of 4 . 5 J/m2·s−1 and 254 nm . Viable cell counts were estimated by plating serial dilutions on LB plates ( overnight at 37°C ) . In order to prevent photoreactivation , after UV treatment , all irradiated cells we kept in dark at 4°C until they were further processed . When necessary , a post-irradiation incubation was performed with its duration depending on the irradiation dose in the following manner: 110 minutes for 120 J/m2 , 248 minutes for 270 J/m2 and 330 minutes for 360 J/m2 . Overnight E . coli cultures were diluted 200-fold in LB medium supplemented with 1 mM trolox ( final concentration ) and grown to the mid-log phase ( OD 0 . 2 ) . According to the following experiment ( UV irradiation , single burst size , mutation rate determination ) , the cells were appropriately prepared ( described elsewhere in the Materials and Methods section ) . It is important to note that for irradiation experiments , the cells were prepared as described above , such that trolox was not present in the medium at the time of irradiation . E . coli cells , exponentially growing in a rich medium and irradiated , were pelleted by centrifugation immediately after irradiation and resuspended in the 10 mM PBS , pH 7 . 4 , supplemented with a mixture of protease inhibitors containing aprotinin bestatin , leupeptin , pepstatin A , E-64 and AEBSFxHCL , EDTA-free ( Pierce ) . Resuspended cells were frozen immediately in liquid nitrogen . Cells were broken by using a mechanical homogenizer and centrifuged 20 min at 12 , 000×g . Samples were then supplemented with 10 mg/100 µl lipid removal agent ( Sigma 13360-U ) , kept 1 hr at room temperature with shaking and centrifuged 15 min at 10 , 000×g . The amount of protein in the supernatant was measured by the Lowry method [27] . Protein extracts diluted to 10 µg/mL were loaded into wells ( Maxisorp , Nunc ) and incubated over night at 4°C to allow proteins to adsorb to the surface , followed by DHR derivatization of adsorbed proteins and detection of derivatized dinitrophenol ( DNP ) -carbonyl by a mouse DNP specific monoclonal antibody conjugated to HRP . Subsequent incubation with enzyme substrate 3 , 3′ , 5 , 5′-tetramethylbenzidine resulted in a colored product that was quantified using a microplate reader with maximum absorbance at 450 nm . Genomic DNA was isolated from E . coli strains by using Qiagen Genomic DNA Purification kit . 100 µL of genomic DNA was loaded onto the Maxisorp ( Nunc ) wells at the concentration of 1 µg/mL and incubated overnight at 4°C . This step was followed by an 1 hour incubation with the primary antibody mouse anti-8-deoxy-guanine and goat anti-mouse the secondary antibody conjugated to HRP . Subsequent incubation with enzyme substrate 3 , 3′ , 5 , 5′-tetramethylbenzidine ( TMB ) resulted in a colored product that was quantified using a microplate reader with maximum absorbance at 450 nm . Protein extracts were prepared as described above and a total of 10 ug of protein extracted from wild type E . coli MG1655 and isogenic Δtig , ΔdnaK , over-expression of GroEL/ES , Tig and DnaK strains were loaded on two SDS-PAGE gels with a 5% stacking and a 10% resolving gel . One gel was silver-stained . From the second gel , proteins were transferred onto a nitrocellulose membrane and visualized by using Amersham ECL Advance chemiluminescence detection system for imaging on autoradiographic film . Anti-MutS rabbit primary antibody ( courtesy of Dr . Robert Wagner , Genecheck , USA ) was used with a goat-anti-rabbit secondary antibody with horse-radish peroxidase . E . coli strains were labeled with 25 µM dihydrorhodamin-123 at indicated growth stage , as well as during UV irradiation . Cells were washed in minimal medium , and their fluorescence was measured ( Victor 3 , Perkin Elmer ) with excitation at 500 nm and emission at 530 nm . E . coli strains were grown to the mid-log phase ( OD600 = 0 . 2 ) in the presence of 1% maltose ( to induce high levels of λ receptor ) , pelleted and resuspended in 1 mL of LB broth supplemented with 30 mM MgSO4 , 15 mM CaCl2 , and 1% maltose ( final concentrations ) . To count the number of phages produced per single infected cell , cells were infected at multiplicity of infection of 0 . 3 and diluted to obtain one cell per three tubes that were incubated in 0 . 4 mL LB for 60 min at 37°C . A drop of chloroform was added to help release eventual intracellular viruses . The content of each tube was mixed with 0 . 7% top agar supplemented with overnight culture of E . coli MG1655 wild type , 30 mM MgSO4 , 15 mM CaCl2 and 1% maltose . Plaques were counted after overnight incubation at 37°C . The experiment was repeated three times . Individual numbers of phage particles obtained in each of the three repetitions were pooled together and a single mean was calculated with a standard deviation , for each of the studied strains . The ranges of single burst size from each of the three experiments are summarized in Table S4 . Strains studied by fluorescence microscopy are listed in Figure 1D . All strains were derived from the sequenced wild-type E . coli MG1655 by P1 transduction and transformation . Cells were grown on standard M9 minimal medium supplemented by 2 mM MgSO4 , 0 . 003% vitamin B1 , 0 . 001% uracil , 0 . 2% casamino acids , 0 . 01% glycerol and the required antibiotic , depending on the selection Overnight cultures of strains expressing fluorescent proteins were grown in supplemented minimal medium ( see above ) diluted 250-fold and re-grown to early exponential phase . Cells were concentrated and spread on agar with supplemented minimal medium , in a cavity slide to obtain a cell monolayer , as described previously [28] . The slide was mounted on Metamorph software ( Universal Imaging ) driven temperature-controlled ( Life Imaging Services ) Zeiss 200M inverted microscope . Images were recorded at 630-fold magnification using CoolSNAP HQ camera ( Princeton Instruments ) , in phase contrast and in fluorescence ( 50% neutral density filter on a 100 W Fluo-Arc Hg-vapor lamp ( Zeiss ) regulated to 100% power ) at wavelength of 500 nm during 20 seconds of exposure time . The images were analyzed by using Image J public domain software . The number of cells with MutL-CFP foci was counted manually . A total of 5000–6000 cells were examined for the wild type and MutH deficient E . coli in the absence of Tig and DnaK chaperones . 10 . 000 cells were examined for wild type and MutH deficient E . coli over-expressing Tig , GroEL/ES and DnaK . In order to determine the spontaneous mutation frequency , strains expressing fluorescent reporter CFP fused to wild-type MutL were grown overnight in supplemented minimal medium ( see above ) . The overnight cultures were then diluted 107-fold and grown to saturation . Dilutions of overnight cultures were plated on selective medium ( LB containing 100 µg/ml rifampicin ) to select rifampicin-resistant ( RifR ) colonies , and on LB to determine the total number of colony forming units . Colonies were scored after 24 h incubation at 37°C . The median of the mutation frequency of each strain was determined from nine independent experiments , each in triplicate . For the determination of UV-induced mutation frequencies , E . coli strains were grown overnight in LB medium . The overnight cultures were then diluted 200-fold until the culture reached an OD 0 . 2 . Cells were then irradiated with 120 , 270 and 360 J/m2 . After irradiation , cells were transferred into the LB medium and incubated at 37°C with shaking ( details described elsewhere in Methods ) . Dilutions were plated on selective medium ( LB containing 100 µg/ml rifampicin ) to select rifampicin-resistant ( RifR ) colonies , and on LB to determine the total number of colony forming units . Colonies were scored after 24 h incubation at 37°C . The median of the mutation frequency of each strain was determined from three independent experiments , each in triplicate . E . coli strains ( wild type , bearing a deletion and an overexpression of the DnaK chaperone ) containing a fusion of sulA and lacZ gene ( encoding β-galactosidase ) were grown until OD600 0 . 3 . An aliquot of cells was exposed to chloroform in the presence of β-merkaptoethanol and incubated for 5 minutes at room temperature . Ortho-Nitrophenyl-β-galactoside ( ONPG ) was added ( final concentration 0 . 6 mg/mL ) and the reaction mixture was mixed by shaking for few seconds . The reaction was stopped by adding Na2CO3 ( final concentration 0 . 5M ) . Tubes were centrifuged at 16 , 000 g for 10 minutes . OD420 was recorded in the supernatant and the units of β-galactosidase were calculated according to the formula: | Cellular life is maintained by the activities of proteins that , together , prevent molecular damage from occurring in the first place and repair damaged DNA , proteins and other damaged cellular components . Cellular fitness decreases due to the fact that these proteins are themselves subject to damage , leading to the progressive degeneracy of cellular functions due to diminishing protein activity and decreased precision . The ultimate liability to protein function is the irreversible oxidative protein modification , protein carbonylation . In our study , we have altered the intrinsic susceptibility of proteins to oxidative damage via alterations of translation fidelity and the accuracy of protein folding . We have found that the increased quality of proteome leads to an improved biosynthetic capacity of cells , as well as to decreased mutation rates . Since cellular aging can be defined as a progressive loss of nearly all vital cellular functions and an increase in mutation rates , this work suggests that oxidative proteome damage may be the most likely cause of aging and age-related diseases . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2013 | Phenotypic and Genetic Consequences of Protein Damage |
The circuit organization within the mammalian brainstem respiratory network , specifically within and between the pre-Bötzinger ( pre-BötC ) and Bötzinger ( BötC ) complexes , and the roles of these circuits in respiratory pattern generation are continuously debated . We address these issues with a combination of optogenetic experiments and modeling studies . We used transgenic mice expressing channelrhodopsin-2 under the VGAT-promoter to investigate perturbations of respiratory circuit activity by site-specific photostimulation of inhibitory neurons within the pre-BötC or BötC . The stimulation effects were dependent on the intensity and phase of the photostimulation . Specifically: ( 1 ) Low intensity ( ≤ 1 . 0 mW ) pulses delivered to the pre-BötC during inspiration did not terminate activity , whereas stronger stimulations ( ≥ 2 . 0 mW ) terminated inspiration . ( 2 ) When the pre-BötC stimulation ended in or was applied during expiration , rebound activation of inspiration occurred after a fixed latency . ( 3 ) Relatively weak sustained stimulation ( 20 Hz , 0 . 5–2 . 0 mW ) of pre-BötC inhibitory neurons increased respiratory frequency , while a further increase of stimulus intensity ( > 3 . 0 mW ) reduced frequency and finally ( ≥ 5 . 0 mW ) terminated respiratory oscillations . ( 4 ) Single pulses ( 0 . 2–5 . 0 s ) applied to the BötC inhibited rhythmic activity for the duration of the stimulation . ( 5 ) Sustained stimulation ( 20 Hz , 0 . 5–3 . 0 mW ) of the BötC reduced respiratory frequency and finally led to apnea . We have revised our computational model of pre-BötC and BötC microcircuits by incorporating an additional population of post-inspiratory inhibitory neurons in the pre-BötC that interacts with other neurons in the network . This model was able to reproduce the above experimental findings as well as previously published results of optogenetic activation of pre-BötC or BötC neurons obtained by other laboratories . The proposed organization of pre-BötC and BötC circuits leads to testable predictions about their specific roles in respiratory pattern generation and provides important insights into key circuit interactions operating within brainstem respiratory networks .
Rhythmic movements such as breathing and locomotion are produced in the nervous system by central pattern generator ( CPG ) networks consisting of interacting inhibitory and excitatory circuits that constitute the basic neural machinery generating motor behavior . Defining the structural-functional arrangement of CPG circuits remains a central problem in many neural motor control systems . In the present studies , we have addressed this problem of delineating the configuration of core circuits of the mammalian brainstem respiratory CPG . The basic pattern of respiratory activity includes three major phases: inspiration , post-inspiration , and late expiration [1 , 2] . This pattern is generated within bilaterally located medullary ventral respiratory columns ( VRC ) , each of which includes two key compartments , the pre-Bötzinger ( pre-BötC ) and Bötzinger ( BötC ) complexes , postulated to represent the core of the respiratory CPG [3–11] . The pre-BötC contains a heterogeneous population of excitatory neurons , including cells with intrinsic bursting properties , with mutual synaptic interconnections that generate rhythmic bursts of inspiratory activity and drive inspiratory motor output [12–17] . This structure also contains circuits of GABAergic and glycinergic inhibitory neurons [18–21] , which are active during the inspiratory phase and provide phasic inhibition that is widely projected to other brainstem compartments , including to BötC circuits to inhibit their expiratory neurons . The BötC mostly contains post-inspiratory and augmenting expiratory neurons that provide post-inspiratory and expiratory inhibition [22–25] . These neurons inhibit pre-BötC inspiratory neurons during expiration to coordinate expiratory and inspiratory phases of breathing . Although post-inspiratory neurons are predominantly located in the BötC , they are widely distributed within the VRC and present in other brainstem compartments including the pre-BötC [9 , 26 , 27] . The current experimental data suggest that generation and shaping of the basic respiratory pattern are based on the intrinsic bursting properties of , and interactions between , the excitatory neurons within the pre-BötC and involve inhibitory interactions between different neuron populations within and between the pre-BötC and BötC regions . The pre-BötC/BötC core circuitry is proposed to interact with other brainstem compartments , including several pontine structures [3 , 28–30] , the retrotrapezoid nucleus/parafacial respiratory group ( RTN/pFRG; [31–36] , and the nucleus tractus solitarii ( NTS , mediating multiple inputs and feedbacks to the respiratory network ) , which modulate and adjust operation of the core respiratory network and the pattern of generated respiratory activity to the current physiological and metabolic conditions [37–39] . The basic connectome of interactions among respiratory neuron populations in the pre-BötC and BötC compartments has been proposed in a series of computational models [3–8 , 10 , 11 , 40] . These models followed the key conceptual suggestions that ( 1 ) although the pre-BötC is capable of autonomous generation of rhythmic bursting when isolated in vitro , in more intact preparations and in vivo this structure is embedded in a larger respiratory network where its operation is modulated by interactions with other brainstem compartments , and particularly depends on mutual inhibitory interactions with the BötC , so that both the pre-BötC and BötC are critically involved in the generation and shaping of the normal ( eupneic ) three-phase rhythmic respiratory pattern , and ( 2 ) inhibitory interactions between and within these core compartments are fundamentally involved in generating the normal rhythmic respiratory neural activity . Accordingly , disruption of these interactions may cause a switch from the eupneic neural pattern of breathing defined by network interactions to either an intrinsic rhythmic activity originating within the pre-BötC , or to apnea or an abnormal apneustic activity . The above concept and the related computational models were indirectly supported by multiple experimental data [2 , 41–44] . However , this concept and , specifically , the critical role of inhibitory interactions in respiratory rhythm generation were challenged by a study from Janczewski et al . [45] , in which fast inhibitory neurotransmission was pharmacologically suppressed in anesthetized rats in vivo . Although these authors acknowledged the important role of inhibition in “shaping the pattern of respiratory motor output , assuring its stability , and in mediating reflex or volitional apnea” , they concluded that: ( 1 ) “postsynaptic inhibition within the pre-BötC and BötC is not essential for generation of normal respiratory rhythm in intact mammals” [45] , and ( 2 ) “bilateral ablation of BötC did not result in apnea” and did not change respiratory frequency and phase durations [45] and hence the BötC does not play a role in respiratory rhythm generation during normal eupneic breathing . Subsequently Sherman et al . [46] , who optogenetically targeted glycinergic pre-BötC neurons in mice to investigate their role in respiratory rhythm generation , similarly concluded that local inhibitory neurons are not essential for rhythm generation . However , the latter study was limited to optogenetic manipulation of only glycinergic neurons in the pre-BötC and did not investigate the potential role of the BötC and its interactions with the pre-BötC . In contrast , another recent study [47] showed that ( a ) optogenetic activation of BötC neurons resulted in depression of the respiratory rhythm , and ( b ) the effects of transient optogenetic stimulation of pre-BötC neurons depended on the phase of the respiratory cycle at which it was applied . Specifically , there was an insensitive post-inspiratory period during which the next inspiratory burst could not be triggered by optogenetic stimulation of the pre-BötC [47] . This insensitive period could result from a post-inspiratory inhibition originating in the BötC or within the pre-BötC . The results of these studies are consistent with the original concept of an important contribution of BötC inhibitory circuits and their interactions with pre-BötC circuits in respiratory rhythm and pattern generation . Resolving these issues requires systematic investigations using site-specific targeting of inhibitory neurons in the pre-BötC and BötC . In the present study , we thus employed an optogenetic strategy with a transgenic mouse line expressing Cre recombinase controlled by the vesicular GABA transporter ( VGAT ) promoter to express Channelrhodopsin-2 ( ChR2 ) in both GABAergic and glycinergic inhibitory neurons . Our experiments were performed with in situ perfused brainstem-spinal cord preparations from adult transgenic mice that readily enable site-specific laser stimulation of inhibitory neurons within the pre-BötC and BötC regions during electrophysiological recording for analysis of perturbations of circuit activity and motor output . Single laser pulses of various intensities delivered at different phases of the respiratory cycle as well as series of pulses or sustained laser stimulations were applied either to the pre-BötC or the BötC regions . The photo-induced perturbations were regionally-specific and were dependent on stimulus intensity , duration , and phase of photostimulation in the respiratory cycle . The results of these studies were consistent with our original concept confirming an important role of the BötC and of inhibitory circuit interactions between the pre-BötC and the BötC for operation of the respiratory CPG . To further examine and explain the results obtained in our study we revised our previous computational model of pre-BötC and BötC microcircuits by incorporating an additional population of post-inspiratory inhibitory neurons ( post-I ) in the pre-BötC which was based on the knowledge that these neurons are widely distributed in the VRC , including the pre-BötC [1 , 2 , 9 , 27 , 48] . The revised model was able to reproduce our experimental perturbations of circuit activity with sustained and phase-dependent short-duration activation of inhibitory neurons in the pre-BötC and BötC . The model was then further tested by its ability to reproduce previously published data showing the effects of site-specific photostimulation of neurons in the pre-BötC and BötC regions [47] , which were reproduced without any additional adjustments to the model parameters . The circuit configuration proposed in the model leads to several testable predictions about respiratory network organization and the specific roles of different brainstem neuron populations and how their interactions contribute to generating and shaping the neural breathing pattern .
We histologically validated the VGAT-Cre driver mouse line by verifying , via confocal fluorescence microscopy in fixed serial sections , Cre-driven tdTomato labeling of neurons in the pre-BötC and BötC ( Fig 1 ) in the double Tg line VGAT-tdTomato ( n = 5 ) , and by confirming co-labeling of glycine antibody ( Fig 2A ) or GABA antibody ( Fig 2B ) in tdTomato-expressing neurons within the pre-BötC and BötC in the VGAT-tdTomato line ( n = 2 each ) . We also confirmed tdTomato labeling of neurons distributed extensively throughout the medullary reticular formation , including in ventrolateral medullary respiratory regions caudal and adjacent to the pre-BötC ( e . g . , in the rostral ventrolateral respiratory group ( rVRG ) ( Fig 1 ) . We subsequently verified Cre-driven ChR2-EYFP expression in VGAT-positive tdTomato labeled neurons within the pre-BötC and BötC regions in the triple Tg line VGAT-tdTomato-ChR2-EYFP ( n = 3 ) by two-photon microscopy in “live” neonatal medullary slices in vitro . The ChR2-EYFP fusion protein was heavily expressed in processes and somal membranes of VGAT-tdTomato labeled neurons within the pre-BötC ( Fig 2C ) and BötC . To verify the efficacy of photostimulation ( 473 nm , 0 . 5–5 mW ) of ChR2-expressing pre-BötC inspiratory VGAT-positive neurons , we performed combined optogenetic stimulation and whole-cell patch-clamp recordings in rhythmically active in vitro neonatal medullary slice preparations from the VGAT-tdTomato-ChR2-EYFP mouse lines ( n = 3 ) . These slices effectively isolate the bilateral pre-BötC along with hypoglossal ( XII ) motoneurons to monitor inspiratory XII activity [17] , allowing whole-cell recordings from rhythmically active inspiratory pre-BötC neurons and simultaneous laser illumination to the recorded neuron ( Fig 3A ) . We performed current-clamp recording from tdTomato-labeled neurons , in which co-expression of ChR2 was confirmed with two-photon single optical plane live images ( Fig 3B ) , and identified rhythmically active pre-BötC inspiratory neurons , which were active in phase with integrated inspiratory XII nerve activity ( ∫XII ) . The membrane potential of these functionally identified VGAT-positive pre-BötC inspiratory neurons was depolarized by 10 . 0 ± 1 . 0 mV at 5 mW laser power ( n = 8 neurons ) with fast kinetics occurring within 12 ms , and recovery within 25 ms after terminating laser illumination ( Fig 3C ) . Summary data ( n = 8 neurons from 3 slice preparations , mean ± SEM ) shown in Fig 3D illustrates that ChR2-mediated depolarization of VGAT-positive inspiratory pre-BötC neurons was laser power-dependent , and importantly demonstrates that laser illumination caused membrane depolarization of VGAT-positive ( inhibitory ) inspiratory neurons in all cases with the applied laser power ranging from 0 . 5 to 5 mW . The observed membrane depolarization of these neurons could also include inhibitory effects on these neurons due to photostimulation of ChR2-expressing terminals from other VGAT-positive neurons with somas located outside of the pre-BötC providing inhibitory synaptic connections to the VGAT-positive neurons within the pre-BötC . However , our results indicate that these possible inhibitory effects were small , and therefore the optogenetic experimental results obtained in our study in slices in vitro were always attributed to excitatory effects ( i . e . , membrane depolarization ) on VGAT-positive inspiratory neurons within the pre-BötC ( as in Fig 3C ) . To investigate the role of inhibitory neurons within the pre-BötC in the generation and control of the respiratory pattern , we used site-specific optogenetic stimulations of VGAT expressing inhibitory neurons by single laser pulses of different durations ( 0 . 3–1 . 0 s ) and power ( 1 . 0 or 2 . 0 mW ) delivered bilaterally to both ( left and right ) pre-BötC regions at different phases of the respiratory cycle . The perturbations of respiratory neural activity induced by these photostimulations were dependent on the phase of application and the intensity of the applied laser pulses . Low intensity ( 1 . 0 mW ) single pulses of 300 ms duration delivered to the pre-BötC during the inspiratory phase did not affect the respiratory rhythm ( n = 9 , Fig 4A and 4G ) . In contrast , the same low-intensity stimuli outlasting the inspiratory phase ( n = 6 , Fig 4B ) or applied during the expiratory phase ( n = 6 , Fig 4C ) , induced rebound activation and thus an advanced onset of the next inspiratory phase . Only when the stimulus occurred very late during the expiratory phase and lasted into the next inspiratory phase , inspiration was delayed until the light stimulus was turned off ( n = 6 , Fig 4D ) . The dependence of the stimulation effect on the respiratory phase was abolished when stronger stimuli ( 2 . 0 mW ) where applied . The 2 . 0 mW laser pulses applied to the pre-BötC during the inspiratory phase reliably terminated the current inspiratory burst ( n = 9 , Fig 4G ) and initiated rebound activation of inspiration leading to an advanced onset of the next inspiratory burst ( n = 9 , Fig 4E ) . Similarly , such stronger stimuli delivered during the expiratory phase caused rebound activation and advanced onset of the next inspiratory burst ( n = 6 , Fig 4F ) . Rebound activation , whenever it occurred , showed a consistent latency of 300 ms on average that was independent of the stimulation intensity ( n = 6 , Fig 4H ) . To investigate how a sustained stimulation of pre-BötC inhibitory neurons affects the respiratory activity , we applied epochs ( duration 10–60 s ) of sustained laser illumination ( 20 Hz , 20 ms pulse trains , laser power of 0 . 5–5 . 0 mW ) allowing a recovery period ( > 4min ) after each epoch . Bilateral laser illumination of pre-BötC regions caused rapid and reversible perturbations of the respiratory frequency ( monitored by phrenic nerve ( PN ) or pre-BötC inspiratory activity recordings ) as a function of laser power ( 0 . 5–5 . 0 mW range ) . Relatively weak sustained stimulation ( 0 . 5–2 . 0 mW ) increased respiratory frequency ( Fig 5A for a representative example of 2 . 0 mW stimulation ) . With further increase in the stimulation intensity , the frequency started to decrease ( Fig 5D ) . The respiratory frequency was significantly reduced at 4 . 0 mW ( Fig 5B ) and the respiratory rhythm was abolished at a stimulation intensity of 5 . 0 mW ( Fig 5C ) . The relationship between the respiratory frequency and stimulation intensity is illustrated in Fig 5D ( n = 10 , mean ± SEM , **p < 0 . 01 , *p < 0 . 05 ) showing that bilateral activation of inhibitory neurons in the pre-BötC region affects the respiratory frequency in a photostimulation intensity-dependent manner . At relatively low stimulation intensity ( laser power of 0 . 5–2 . 0 mW ) frequency increased monotonically ( linear regression fit: slope = 27 . 63 ± 7 . 39% per mW , r = 0 . 80 ) , However , no significant increase of respiratory frequency was observed at 3 . 0 mW , and the stronger laser stimulation ( 4 . 0–5 . 0 mW ) caused a significant decrease of frequency with termination of respiratory activity for the duration of the photostimulation at the maximum laser power ( 5 . 0 mW ) . To investigate the involvement of the BötC inhibitory neurons in the generation and control of the respiratory pattern , we applied bilateral site-specific photostimulation of VGAT expressing BötC neurons . In contrast to the intensity- and phase-dependent effects of pre-BötC stimulations , the effect of single light pulses applied to the BötC terminated inspiratory activity with photostimulation at low laser power ( 1 . 0 mW , n = 5 , Fig 6 ) , independent of the phase of application . The respiratory activity was terminated for the entire duration of the photostimulation epoch ( Fig 6A–6D ) and a release of light stimulation induced rebound activation of inspiration after a fixed latency of approximately 300 ms ( Fig 6E ) , very similar to the rebound activation after stimuli applied to the pre-BötC . Sustained optogenetic activation of BötC ChR2-expressing inhibitory neurons ( 20 Hz pulses , 0 . 5–3 . 0 mW , 30–60 sec duration ) caused a rapid and reversible reduction of inspiratory frequency ( Fig 7A for a representative example of the effect of 2 . 0 mW stimulation ) with complete cessation of the rhythm at a maximum laser power of 3 . 0 mW ( Fig 7B ) . Fig 7C ( n = 12 , mean ± SEM , ** p < 0 . 01 ) illustrates that bilateral sustained activation of BötC inhibitory neurons monotonically reduced inspiratory frequency in a laser power-dependent manner ( 0 . 5–3 . 0 mW , linear regression fit: slope = -35 . 36 ± 2 . 28% per mW , r = -0 . 89 ) . To investigate the effects of photostimulation ( 20 Hz pulses , 0 . 5–3 . 0 mW ) in the BötC region on the activity of BötC and pre-BötC respiratory neurons , we recorded extracellular population activity from BötC post-I or aug-E type of expiratory neurons as well as pre-BötC pre-I/I type inspiratory neurons in perfused brainstem-spinal cord in situ preparations from adult VGAT-ChR2 mouse lines . Extracellular recordings from post-I expiratory neurons in the BötC ( n = 3 , Fig 8A ) showed a rapid and reversible augmentation , typically tonic excitation of neuron activity during laser application at ≥ 2 mW laser power to the BötC , which also caused reduction of inspiratory frequency and inhibition of inspiratory PN activity . In aug-E type of expiratory neurons in the BötC ( n = 4 , Fig 8B ) , photostimulation to the BötC similarly caused augmentation of neuron activity along with reduction of inspiratory frequency and inhibition of inspiratory PN activity . These results indicate that laser illumination ( 0 . 5–3 . 0 mW ) in the BötC region induced excitatory effects on expiratory VGAT-positive inhibitory neurons in the BötC . On the other hand , extracellular recordings from inspiratory ( pre-I/I ) neuron activity in the pre-BötC ( n = 6 , Fig 8C ) showed inhibition , typically complete silencing , of neuron activity during laser application to the BötC , which also caused a reduction of inspiratory frequency and inhibition of inspiratory PN activity . These results suggest functional inhibitory connections from BötC to pre-BötC circuits , consistent with the core respiratory network configuration proposed in our computational model ( see below ) . Previous models of the brainstem respiratory circuits [3 , 4 , 6–8 , 10] proposed neural interactions within and between the pre-BötC and BötC as shown in Fig 9A . In these models , the pre-BötC contained only populations of inspiratory neurons: the excitatory pre-inspiratory/inspiratory ( pre-I/I ) and the inhibitory early-inspiratory ( early-I ) populations . The BötC contained the inhibitory post-inspiratory ( post-I ) and augmenting expiratory ( aug-E ) populations . Hence , post-I neurons were suggested to be only present in the BötC but not in the pre-BötC . However , our analysis has shown that this architecture of the pre-BötC/BötC core network is unable to reproduce and explain the above experimental data on phase- and stimulation intensity-dependent effects of site-specific activation of inhibitory neurons within the pre-BötC . This conclusion was based on the following . First , optogenetic stimulation of the pre-BötC inhibitory neurons revealed a rebound activation of inspiration after the end of a sufficiently strong photostimulus ( Fig 4B–4F ) . The excitatory pre-I/I population could , in principle , generate a rebound activation , because of the persistent sodium-dependent mechanism . However , the pre-BötC compartment in previous models contained only the early-I inhibitory population ( Fig 9A ) which was active during inspiration only , did not inhibit pre-I/I neurons , and thus could not cause rebound activation of pre-I/I neurons . Second , we showed that the effect of low-intensity photostimulation of inhibitory pre-BötC neurons was suppressed during inspiration ( Fig 4A ) , but this suppression could be overcome by a stronger light activation ( Fig 4E ) . Together these observations suggest that the neurons responsible for termination of inspiration could be inhibited during the inspiratory phase unless the intensity of laser stimulation overcame this inhibition . Finally , additional activation of early-I inhibitory neurons in the previous models was known to increase respiratory frequency ( e . g . [10] ) and thus could not account for the biphasic shape of the frequency response to sustained light activation of inhibitory pre-BötC neurons ( Fig 5A ) . Further analysis of the previous models in the context of our new data led us to the suggestion that the above effects can be reproduced and explained by having post-inspiratory neurons located in both the BötC and the pre-BötC . Indeed , neurons with post-inspiratory firing characteristics have been previously found in the pre-BötC [9 , 26 , 27] . Therefore , we extended the previous models by incorporating an additional population of post-inspiratory inhibitory neurons in the pre-BötC and called it post-IpBC , ( Fig 9B ) . Except for its location , the proposed inhibitory post-IpBC population was similar to the post-I population in the BötC: it had adapting intrinsic properties and inhibited all other populations in the network except for its partner post-I population in the BötC . Like the BötC's post-I population , the post-IpBC population had mutual inhibitory interactions with the early-I population of the pre-BötC and the aug-E population of the BötC ( Fig 9B ) . To match model behavior to the experimental data , we assumed that the post-IpBC strongly inhibits the pre-I/I and other respiratory populations leading to rebound excitation of inspiration after ending the post-IpBC photostimulation ( Fig 4B–4D ) . The dependence of network behavior on the intensity of the applied photostimulation led us to the suggestion that the observed different ( and sometimes opposite ) consequences of weak vs . strong stimulations may result from mutual inhibitory interactions between the post-IpBC and early-I populations of the pre-BötC , so that the early-I inhibits post-IpBC at lower stimulation intensities , whereas post-IpBC overcomes this inhibition at higher stimulation intensities . Activation of ChR2 by photostimulation was implemented in the model as opening an additional excitatory "ChR2" channel in the populations of inhibitory neurons within the compartment ( pre-BötC or BötC ) stimulated by light . The light sensitivity of each inhibitory population was defined by the conductance of this channel ( gChR , which was the same in each inhibitory population , see Methods ) . The intensity of light stimulation was characterized by the parameter “stimChR“ . Incorporating the gChR , conductance in all inhibitory neurons allowed the above mutual inhibitory interactions between the post-IpBC and early-I populations . Specifically , at relatively low stimulation intensity during inspiration , the early-I population ( receiving excitation from the pre-I/I population ) kept the post-IpBC population inhibited , and thus did not allowing the applied stimulation to terminate inspiration . In this case , the early-I population became highly active and close to saturation of its activity . Consequently , an additional increase of stimulation intensity affected early-I activity less than that of the post-IpBC population . These different degrees of additional activation by photostimulation allowed the post-IpBC population to overcome early-I inhibition at higher stimulation intensity and to inhibit both early-I and pre-I/I populations , hence terminating inspiration . Network operation under normal conditions is shown in Fig 9C . During expiration , the BötC post-I population demonstrates adapting ( decrementing ) activity ( see Methods ) . The corresponding decline in inhibition from this population shapes the augmenting profile of aug-E activity . The reduction in post-I inhibition also produces slow depolarization of pre-I/I and early-I neurons in the pre-BötC . In addition , the pre-I/I population is depolarizing because of the slow deinactivation of the persistent ( slowly inactivating ) sodium current ( INaP , see Methods ) . At some moment during expiration , the pre-I/I population rapidly activates , providing excitation of early-I . The latter inhibits post-I , aug-E and post-IpBC populations , completing the switch from expiration to inspiration . During inspiration , the early-I population of the pre-BötC demonstrates adapting ( decrementing ) activity . The decline in inhibition from this population produces a slow depolarization of the post-I , aug-E and post-IpBC populations . At some moment , the post-I populations of BötC rapidly activates and inhibits both inspiratory neurons ( pre-I and early-I ) and the aug-E neuron ( initially ) , producing the switch from inspiration to expiration and the release of the fast adapting post-IpBC population in the pre-BötC . The model reproduced our experimental findings on the effects of selective optogenetic activation of inhibitory neurons in the pre-BötC ( Fig 10 ) . Similar to our experimental data , activating both inhibitory populations in the pre-BötC ( early-I and post-IpBC ) in the model by the simulated weak pulses ( 300 ms , stimChR = 1 ) did not terminate inspiration and did not affect the respiratory rhythm ( Fig 10A ) . The weak pulses induced rebound activation advancing the next inspiratory activity if the stimulus outlasted the inspiratory phase ( 1 s , stimChR = 1 , Fig 10B ) or was applied during the expiratory phase ( 300 ms , stimChR = 1 , Fig 10C ) . Also , inspiratory phase activity was delayed until the light stimulus was turned off , when the stimulus occurred very late during the expiratory phase and lasted into the next inspiratory phase ( 300 ms , stimChR = 1 , Fig 10D ) . Similar to our experimental data , stronger stimuli ( 300 ms , stimChR = 2 ) applied during the inspiratory phase terminated inspiratory activity ( Fig 10E ) . Strong stimuli during either the inspiratory or expiratory phase initiated rebound activation of inspiration and advanced the next inspiratory activity phase ( Fig 10E and 10F ) . As seen in Fig 10B–10F termination of the stimulus first produced a release of the aug-E population of the BötC , whose activity created a fixed latency for release of activity of the pre-BötC pre-I/I population that defines the inspiratory phase . The model also reproduced the biphasic frequency response of a sustained activation of inhibitory neurons within the pre-BötC . Sustained activation of early-I and post-IpBC inhibitory populations increased the frequency at low stimulation intensities ( stimChR ≤ 1 , Fig 11A ) , whereas further increase of stimulus intensity reduced the frequency ( stimChR > 1 , Fig 11B ) and finally terminated respiratory oscillations ( stimChR > 1 . 8 , Fig 11C ) . The intensity range of the sustained stimulation in the model was reduced by a factor of two ( Fig 11C ) , which is consistent with the level of inactivation under continuous light stimulation of ChR2 channels in the biological system [49 , 50] . Similar to our experimental studies , sustained activation of inhibitory neurons within the BötC ( aug-E and post-I ) by short ( 300 ms ) and long ( 5 s ) single pulses inhibited the respiratory rhythm for the duration of each pulse ( Fig 12 ) and the termination of each stimulus first produced a transient release of activity of the BötC aug-E population which delayed the onset of the next inspiratory phase with a latency ( Fig 12A–12D ) consistent with our experimental studies . Also , similar to the experimental results , a sustained activation of inhibitory populations within the BötC decreased the respiratory frequency with low stimulation intensities ( stimChR < 0 . 18 , Fig 13A and 13C ) and suppressed all activity with higher intensities ( stimChR ≥ 0 . 18 , Fig 13B and 13C ) , albeit at a lower intensity range when compared to the short pulses and to the pre-BötC stimulation . The computational model presented here reproduced our experimental findings on the effects of sustained and phase-dependent short-duration activation of inhibitory neurons in the pre-BötC . To validate the model performance further , we simulated the simultaneous optogenetic stimulation of all , inhibitory and excitatory , neurons within the pre-BötC and compared the modeling results with findings in a recently study of Alsahafi et al . [47] . They unilaterally activated pre-BötC or BötC neurons using an adeno-associated virus expressing ChR2 under the synapsin promoter to selectively photoactivate all ( excitatory and inhibitory ) neurons in the pre-BötC or the BötC [47] . Specifically , Alsahafi et al . [47] showed that the sustained stimulation of pre-BötC increased respiratory frequency ( Fig 14A , top diagram ) and that the network could be entrained by repetitive application of short light pulses ( Fig 14B , top diagram ) . To test the model against the experimental data published by Alsahafi et al . [47] we set the light sensitivity of the excitatory pre-I/I population equal to that of post-IpBC and early-I populations ( see Methods ) . Similar to the experimental results of Alsahafi et al . [47] , sustained stimulation of all populations in the pre-BötC increased respiratory frequency in our model at stimChR = 0 . 18 ( Fig 14A , bottom diagram ) . In the model , entrainment of the rhythm was possible down to a respiratory cycle period of 1 . 5 s with 300 ms pulses of stimChR = 0 . 2 to all populations of the model pre-BötC ( Fig 14B , bottom diagram ) . Alsahafi et al . [47] also demonstrated that optogenetic stimulation of pre-BötC neurons with short stimuli could induce a reliable reset of inspiratory activity but found a post-inspiratory period during which the rhythm could not be reset ( Fig 15A ) . The reset of inspiration in their experiments occurred with a short delay ( about 100 ms ) after the stimulus onset . The post-inspiratory period during which optogenetic stimulation of the pre-BötC was unable to initiate the next inspiratory event was about 200 ms ( Fig 15A ) . In our model , stimulating all neuron populations within the pre-BötC also initiated a reset of the inspiratory burst ( with a slightly longer latency of ~200 ms ) when stimulated at stimChR = 0 . 2 ( Fig 15B ) . The post-I period insensitive for inspiratory resetting was about 300 ms and was mainly defined by the activity of the post-I population of the BötC and the time constant of its adaptation . Our experimental and modeling results were also consistent with other data reported by Alsahafi et al . [47] . In some experiments , they used optogenetic excitation of BötC neurons which resulted in a decrease in respiratory frequency or complete suppression of inspiratory activity during the applied stimuli ( Fig 4B and 4C in their paper ) . This corresponds to our experimental ( Fig 7 ) and modeling ( Fig 11 ) results . Thus , our model reproduces multiple results of experimental perturbations that employed selective optogenetic activation of either inhibitory neurons within the pre-BötC and BötC or of all neuron types in each of these compartments .
We established a VGAT-tdTomato-ChR2-EYFP transgenic mouse line with Cre-driven ChR2 expression selectively in VGAT-positive neurons for temporally controlled , specific optogenetic manipulation of pre-BötC and BötC inhibitory neuron population activity . To validate this triple Tg line , we ( 1 ) verified the Cre-driver line ( VGAT-Cre ) used to derive this strain by tdTomato reporter expression , ( 2 ) performed immunocytochemical analyses of antibody co-labeling for glycine/GABA establishing that the majority of these tdTomato-labeled neurons were labeled by glycine or GABA antibodies , and subsequently ( 3 ) confirmed ChR2-EYFP fusion protein expression in the majority of VGAT-tdTomato labeled neurons in the pre-BötC and BötC regions as illustrated in Results ( Figs 1 and 2 ) . An important consideration for interpreting our experimental results is whether our transgenic approach to drive expression of ChR2 is efficient for site-specific photostimulation of inhibitory neuron depolarization , which is a basic assumption of our modeling studies . In the mouse line used , the VGAT Cre-driven expression of ChR2 occurs not only in neuronal soma but also in processes/terminals of inhibitory neurons . This could confound the interpretation of our experimental results , since photostimulation of inhibitory synaptic processes on targeted inhibitory neurons could inhibit postsynaptic activity rather than stimulate somatic spiking of these neurons . To address this problem and validate our approach , we recorded neuronal responses to photostimulation by somatic whole-cell recording in vitro , and we also performed extracellular recordings of neuronal activity in more intact in situ brainstem-spinal cord preparations . The responses to photostimulation were always depolarization of rhythmic inspiratory VGAT-expressing neurons in the pre-BötC in vitro , as well as augmented spiking of rhythmic inhibitory neurons recorded extracellularly . For example , expiratory neurons in the BötC always extended their spiking when the optical cannulae are positioned directly over this region in situ . If activation of ChR2 in terminals from inhibitory neurons located elsewhere , or within local circuits , that synapse on the inhibitory neurons in the BötC or pre-BötC regions was the predominant response to photostimulation , then we would expect to observe neuronal hyperpolarization or inhibition of neuronal spiking activity , which is clearly not the case . If photoactivation of inhibitory terminals is involved to some extent , it is certainly possible that the light-induced depolarization/spiking that we consistently observed may be somewhat attenuated , relative to the case where only ChR2 channels in neuronal somal membranes are activated . In general , however , our experimental tests indicate that with the VGAT-Cre transgenic approach , we achieve a net photo-induced depolarization/spiking of inhibitory neurons , which is consistent with the assumptions of the modeling . It should be noted that increasing the intensity of photosimulation of inhibitory pre-BötC neurons in slices did not result in a reduction or termination of the activity of VGAT-positive inspiratory neurons by a possible activation of some other ( non-inspiratory ) inhibitory neurons located within the pre-BötC ( e . g . , such as the pre-IpBC in our model ) at higher stimulation intensities ( as seen in Fig 3C and 3D ) , as could be expected considering our data obtained in situ ( Figs 4E , 4F and 5B–5D ) , as well as the proposed network interactions in the model ( see Fig 9B ) and the results of our simulations ( see Figs 10E , 10F and 11B–11D ) . Our explanation for this lack of inspiratory activity termination in vitro is that , in contrast to the more intact preparations , slices operate in very different physiological conditions engaging different rhythmogenic mechanisms [3 , 4 , 6 , 8 , 51 , 52] . Moreover , these slice preparations contain highly reduced respiratory circuitry , lacking many important structures , such as pontine circuits , which in the more intact preparations ( in situ and in vivo ) are known to provide excitatory drive to , and increased excitability of , post-inspiratory neurons [3 , 4 , 6 , 29 , 53–55] . Based on the results of our previous modeling studies [4 , 6 , 10] , all post-I neurons in such conditions ( without additional excitatory drives from the pons and other structures not present in the slice ) remain silent and cannot escape from inhibition provided by the early-I type inspiratory inhibitory neurons during applied photostimulation . To verify site-specificity of perturbations in the pre-BötC or BötC regions targeted for optical activation of VGAT-ChR2 expressing inhibitory neurons in situ , we systematically repositioned the optical cannula bilaterally in reticular formation sites outside of these regions . Laser illumination ( 2–5 mW ) in the rVRG immediately ( ~ 200 μm ) caudal to the pre-BötC region did not cause optical perturbations of the respiratory rhythm , consistent with the coordinates we used for bilateral positioning the optical cannulae in the pre-BötC from our extracellular recordings of respiratory neuron activity within the ventral respiratory column in the in situ experimental preparation . We accordingly positioned the cannulae more rostrally at our electrophysiologically established coordinates with predominant post-I and aug-E neuron activity to target the BötC . Furthermore , the perturbations produced by photostimulation at low laser intensities ( ≤ 2 . 0 mW ) where light scattering is reduced in the pre-BötC and BötC sites targeted were opposite , with augmentation of the inspiratory frequency with photostimulation in the pre-BötC region and inhibition of the rhythm with BötC regional photostimulation , indicating site-specificity of the perturbations . There has been a long-lasting debate concerning the role of the BötC and inhibitory circuit interactions within and between the pre-BötC and BötC for respiratory rhythm and pattern generation [2–4 , 6–8 , 42–46 , 56] . In this study , we have shown that selective optical activation of inhibitory neurons within either the pre-BötC or the BötC can terminate inspiration and reset the rhythm after a latency that according to our model may be defined by the activity of the inhibitory expiratory ( aug-E ) neurons located in the BötC . Our model also suggests that the post-inspiratory period insensitive to activation of all pre-BötC neurons [47] depends on the activity of post-I neurons in the BötC . Hence both types of inhibitory neurons in the BötC ( post-I and aug-E ) as well as the post-IpBC population in the pre-BötC could be specific targets of different inputs controlling breathing and , therefore , these populations may play an important role in respiratory pattern generation and control under various behavioral conditions . In previous compartmental models of the pre-BötC/BötC respiratory core network only one type of inhibitory post-I neurons was usually considered ( whether these neurons were called post-I or E-Dec or dec-E ) . The population of these neurons was assumed to be located within the BötC [3–8 , 10 , 11] , since this compartment does contain a larger number of neurons with post-inspiratory activity ( e . g . [23 , 25 , 57] ) . In the present model , we incorporated two different types of post-I neuron populations: a BötC post-I population with long decrementing activity during expiration and a post-IpBC with short transient post-inspiratory activity ( see Fig 9B and 9C ) . There are two issues here that require a more careful consideration and discussion . The first issue concerns the heterogeneity of the post-I neuron population . Two types of neurons with post-I activity patterns have been described: the post-I neurons of so-called E-DEC ( or dec-E ) type with activity decrementing over the entire , or a large part of , expiration [23 , 25 , 57–63] , and the post-I neurons with very short or transient firing during the first half of expiration ( during the post-inspiratory phase ) that were first described by Diethelm Richter [1 , 64] . The debated issue has been whether these two types of post-I neurons represent functionally the same or distinct neuron populations [23 , 25 , 65] . This issue remained unresolved although both types of these neurons were recorded , often in the same experiments ( e . g . , [26 , 66 , 67] ) and some earlier models included these two post-I neuron types ( dec-E and post-I ) as separate populations to perform different roles in the network [65] . Taking into account previous evidence and experimental and modeling data presented here , it is likely that post-I is not a single , functionally unique neuron population , but rather multiple ( at least two , as in our model ) populations that are involved in termination of inspiration or activated right after inspiration and have short or long lasting decrementing activity , including recently discovered post-I neurons with intrinsic bursting properties [68 , 69] . These different post-I populations may mediate different inputs ( e . g . from the pons and from NTS ) and/or contribute to termination of inspiration during different behavioral processes ( e . g . , during eating , drinking , swallowing , vocalization , etc . ) . The second issue concerns the location of neurons exhibiting post-I activity . Even though the majority of these neurons are located within the BötC [23 , 25 , 57] , different types of post-I/E-DEC neurons are widely distributed over the ventral respiratory column [9 , 70] and are present in the pre-BötC [26] . Moreover , based on some earlier investigations , the post-I neurons with short firing patterns are located more rostrally in the area that likely corresponds to the pre-BötC [25 , 66 , 71] , which is consistent with the model presented here . Selective optogenetic activation of inhibitory ( glycinergic only ) neurons in the BötC was previously performed by Sherman et al . [46] . In their in vivo experiments in mice , optical stimulations applied during inspiration always terminated inspiratory activity , whereas stimulations applied during the expiratory phase delayed the onset of the next inspiration . Sustained stimulations of glycinergic pre-BötC neurons produced apnea . At first glance , these results look contradictory to our results , but can be reconciled with our data in the following way . The in vivo preparations used by Sherman et al . [46] exhibited a respiratory rhythm that was considerably faster ( ~ 500 ms cycle period ) than what is reported here ( ~ 2 s ) for in situ preparations . They , however , reported a similar fixed latency between the end of the photostimulus and the beginning of the subsequent inspiratory burst ( ~ 300 ms ) . Taking into account an oscillation period of ~ 500 ms ( ~ 300 ms of expiratory phase ) and ~ 100 ms stimulus duration , this would naturally result in a phase delay ( which is what they reported ) and not in a phase advance which we found for the much slower respiratory rhythm in our experiments . This is further corroborated by the delay of inspiration in our experiments when stimuli were applied at the end of expiration ( see Figs 4D and 10D ) . Thus , in both Sherman et al . [46] and our experiments , the underlying mechanism can be a rebound activation of inspiration after a fixed latency that is independent of the respiratory frequency . Further , Sherman et al . [46] reported that optogenetic stimulation of glycinergic neurons terminated inspiration , which is consistent with our data and model simulations for high stimulation intensities ( laser power ≥ 2 . 0 mW in our experiments ) . The same is true for strong sustained photostimulation , where Sherman et al . [46] reported apnea for the duration of the stimulation ( Figs 5C and 11C ) . Biphasic responses to optical stimulations depending on laser intensity ( as those we observed ) could be absent in the study by Sherman et al . [46] because: ( 1 ) they only used “strong” light stimulations corresponding to ≥ 2 . 0 mW in our experiments; ( 2 ) they used viral transfection to express ChR2 which might lead to stronger expression and thus a stronger optogenetic excitation of neurons; or ( 3 ) the selective targeting of glycinergic neurons might change the balance of activation of specific populations depending on the transmitter phenotype composition in each respiratory population . For example , if the population that we labeled as post-IpBC were predominantly glycinergic while other populations have a higher proportion of GABAergic inhibitory neurons , then our model prediction would be very similar to the results shown by Sherman et al . [46] even under lower intensity stimulations . Interestingly , single unit recordings from various types of pre-BötC neurons in the Sherman et al . [46] study showed inhibition of these neurons during photostimulation , although there should be some inhibitory ( glycinergic ) neurons expressing ChR2 within pre-BötC that were activated by light to inhibit all other neurons . Our modeling studies suggest that the neurons active during optogenetic stimulation should be a population of post-I neurons located in the pre-BötC ( post-IpBC ) which strongly inhibits all other pre-BötC neurons ( Fig 10D–10F ) . The above analyses further support the important role of inhibition within the pre-BötC and between the pre-BötC and BötC . In a recent modeling study , Oku and Hülsmann [72] also analyzed the results published by Sherman et al . [46] using a computational model that reproduced the neuronal composition and connectivity within and between the pre-BötC and BötC from our previous models [3 , 4 , 6–8 , 10] . In line with our modeling studies , they concluded that sole activation of the inhibitory early-I population in the pre-BötC ( called I-DEC ) could not reproduce the Sherman et al . [46] results . They , however , did not add an additional post-I population to the pre-BötC as we did , but simply assumed that in Sherman et al . [46] experiments , glycinergic BötC neurons were also activated by light ( in addition to those in the pre-BötC ) due to wider light scattering and/or spreading of the virus driving expression of ChR2 into the BötC . The latter could also explain the lack of extracellular recordings from pre-BötC neurons activated by light stimulation in Sherman et al . [46] ( see above ) . In addition , the model of Oku and Hülsmann [72] could reproduce the termination of inspiration and rebound activation of the next inspiration by stimulating both pre-BötC and BötC inhibitory populations . However , their model did not reproduce the fixed latency before the inspiration onset observed in both Sherman et al . [46] and our present study . As described in Results , Alsahafi et al . [47] used site-specific optogenetic stimulation of pre-BötC and BötC neurons in vivo . They stimulated both inhibitory and excitatory neurons due to the viral vector construct with a general neuronal promoter ( synapsin ) employed . To test the ability of our model to reproduce their results , we used our model without adjusting any of the model parameters that were used to reproduce our data . As is , the model was able to reproduce many of their results , assuming simultaneous activation of inhibitory and excitatory neuron populations . Specifically: ( 1 ) sustained stimulation of the pre-BötC ( all pre-BötC populations in the model ) increased respiratory frequency ( Fig 14A ) , ( 2 ) respiratory activity could be entrained 1:1 by repetitive short light pulses activating the pre-BötC ( Fig 14B ) , ( 3 ) sustained activation of the BötC ( all populations in the model ) produced apnea ( Fig 11 ) , and ( 4 ) short phase-dependent simulations reset the respiratory rhythm except for a short post-inspiratory period during which the rhythm could not be reset ( Fig 15 ) . These simulations provided an additional validation of our model and further evidence for the concept that the BötC and the inhibitory interactions within and between the pre-BötC and BötC play a significant role in respiratory rhythm and pattern generation and control . Sufficiently strong optogenetic stimulation of inhibitory neurons either within the pre-BötC or the BötC suppressed rhythmic activity in the network for the duration of the stimulus and elicited a rebound activation of inspiration with a fixed latency of about 300 ms after the end of the stimulus ( Figs 4 and 6 ) . As discussed above , a similar latency has been reported for the photoactivation of glycinergic neurons in the pre-BötC by Sherman et al . [46] . Our modeling studies suggest that one possible explanation for this constant latency between the end of the photostimulation and the onset of the next inspiratory burst is a transient activation of the aug-E population and subsequent rebound activation of the pre-I/I population . An alternative explanation could be a delayed activation after inhibition of the pre-I/I population that is inherent to the pre-I/I population itself . Both options could be tested in future studies by either recording from aug-E neurons during photostimulation of the inhibitory pre-BötC neurons or by optogenetically inhibiting the excitatory pre-I/I population directly and analyzing rebound characteristics of this population . The main limitation of our experimental study is a lack of cellular-level recordings from all of the different neuron types within the pre-BötC or the BötC during different site- and respiratory phase-specific photostimulations . In the present studies , we have assessed from simultaneous extracellular recordings the perturbations of pre-BötC inspiratory population activity at the level of the inspiratory oscillator . This included our recordings of this activity during photostimulation of the BötC region , which reliably caused inhibition of the pre-BötC neurons , consistent with inhibitory input connections from the BötC . Our model simulations make specific predictions about activity perturbations of the different types of pre-BötC and BötC neurons and it remains an important problem to compare these predictions to neuronal activity recordings to further test the validity of the circuit interactions represented by the model . We also note that our transgenic approach with the VGAT promoter-driven expression of ChR2 in all inhibitory neurons did not allow us to separately investigate the roles of GABAergic and glycinergic neuron populations . We have previously proposed more complex model circuit configurations that make functional distinctions between GABAergic and glycinergic populations ( e . g . , [43] ) , and while there is a large population of neurons co-expressing GABA and glycine ( e . g . , [21] ) , it will be of considerable interest to employ transgenic approaches allowing selective manipulations of GABAergic or glycinergic neuron activity to possibly differentiate their functional roles in respiratory pattern generation . Moreover , additional tests of the inhibitory neuron circuit architecture and function that we propose should also be done using regional photoinhibition to assess activity perturbations associated with attenuation/loss of inhibitory circuit function . Some comparisons have been performed for neuronal activity recording data from studies where glycinergic postsynaptic inhibition was pharmacologically blocked [43] . It is likely that experiments utilizing VGAT promotor-driven expression of inhibitory opsins , or expression driven by other promoters ( e . g . , GlyT2 to target glycinergic neurons [46] ) , will provide new insights . There are certain limitations in the current modeling approach . Specifically , in contrast to some of our previous models [3 , 4 , 6 , 7] , we used a simplified description of neural population activity , which has computational and conceptual advantages [10 , 11] , but did not allow us to consider in detail the roles of neuronal recruitment and synchronization in each population under different conditions . This modeling approach also did not allow us to investigate the potential roles of intrinsic neuronal properties , known to be present in the respiratory neurons , such as different calcium , calcium-dependent potassium , and calcium-activated non-selective cationic currents , that are present in different respiratory neurons [14 , 15 , 73–76] . These issues can ultimately be addressed by implementing populations of neurons described by more detailed conductance-based models to represent the interacting populations within the respiratory CPG core circuits that we have proposed in the present study .
All animal procedures were approved by the Animal Care and Use Committee of the National Institute of Neurological Disorders and Stroke . For our optogenetic experiments photostimulating inhibitory neurons , we used Tg mice expressing Channelrhodopsin-2 ( ChR2 ) and reporter fluorescent proteins in glycinergic and GABAergic neurons driven by Cre recombinase expressed in these neurons under control of the vesicular GABA transporter ( VGAT ) promoter [77–79] . To produce this mouse line , we first crossed a Slc32a1tm2 ( Cre ) Lowl knock-in Cre-driver mouse line ( VGAT-ires-Cre , the Jackson Laboratory , Bar Harbor , ME ) with a Cre-dependent reporter mouse strain [B6;Cg-Gt ( ROSA ) 26Sortm9 ( CAG-tdTomato ) Hze , Rosa-CAG-LSL-tdTomato-WPRE , the Jackson Laboratory] to obtain offspring expressing red fluorescent protein variant tdTomato in Cre-expressing VGAT-positive neurons . This double Tg line ( VGAT-tdTomato ) was analyzed histologically to verify tdTomato expression in glycinergic and GABAergic neurons labeled by antibodies ( below ) , and subsequently crossed with a Cre-dependent optogenetic mouse strain [B6;129S-Gt ( ROSA ) 26Sortm32 ( CAG-COP4*H134R/EYFP ) Hze , Rosa-CAG-LSL-ChR2 ( H134R ) -EYFP-WPRE , the Jackson Laboratory] to obtain a triple Tg mouse line ( VGAT-tdTomato-ChR2-EYFP ) , in which Channelrhodopsin-2 ( ChR2 ) -EYFP fusion proteins are specifically expressed in VGAT-Cre positive neurons . Fluorescence immuno-labeling with glycine and GABA antibodies was used to verify glycine and GABA expression in VGAT-tdTomato neurons within the pre-BötC and BötC in immersion or transcardially perfusion fixed brainstem tissue sections from neonatal and mature VGAT-tdTomato double Tg mouse lines . The medulla oblongata used for histological analyses was fixed in 4% paraformaldehyde ( wt/vol ) in phosphate-saline buffer , and cryoprotected overnight at 4°C in 30% sucrose , 0 . 1 M PBS solution and sectioned coronally or parasagittally at 30 or 50 μm with a freezing microtome . For fluorescence immunohistochemistry , floating sections were incubated with 10% donkey serum in PBS with Triton X-100 ( 0 . 3% ) and subsequently incubated for 48–72 hours at room temperature with primary antibodies for GABA ( chicken anti-GABA , 1:300 , ab62669 , Abcam , Cambridge , MA ) , for glycine ( rabbit anti-glycine , 1:5000 , IG1001 , ImmunoSolution; MP Biomedicals , Solon , OH ) , and also for choline acetyltransferase ( ChAT ) ( goat anti-ChAT , 1:200 , AB144P , EMD Millipore , Billercia , MA ) to label motoneurons . Individual sections were then rinsed with PBS and incubated for 2 hours with secondary antibodies ( donkey anti-rabbit Alexa Fluor 647 , 1:500 for glycine labeling; donkey anti-chicken Alexa Fluor 647 , 1:500 for GABA labeling; donkey anti-goat Alexa Fluor 488 , 1:500 for ChAT , Jackson ImmunoResearch , West Grove , PA ) . Individual sections were mounted on slides and covered with an anti-fading medium ( Fluoro-Gel; Electron Microscopy Sciences , Hatfield , PA ) . Fluorescent labeling of neurons was visualized with a laser-scanning confocal imaging system ( LSM 510 , Zeiss , Oberkochen , Germany ) or two-photon microscopy ( TCS SP5 II MP , Leica , Buffalo Grove , IL ) . All images were color/contrast enhanced and adjusted with a thresholding filter in Adobe Photoshop . We validated specificity of the glycine and GABA antibodies used by analyzing antibody labeling in transgenic ( Tg ) mouse lines , in which fluorescent proteins ( tdTomato or GFP ) are expressed in excitatory or inhibitory neurons by the cell type-specific promoters VgluT2 for glutamatergic , GlyT2 for glycinergic , and GAD67 for GABAergic neurons . We verified glycine antibody co-labeling in all of the GlyT2-tdTomato positive glycinergic neurons within the pre-BötC and BötC regions in our GlyT2-tdTomato Tg mouse line , which was obtained by crossing a GlyT2-Cre line [B6 . FVB ( cg ) -Tg ( Slc6a5-cre ) KF109Gsat/Mmucd , MMRRC , UC Davis] with a Cre-dependent tdTomato reporter strain [B6 . Cg-Gt ( ROSA ) 26Sortm9 ( CAG-tdTomato ) Hze/J , Jackson Laboratories] . GlyT2 has been regularly used as a specific promotor for glycinergic neurons in the field ( e . g . , [19 , 46] ) . We also verified GABA antibody co-labeling in all of the GAD-67 positive GABAergic neurons , but no labeling in VgluT2-tdTomato positive neurons within the pre-BötC and BötC regions in the triple Tg mouse line ( VgluT2-tdTomato-GAD67-GFP ) , which was obtained by crossing the VgluT2-tdTomato mouse line with a GAD67-GFP knock-in mouse line ( [80] ) as we have previously reported ( [81] ) . In this Tg mouse line , we also verified no glycine antibody labeling in VgluT2-tdTomato positive neurons ( [81] ) . In addition , we verified the same labeling patterns of the pre-BötC and BötC neurons with different GABA antibodies [i . e . , chicken anti-GABA ab62669 , Abcam reported in this study , and rabbit anti-GABA A2052 , Sigma ( e . g . , [21 , 82] ) ] . We also note that the VGAT Cre-driver mouse line used in our study ( VGAT-ires-Cre , the Jackson Laboratory ) has been validated with in situ hybridization analyses for VGAT mRNA , which has verified that Cre activity is expressed in all of VGAT mRNA-positive neurons analyzed ( [83] ) . We performed combined optogenetic stimulation and whole-cell patch-clamp recordings with rhythmically active in vitro slice preparations ( 300–400 μm thick ) of the medulla oblongata that were cut from neonatal [postnatal day 3 ( P3 ) to P8] triple Tg mouse line ( VGAT-tdTomato-ChR2-EYFP ) of either sex ( Fig 3A ) . These slices contain the active bilateral pre-BötC and rostral end of the hypoglossal motor nucleus ( XII ) with intact XII nerve rootlets for recording inspiratory motor output ( [21] ) . The slice was superfused ( 4 ml/min ) in a recording chamber ( 0 . 2 ml ) mounted on the stage of an upright laser scanning microscope with artificial cerebrospinal fluid ( ACSF ) containing the following ( in mM ) : 124 NaCl , 25 NaHCO3 , 3 KCl , 1 . 5 CaCl2 , 1 . 0 MgSO4 , 0 . 5 NaH2PO4 , 30 D-glucose equilibrated with 95% O2 and 5% CO2 ( pH = 7 . 35–7 . 40 at 27°C ) . During experiments , rhythmic respiratory network activity was maintained by elevating the superfusate solution K+ concentration to 8–9 mM . Electrophysiological signals recorded from XII nerve rootlets with fire-polished glass suction electrodes ( 50–100 μm inner diameter ) were amplified ( 50 , 000–100 , 000X; CyberAmp 380 , Molecular Devices Sunnyvale , CA ) , band-pass filtered ( 0 . 3–2 kHz ) , digitized ( 10 kHz ) with an AD converter ( PowerLab , AD Instruments , Inc . , Colorado Springs , CO ) , and then rectified and integrated digitally with Chart software ( AD Instruments ) . Whole-cell current-clamp data were recorded with an EPC-9 patch-clamp amplifier ( HEKA Electronics Inc . , Mahone Bay , Nova Scotia , Canada ) controlled by PatchMaster software ( HEKA; 2 . 9 kHz low-pass filter , sampled at 100 kHz ) . Whole-cell recording electrodes ( borosilicate glass pipette , 4–6 MΩ ) , positioned with 3-dimensional micromanipulator ( Scientifica ) , contained the following ( in mM ) : 130 . 0 K-gluconate , 5 . 0 Na-gluconate , 3 . 0 NaCl , 10 . 0 HEPES , 4 . 0 Mg-ATP , 0 . 3 Na-GTP , and 4 . 0 sodium phosphocreatine , pH 7 . 3 adjusted with KOH . In all cases , measured potentials were corrected for the liquid junction potential ( -10 mV ) . Series resistance was compensated on-line by 80% , and the compensation was periodically readjusted . Optical live imaging of tdTomato-EYFP labeled neurons in the slices was performed with a Leica multi-photon laser scanning upright microscope ( TCS SP5 II MP with DM6000 CFS system , a 20X water-immersion objective , N . A . 1 . 0 , Leica ) . A two-photon Ti:Sapphire pulsed laser ( MaiTai , Spectra Physics , Mountain View , CA ) was used at 800–880 nm with DeepSee predispersion compensation . Combined photostimulation and electrophysiological recording experiments were performed with in situ arterially perfused brainstem-spinal cord preparations ( Fig 16A ) from the mature triple Tg mouse line ( VGAT-tdTomato-ChR2-EYFP ) of either sex ( 20 – 30g , 2–4 months old ) , in which complex cellular and circuit interactions involving the pons , BötC , pre-BötC and rVRG generate a normal 3-phase rhythmic respiratory neural activity pattern [6 , 84] . Preheparinized ( 1 , 000 units , given intraperitoneally ) mice were anaesthetized deeply with 5% isoflurane until loss of the paw withdrawal reflex , and the portion of the body caudal to the diaphragm was removed . The head and thorax were immersed in ice-chilled carbogenated ACSF solution containing the following ( in mM ) : 1 . 25 MgSO4 , 1 . 25 KH2PO4 , 5 . 0 KCl , 25 NaHCO3 , 125 NaCl , 2 . 5 CaCl2 , 10 dextrose , 0 . 1785 polyethylene glycol . The brain was decerebrated at a precollicular level , and the descending aorta , thoracic phrenic nerve and cervical vagus nerves were surgically isolated . The dorsal brainstem was exposed by craniotomy and cerebellectomy . The preparation was transferred to a recording chamber and secured in a stereotaxic head frame with dorsal side up . The descending aorta was cannulated with a double lumen catheter ( DLR-4 , Braintree Scientific , Braintree , MA ) for ACSF perfusion with a peristaltic roller pump ( 505D , Watson-Marlow , Wilmington , MA ) and for recording of perfusion pressure with a pressure transducer . The ACSF perfusate was gassed with 95% O2-5% CO2 and maintained at 31°C . Rocuronium bromide ( 2–4 μg/ml; SUN Pharmaceutical Industries , Cranbury , NJ ) was added to the perfusate to block neuromuscular transmission . Throughout the experiments , the perfusion pressure was maintained between 60 and 80 mmHg with vasopressin application ( 200–400 pM as required; APP Pharmaceuticals , Los Angeles , CA ) and by adjusting the perfusion pump speed to avoid the possible effects of pressure changes on respiratory activity . To monitor respiratory network activity and motor output , we recorded inspiratory activity from phrenic nerves ( PN ) and/or in some cases extracellular population activity from pre-BötC inspiratory neurons as described previously [81] that allowed us to directly analyze activity perturbations at the level of the inspiratory rhythm generator . We also recorded extracellular population activity from BötC post-I or aug-E type of expiratory neurons to verify photostimulation of these BötC inhibitory neurons by laser illumination in the BötC region . Extracellular population activity from the pre-BötC and BötC ( Figs 8 and 16 ) was recorded by a dorsal approach with a fine-tipped glass pipette ( 3–5 MΩ resistance ) filled with 0 . 5 M sodium acetate ( Sigma ) positioned by a computer-controlled 3-dimensional micromanipulator ( MC2000 , Märzhäuser ) based on our established anatomical coordinates for the pre-BötC and BötC [81] . Electrophysiological signals were amplified ( 50 , 000–100 , 000X , CyberAmp 380 , Molecular Devices , Sunnyvale , CA ) , band-pass filtered ( 0 . 3–2 kHz ) , digitized ( 10 kHz ) with an AD converter [Cambridge Electronics Design ( CED ) , Cambridge , UK] , and then rectified and integrated digitally with Spike2 software ( CED ) . Laser illumination for photostimulation experiments was performed with a blue laser ( 473 nm; OptoDuet Laser , IkeCool , Los Angeles , CA ) , and laser power ( 0 . 5–5 . 0 mW ) was measured with an optical power and energy meter ( PM100D , ThorLabs , Newton , NJ ) . Illumination epochs were controlled by a pulse stimulator ( Master-8 , A . M . P . I . , Jerusalem , Israel ) . To apply the laser during the specific respiratory phase ( i . e . , I-phase or E-phase ) , the onsets of inspiratory phrenic activity were detected as the times the integrated PN signals crossed a defined threshold ( normally 20% of the peak amplitude ) , and photostimulation epochs controlled by computer-triggered TTL pulses were delivered after pre-programmed delays from the detected onset of PN activity . Optical fiber ( s ) , from a bifurcated fiberoptic patch cable , each terminated by an optical cannula ( 100 μm diameter , ThorLabs ) were positioned unilaterally on the surface of the pre-BötC in the in vitro rhythmically active slice preparations ( Fig 3A ) , and implanted bilaterally to a depth immediately dorsal to the pre-BötC or BötC in the brainstem-spinal cord preparations in situ ( Fig 16A ) . Positioning of the optical fibers by micromanipulators was based on coordinates for the pre-BötC and BötC from our extracellular recordings of respiratory neurons within the ventral respiratory column in the present and previous combined electrophysiological and optogenetics-based studies [81] . We conducted control experiments with in situ preparations from the VGAT-tdTomato ( non-ChR2-expressing ) Tg mouse strains ( n = 2 ) , and found that no light-induced perturbations of the respiratory frequency or burst amplitude of inspiratory PN activity with the optical cannula positioned bilaterally in the pre-BötC or the BötC . All digitized electrophysiological signals were analyzed by automated procedures to extract respiratory parameters from integrated PN or pre-BötC inspiratory activities , performed with Matlab ( R2016a ) and Python ( 3 . 5 ) software utilizing the computational resources of the NIH HPC Biowulf cluster . Inspiratory events were detected from the smoothed integrated PN signals via a 200 ms window moving average and peak detection algorithm based on detecting ridges in the continuous wavelet transform with appropriate peak width boundaries for each recording . Following peak detection , interburst interval was computed to obtain the respiratory frequency . Inspiratory amplitude was calculated as the difference between peak value and the local baseline voltage , which was calculated as the value corresponding to the peak of the Gaussian kernel density estimation in a local 10 second window . Latencies were calculated as the time difference between the end of the light stimulus and the onset of the next PN inspiratory burst , detected by the integrated PN signals crossing the threshold at 20% of the peak amplitude as described above . For statistical analysis ( significant p-value < 0 . 05 ) , experimental conditions were compared with a one sample Student’s t-test . Summary data are presented as means ± SEM . The schematic of the model of the core pre-BötC and BötC microcircuits developed in this study is shown in Fig 9B . The model is based on network interactions within pre-BötC and BötC proposed in a series of previous models [3–8 , 10 , 11] and specifically represents an extension of the model of Rubin et al . [10] ( Fig 9A ) . Synaptic interactions between neuron populations included in that earlier model had been justified in previous publications [3 , 4 , 6] . To reproduce the experimental data presented in this study we extended the Rubin et al . [10] model ( Fig 9A ) by incorporating an additional population of post-inspiratory inhibitory neurons in the pre-BötC ( post-IpBC , Fig 9B ) . The reasons for including this additional post-I population are described in the Results section and further justification for the presence of two separate post-I populations , including the one within the pre-BötC , is presented in the Discussion . All connections of the incorporated post-IpBC population were implemented the same as the connections of the post-I population in the BötC . Weights of synaptic connections in the present model have been modified from the original Rubin et al . [10] model to match our experimental data and/or to accommodate the inclusion of the additional post-IpBC , population but not altered otherwise . The model includes five neuron populations ( i = 1 , 2 , 3 , 4 , 5 ) , three of which are located in the pre-BötC ( the excitatory pre-inspiratory , pre-I/I , i = 1 , the inhibitory early-inspiratory , early-I , i = 2 , and the novel inhibitory post-IpBC , i = 5 ) and two inhibitory populations in the BötC ( the augmenting-expiratory , aug-E , i = 3 , and the post-I , i = 4 ) ( Fig 9B ) . Each neural population is described by an activity-based model , in which the dependent variable Vi represents an average voltage for population i and each output fi ( Vi ) ( 0 ≤ fi ( Vi ) ≤ 1 ) represents the average or integrated population activity at the corresponding average voltage [10 , 85] . This description allows an explicit representation of ionic currents , specifically of the persistent ( slowly inactivating ) sodium current INaP [10] responsible for state-dependent generation of intrinsic bursting activity in the pre-BötC [86–88] . Because we consider regimes in which neurons within each population switch between silence and active spiking in a generally synchronized way , we assume that the dynamics of the average voltages in the model can be represented by a conductance-based framework but without fast membrane currents responsible for spiking activity , which tremendously simplifies the model description and analysis [10] . While such a simplified activity-based description of neuron populations allows for the qualitative reproduction of experimental data , it does not imply an exact correspondence between experimental and modeling results with regard to single neuron activity and synaptic interactions . The membrane potential ( V1 ) of the excitatory pre-I/I population ( i = 1 ) with INaP-dependent intrinsic oscillatory properties obeys the following differential equation: C∙dV1dt=−INaP−IK−IL1−ISynE1−ISynI1−IChR1 , ( 1 ) where IK represents the potassium delayed rectifier current included in the pre-I/I population . The other four populations do not have INaP and IK , but have adaptive properties defined by the outward potassium currents IADi . For each of these populations ( i ∈ {2 , 3 , 4 , 5} ) , the average membrane potential is described as: C∙dVidt=−IADi−ILi−ISynEi−ISynIi−IChRi . ( 2 ) In Eqs ( 1 ) and ( 2 ) , C is the neuronal membrane capacitance; ILi is the leak current; ISynEi and ISynIi are the excitatory and inhibitory synaptic currents , respectively , and IChRi is the excitatory ChR2 current . The currents are described as follows: INaP=g¯NaP∙mNaP∙hNaP∙ ( V1−ENa ) ; IK=g¯K∙mk4∙hNaP ( V1−EK ) ; IADi=g¯AD∙mADi∙ ( Vi−EK ) ; ILi=g¯L∙ ( Vi−EL ) ; ISynEi=g¯SynE∙drivei∙ ( Vi−ESynE ) , fori≠2; ISynE2=g¯SynE∙ ( drive2+a12∙f1 ( V1 ) ) ∙ ( V2−ESynE ) ; ISynIi=g¯SynI∙∑j=2j≠i5bji∙fj ( Vj ) ∙ ( Vi−ESynI ) ; IChRi=g¯ChRi∙stimChRi ( t ) ∙ ( Vi−EChR ) , ( 3 ) where IX is the current , g¯X is the maximal conductance and EX is the reversal potential of the corresponding channel x; a12 defines the weight of the excitatory synaptic input from the pre-I/I to the early-I population; bji defines the weight of the inhibitory input from population j to population i ( i ∈ {1 , 2 , 3 , 4 , 5} , j ∈ {2 , 3 , 4 , 5} ) ( Fig 9B ) ; drivei defines the excitatory drive to population i; stimChRx ( t ) represents the laser intensity of the ChR2 stimulation . The nonlinear function fi ( Vi ) defines the output activity of each population unit ( indirectly representing the rate of spiking activity ) fi ( Vi ) ={1/ ( 1+exp[− ( Vi−V1/2 ) /kVi] ) , ifVi≥−60mV;0 , ifVi<−60mV , ( 4 ) where V1/2 is the half-activation voltage and kVi defines the slope of the output function for each population i . Two types of slow variables are implemented in the model . One variable ( hNaP ) represents the slow inactivation of the persistent sodium current [86] in the pre-I/I population; the other variables ( mADi , i∈{2 , 3 , 4 , 5} ) describe the level of adaptation in the other populations: τhNaP ( V1 ) ∙ddthNaP=h∞NaP ( V1 ) −hNaP , τADi∙ddtmADi=kADi∙fi ( Vi ) −mADi . ( 5 ) Voltage dependent activation and inactivation variables and time constants for the persistent sodium and potassium rectifier channels in pre-I/I population are described as: mNaP=1/ ( 1+exp[ ( −V1+40 ) /6] ) ; h∞NaP=1/ ( 1+exp[ ( V1+48 ) /6] ) ; τhNaP=τhNaPmax/cosh[ ( V1+48 ) /12]; mK=1/ ( 1+exp[ ( −V1+30 ) /4] ) . ( 6 ) Model parameters were not based on direct experimental measurements , but adapted from the model of Rubin et al . [10] and other previous models [38 , 40 , 89] and modified/adjusted to accommodate the incorporation of the additional post-IpBC population . These adjustments were done carefully to allow , on one hand , reproduction of major features of the previous models ( and related experimental data ) and , on the other hand , to qualitatively reproduce the experimental data presented here . The following parameter values were used: Simulations were performed using Matlab R2015b . Differential equations were solved using a variable order multistep differential equation solver ode15s available in Matlab . | We investigated the organization of key circuits within the brainstem respiratory central pattern generator ( CPG ) responsible for generating the basic breathing pattern in mammals . Our study focused on the organization and interactions of two core components of this CPG: the pre-Bötzinger ( pre-BötC ) and Bötzinger ( BötC ) complexes . We used genetically transformed mice with inhibitory neurons containing light activated membrane channels , allowing us to investigate perturbations of respiratory activity during site-specific activation of pre-BötC or BötC inhibitory neurons by applying laser photostimuli of different intensities and durations at different phases of the respiratory cycle . Photostimulation of inhibitory neurons caused major disturbances of circuit activity . The obtained experimental results were used to extend and refine a computational model simulating circuit interactions within and between the pre-BötC and BötC . A major feature of the model is a novel inhibitory post-inspiratory population within the pre-BötC and its interactions with other populations in the network . We validated the model by successfully testing its ability to reproduce our experimental results as well as results of previous studies that employed photoactivation of pre-BötC or BötC neurons . Our study provides important insights into the circuit organization in the brainstem respiratory CPG with implications for other circuits controlling rhythmic behaviors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"optogenetics",
"chemical",
"compounds",
"neural",
"networks",
"engineering",
"and",
"technology",
"aliphatic",
"amino",
"acids",
"lasers",
"brain",
"light",
"neuroscience",
"organic",
"compounds",
"electromagnetic",
"radiation",
"brain",
"mapping",
"amino",
"acids",
"bioassays",
"and",
"physiological",
"analysis",
"optical",
"equipment",
"neurotransmitters",
"research",
"and",
"analysis",
"methods",
"glycine",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"proteins",
"chemistry",
"light",
"pulses",
"brainstem",
"neurophysiological",
"analysis",
"gamma-aminobutyric",
"acid",
"physics",
"biochemistry",
"cellular",
"neuroscience",
"cell",
"biology",
"anatomy",
"organic",
"chemistry",
"equipment",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences"
]
| 2018 | Organization of the core respiratory network: Insights from optogenetic and modeling studies |
Invariant Natural Killer T cells ( iNKT ) are a versatile lymphocyte subset with important roles in both host defense and immunological tolerance . They express a highly conserved TCR which mediates recognition of the non-polymorphic , lipid-binding molecule CD1d . The structure of human iNKT TCRs is unique in that only one of the six complementarity determining region ( CDR ) loops , CDR3β , is hypervariable . The role of this loop for iNKT biology has been controversial , and it is unresolved whether it contributes to iNKT TCR:CD1d binding or antigen selectivity . On the one hand , the CDR3β loop is dispensable for iNKT TCR binding to CD1d molecules presenting the xenobiotic alpha-galactosylceramide ligand KRN7000 , which elicits a strong functional response from mouse and human iNKT cells . However , a role for CDR3β in the recognition of CD1d molecules presenting less potent ligands , such as self-lipids , is suggested by the clonal distribution of iNKT autoreactivity . We demonstrate that the human iNKT repertoire comprises subsets of greatly differing TCR affinity to CD1d , and that these differences relate to their autoreactive functions . These functionally different iNKT subsets segregate in their ability to bind CD1d-tetramers loaded with the partial agonist α-linked glycolipid antigen OCH and structurally different endogenous β-glycosylceramides . Using surface plasmon resonance with recombinant iNKT TCRs and different ligand-CD1d complexes , we demonstrate that the CDR3β sequence strongly impacts on the iNKT TCR affinity to CD1d , independent of the loaded CD1d ligand . Collectively our data reveal a crucial role for CDR3β for the function of human iNKT cells by tuning the overall affinity of the iNKT TCR to CD1d . This mechanism is relatively independent of the bound CD1d ligand and thus forms the basis of an inherent , CDR3β dependent functional hierarchy of human iNKT cells .
Invariant Natural Killer T ( iNKT ) cells are a conserved subset of highly potent and versatile T-cells which specifically recognize the non-polymorphic lipid-presenting molecule CD1d [UniprotKB P15813] [1] . iNKT cells co-express a unique T-Cell Receptor ( iNKT TCR ) , which mediates recognition of CD1d , and the pan-NK receptor NKR-P1A ( CD161 ) . Human and mouse iNKT TCRs feature a homologous invariant TCRα chain , i . e . Vα24-Jα18 in humans and Vα14-Jα18 in mice . In addition , all human iNKT TCRs make use of a single TCR Vβ family , Vβ11 , whereas mouse iNKT TCRs utilize several different TCR Vβ families . The current paradox of iNKT biology lies in the fact that , despite their apparent innate-like simplicity , they can exert directly conflicting functions . On the one hand , several in vivo studies have demonstrated an essential role for iNKT cells in the induction and maintenance of immunological tolerance [2] , [3] . Consistent with this , iNKT cells exert a protective role in animal models of spontaneous autoimmunity [4] , [5] , and numerical and functional defects of iNKT cells are observed in different human autoimmune diseases [6] . In contrast to these tolerogenic functions , iNKT cells can exert potent cytotoxic functions and contribute to host defense against tumors and various infectious pathogens [7] , [8] , [9] . Whether different subsets of iNKTs are involved in these opposed roles or whether individual iNKT clones fulfill both of these functions under different conditions is unknown . Several mechanisms underpin iNKT activation during host defense , such as TLR [10] , [11] , [12] and PPAR-γ activation [13] , co-stimulatory molecule signaling [14] , and inflammatory cytokines [15] , [16] . However , it is unknown how iNKT cells are induced to mediate their tolerogenic functions under non-inflammatory conditions . Some iNKT clones exhibit substantial activation in response to CD1d-expressing antigen-presenting cells in the absence of exogenous antigens . This autoreactive function is essential for both iNKT selection [17] and tolerogenic activity [18] . While iNKT TCR binding to CD1d is absolutely required [19] , the mechanistic basis of iNKT cell autoreactivity is largely unresolved . In particular , the importance of specific CD1d-presented endogenous lipid antigens for the autoreactive interaction of the iNKT TCR with CD1d is contentious . Studies in mice have suggested that the iNKT repertoire displays clonal heterogeneity with regard to recognition of weaker stimulatory lipid antigens , such as the α-galactosylceramide analogue OCH . These differences can be explained by the differential Vβ family usage in mouse iNKT TCRs [20] , [21] , [22] . However , human iNKT TCRs use a single Vβ family and so the short hypervariable complementarity determining region ( CDR3β ) loop in human iNKT TCRs is their only truly adaptive element . It is not known whether this is sufficient to endow the human iNKT TCR with meaningful ability to discriminate a diverse range of human CD1d-presented antigens . Here we examined a large panel of human iNKT cell lines and clones for their binding to different CD1d-ligand tetramers and related this both to the affinity of their TCRs to different CD1d-ligand complexes and to their functional recognition of diverse antigens . The results presented here demonstrate that variations in the CDR3β loop have a profound , antigen-independent , impact on the iNKT TCR's affinity to CD1d and on iNKT cell autoreactive function .
Previous studies have shown that the CDR3β loop is dispensable for the ability of human iNKT cells to strongly react to the α-galactosylceramide antigen KRN7000 ( K7 ) , a xenobiotic glycolipid which can be presented to iNKT cells by CD1d . In fact , K7-CD1d tetramer staining does not allow discrimination of different human iNKT cell subsets by flow cytometry . We hypothesized that CD1d-tetramers loaded with weaker antigens might be better able to reveal the existence of CDR3β-dependent variation among human iNKT cells . Therefore , we first examined whether different human iNKT subsets could be segregated by their binding to CD1d tetramers that were loaded with the synthetic iNKT partial agonist antigen OCH . For this purpose , polyclonal iNKT lines , generated from healthy donors by in vitro stimulation with K7 , were tested for their binding to both K7- and OCH-CD1d tetramers . In all of these lines , K7-CD1d tetramers stained a single , clearly distinct , homogeneous , and strongly fluorescent population of iNKT lymphocytes ( Figure 1A ) . In contrast , staining of the same lines with OCH-CD1d tetramers revealed a considerable degree of variation in fluorescence , suggesting the presence of distinct iNKT subpopulations ( Figure 1A ) . Importantly , similar qualitative differences between K7- and OCH-CD1d tetramer staining of iNKT cells could also be observed ex vivo ( Figure 1B ) , indicating that these differences were not due to an artifact of previous in vitro stimulation with K7 . In order to examine whether the broadly heterogeneous OCH-CD1d tetramer staining of human iNKT cells resulted from stable clonal variation or from transient changes in TCR expression levels , we generated a large panel of “K7/OCH-naïve” human iNKT cell clones and lines . For this purpose , Vα24+/Vβ11+ T cells were directly sorted ex vivo from healthy human donors and expanded using the non-specific T cell mitogen phytohaemagglutinin . Ninety-seven different human Vα24+/Vβ11+ T cell lines and 256 Vα24+/Vβ11+ T cell clones from 13 different healthy donors were established and analyzed by flow cytometry with K7- and OCH-CD1d tetramers . All Vα24+/Vβ11+ T-cell clones and lines showed bright , homogeneous staining with K7-tetramers ( Figure 2 ) , thereby confirming them as iNKT cells . Individual iNKT clones showed modest variation , up to 6-fold , in K7-CD1d tetramer mean fluorescence intensity ( MFI ) . In contrast , multiple iNKT cell subpopulations with differing fluorescence intensities were revealed by OCH-CD1d tetramer staining in 31 of the 97 iNKT lines ( Figure 2A ) , thereby mirroring the above described findings in K7 stimulated iNKT lines . As expected , all 256 iNKT clones stained homogeneously with OCH-CD1d tetramers . However , substantial differences , up to 200-fold , in OCH-CD1d tetramer MFI were observed between individual clones ( Figure 2B ) . Based on the observed large differences in OCH-CD1d tetramer MFI , the 256 human iNKT clones were categorized as OCHHIGH ( MFI>300; n = 41 ) , OCHINT ( MFI>50 and <300; n = 164 ) , or OCHLOW ( MFI<50; n = 51 ) . Importantly , the differences in OCH-CD1d tetramer staining could not be explained by differences in either TCR or CD4 co-receptor expression . Whereas K7-CD1d tetramer binding significantly correlated with surface expression levels of the Vα24 and Vβ11 TCR chains , no such association was observed for OCH-CD1d tetramer staining ( Figure 2C ) . Furthermore , CD4 co-receptor usage was not related to the intensity of the iNKT clones' OCH or K7-CD1d tetramer staining ( unpublished results ) . The results of these experiments revealed that the human iNKT repertoire is broadly heterogeneous with regard to the ability of individual clones to bind OCH-CD1d tetramers , independent of either CD4 co-receptor or TCR expression levels . The above results indicated that clonally distributed qualitative differences in iNKT TCRs were responsible for the considerable variation in OCH-CD1d tetramer binding . However , differences in iNKT TCR mediated recognition of an unnatural compound like OCH would be physiologically irrelevant if they simply reflected random differences in OCH-specific antigen selectivity . To explore this possibility , 18 iNKT clones of broadly varying OCH-CD1d MFI were tested for their ability to bind CD1d tetramers loaded with the common mammalian glycolipid β-glycosylceramide ( βGC ) . These 18 iNKT clones displayed significant variation , up to 50-fold , in βGC-CD1d tetramer staining ( Figure 3A ) . Importantly , a strong association was evident between OCH-CD1d tetramer staining and βGC-CD1d tetramer staining , while no correlation was seen between βGC-CD1d tetramer staining and Vα24 TCR chain surface expression ( Figure 3B ) . These results demonstrated that the observed broad variation in OCH-CD1d tetramer binding between individual human iNKT clones was not simply due to their antigen selectivity but was a reflection of a general variability in human iNKT TCR binding to CD1d loaded with weak antigenic lipids . Furthermore , they indicated that OCH-CD1d tetramer binding can act as a surrogate marker for human iNKT cell recognition of endogenous CD1d antigens . Based on the above results we hypothesized that the observed substantial differences in tetramer staining between OCHHIGH and OCHLOW iNKT clones resulted from significant variations in TCR:CD1d binding affinity . As expected , sequencing of the TCR Vα24 and Vβ11 chains demonstrated the usage of the known invariant Vα24-Jα18 rearrangement in all clones , while Vβ11 in these clones was rearranged with several different Jβ families , resulting in highly variable CDR3β sequences . This indicated that , in human iNKT TCRs , structural differences of the CDR3β loop have a substantial impact on iNKT TCR binding to CD1d . To test this in a cell-free system we cloned the extracellular domains of the TCR Vβ11 chains from a panel of seven OCHHIGH and OCHLOW iNKT cell clones ( Table 1 ) , as well as the invariant TCR Vα24 chain from one iNKT clone , and used them to generate soluble Vα24/Vβ11 iNKT TCRs . Binding of these recombinant iNKT TCRs to K7- , OCH- , as well as βGC- and lactosylceramide ( LacCer- ) loaded recombinant human CD1d complexes was measured using surface plasmon resonance ( Figure 4A; Table 2 ) . The results of these experiments showed a striking variation , up to 40-fold , between the different iNKT TCRs in their binding affinity ( KD ) to a given ligand-CD1d complex ( for K7-CD1d , KD: 0 . 24–3 . 67 µM; for OCH-CD1d , KD: 2 . 17–38 . 3 µM; for βGC-CD1d , KD: 2 . 17–85 µM; for LacCer-CD1d , KD: 2 . 1–54 µM; see Table 2 ) . These findings clearly showed that the CDR3β loop of human iNKT TCRs can strongly impact on their binding to ligand-CD1d complexes . Importantly , the binding affinities of all seven recombinant iNKT TCRs to OCH-CD1d strongly correlated with the OCH-CD1d tetramer staining ( MFI ) of their corresponding original iNKT clones ( Figure 4B ) . Moreover , the binding affinity of a given iNKT TCR to OCH-CD1d also correlated closely with its affinity to either βGC- or K7-CD1d ( Figure 4C ) . Therefore , the wide variation in affinity between our seven human iNKT TCRs contrasted to the lack of variation in antigen selectivity . In other words , the CDR3β loop of human iNKT TCRs modulated the overall binding affinity to different human ligand-CD1d complexes irrespective of the bound ligand . Based on these findings we hypothesized that the TCRs of OCHHIGH iNKT clones could also mediate enhanced functional recognition of endogenous ligand-CD1d complexes . We tested this hypothesis by comparing autoreactive responses of OCHHIGH and OCHLOW iNKT clones to CD1d-expressing antigen-presenting cells . We directly compared the extent of proliferation , cytokine secretion , and cytotoxicity of human OCHHIGH and OCHLOW iNKT cells in response to CD1d expressing human cell lines presenting either endogenous or specific exogenous ( “pulsed” ) glycolipids . Because functional responses of iNKT cells might change during long term in vitro culture , we compared different donor-matched pairs of OCHHIGH and OCHLOW iNKT cell clones with identical in vitro history , i . e . each pair was sorted from a given donor 3 wk prior to the experiment and kept under identical cell culture conditions until the day of the experiment . The selected clones were all CD4+ and were additionally matched for TCR expression levels . For all pairs , OCHHIGH iNKT clones exhibited significantly greater proliferation than OCHLOW iNKT clones in response to either unpulsed or OCH-pulsed T2-CD1d lymphoblasts . In contrast , when T2-CD1d were pulsed with the strong agonist ligand K7 , both OCHHIGH and OCHLOW iNKT clones proliferated vigorously , and to similar extent ( Figure 5A ) . Next , we measured CD1d-dependent secretion of a panel of cytokines by OCHHIGH and OCHLOW iNKT clones . The OCHHIGH iNKT clones secreted considerably greater quantities of cytokines than their OCHLOW counterparts in response to either unpulsed or OCH-pulsed T2-CD1d cells ( Figure 5B , C ) , while no significant differences in cytokine secretion were observed between OCHHIGH and OCHLOW iNKT clones upon stimulation with K7-pulsed T2-CD1d cells . A general Th0-type cytokine secretion pattern was observed in response to stimulation with either K7 or OCH , while a Th1 pattern was often produced by autoreactive stimulation of OCHHIGH iNKT ( Figure 5C ) . Although most OCHLOW iNKT clones did not exhibit autoreactive cytokine release , two OCHLOW iNKT clones reproducibly secreted significant amounts of IL-13 and either IL-4 or IL-5 , but no IFNγ or TNF-α , while one OCHLOW iNKT clone secreted measurable amounts of IFNγ and TNF-α , but no Th2 cytokines . None of the tested iNKT clones secreted detectable amounts of cytokines in response to CD1d-deficient T2-lymphoblasts , and blocking of surface CD1d molecules on T2-CD1d by the monoclonal antibody CD1d42 effectively prevented autoreactive secretion of cytokines by OCHHIGH or OCHLOW iNKT cells ( unpublished data ) . Therefore , autoreactive cytokine secretion by these iNKT clones was wholly dependent on their recognition of surface CD1d . Finally , in Cr51 release assays , OCH-pulsed T2-CD1d were much more efficiently killed by OCHHIGH iNKT clones than their corresponding OCHLOW iNKT clones ( Figure 6D ) . In contrast , K7-pulsed T2-CD1d were efficiently lysed by both OCHHIGH and OCHLOW iNKT clones , whereas neither OCHHIGH nor OCHLOW iNKT clones showed relevant cytotoxicity towards unpulsed T2-CD1d lymphoblasts . Together , these results demonstrated that OCH-CD1d tetramer staining allows for identification of distinct human OCHHIGH and OCHLOW iNKT clones , which exhibit differential functional ability to respond to endogenous ligand-CD1d complexes . The above results indicated that the autoreactive potential of human iNKT clones is governed by the affinity of their iNKT TCR to CD1d , and therefore the structure of their CDR3β loop . In order to test our hypothesis that OCHHIGH and OCHLOW iNKT TCRs differed in their binding to endogenous ligand-CD1d complexes , we generated soluble fluorescent iNKT TCR-tetramers derived from an autoreactive OCHHIGH iNKT clone and a non-autoreactive OCHLOW iNKT clone . As shown in Figure 6 , both iNKT TCR tetramers bound well to K7-pulsed T2-CD1d . In contrast , only the OCHHIGH-derived iNKT TCR tetramer was able to effectively stain unpulsed T2-CD1d . These results further substantiated our hypothesis that autoreactive recognition of CD1d by human iNKT cells is primarily determined by the structure of their iNKT TCRs' CDR3β loop . All together , these studies demonstrated that the human iNKT cell repertoire exhibits considerable clonally distributed CDR3β-dependent differences in overall TCR affinity to CD1d , irrespective of the bound ligand , and that these inherent structural differences control iNKT autoreactive activation .
iNKT cells are a conserved subset of highly potent regulatory T cells at the innate-adaptive interface . The hallmark of human iNKT cells is their unique TCR , which is composed of an invariant TCR Vα24-Jα18 alpha chain and a semi-invariant TCR Vβ11 chain . The only variable , and therefore potentially adaptive , element in human iNKT TCRs is their hypervariable CDR3β loop . The results of the present study demonstrate for the first time , to our knowledge , that the structure of the hypervariable CDR3β loop in human iNKT TCRs exerts a strong impact on CD1d binding and is a key determinant of iNKT cell autoreactivity . The magnitude of the effect of CDR3β variations on human iNKT TCR:CD1d binding observed here was unexpected as previous studies with mouse iNKT TCRs have reported only minor effects of CDR3β mutations on CD1d binding . Furthermore , they strongly suggest that CDR3β loops in autoreactive iNKT TCRs make functionally important direct protein-protein contacts with human CD1d , rather than contacts with CD1d-bound ligands , thereby affecting overall affinity rather than antigen specificity . The role of the hypervariable CDR3β loop in human iNKT TCRs is currently unresolved . It is dispensable for binding to CD1d molecules that are loaded with the strong agonist ligand K7 , and hence K7-CD1d tetramers do not support subset differentiation of human iNKT cells . Consistent with this , the recently solved structures of one human and two mouse iNKT TCR:K7-CD1d co-crystals have found no relevant contacts between CDR3β and the K7-CD1d complex [20] , [23] . In contrast , recent mutagenesis studies have indicated that the CDR3β loop of mouse iNKT TCRs may exert some impact on the affinity to CD1d , particularly when CD1d was loaded with weaker antigens [24] , [25] , [26] . We found that human iNKT cells were surprisingly heterogeneous in their binding to CD1d tetramers loaded with the partial agonist ligand OCH , which is a synthetic analogue of K7 . Up to 200-fold differences in OCH-CD1d tetramer staining were observed between individual iNKT clones , independent of variations in TCR expression . The same clones exhibited only modest differences in K7-CD1d tetramer staining , which could largely be explained simply by variations in TCR expression . Importantly , we found that the clonal variation in OCH-CD1d tetramer binding was directly related to OCH-CD1d dependent functional responses , while no such linkage was observed between K7-CD1d tetramer staining and K7-dependent functional iNKT activation . These data underpinned the notion that the five germline encoded CDR loops in human iNKT TCRs , i . e . CDR1α-3α and CDR1β-2β , are sufficient for effective iNKT cell interaction with K7-CD1d [26] . Importantly , they strongly indicated that productive iNKT TCR interactions with OCH-CD1d require additional binding energy provided by certain CDR3β loop structures . We tested this hypothesis by directly measuring the binding of K7- and OCH-CD1d complexes to a panel of seven recombinant human iNKT TCRs , which were derived from selected OCHHIGH and OCHLOW iNKT clones . These recombinant iNKT TCRs differed only in their CDR3β structure . The results of these experiments demonstrated that the broad clonal heterogeneity in OCH-CD1d tetramer staining is indeed directly determined by the iNKT clones' TCRs binding affinities to OCH-CD1d , and hence the structure of the CDR3β loop . Conversely , while all tested recombinant iNKT TCRs bound approximately 10-fold better to K7-CD1d than to OCH-CD1d , the fold-differences in affinity between the strongest and the weakest binding iNKT TCRs were similar for binding to either OCH- or K7-CD1d . Together with the above discussed tetramer-based and functional studies , this indicates that the synthetic CD1d ligand K7 pushes the interaction between human CD1d and iNKT TCRs beyond a physiological range . This is consistent with numerous in vivo and in vitro studies which showed that K7 induces concurrent massive iNKT cell secretion of TH1- , TH2- , and TH17-type cytokines , whereas OCH causes a clearly TH2-biased cytokine secretion pattern [27] . Also , addition of K7 to mouse fetal thymic organ cultures leads to effective deletion of iNKT cells [28] , and K7 stimulation induces a prolonged anergy in iNKT cells [29] , which supports the view that K7 is not a physiological ligand for iNKT cells . Hence , a full understanding of the biological role of CDR3β loop polymorphism will require more studies with weaker agonistic antigens , and the results of this study suggest that OCH is a good surrogate for endogenous weak agonist antigens . There are two competing models to explain how differences in CDR3β loop structure could translate into variations of weak antigen recognition . In an “antigen-dependent” or “adaptive” model , the CDR3β loop bestows upon iNKT cells a degree of lipid selectivity by controlling iNKT TCR affinity to CD1d in a lipid antigen-specific manner . Alternatively , in an “antigen-independent” or “innate-like” model , the CDR3β loop structure modulates iNKT TCR binding affinity to CD1d via protein-protein interactions . This model would allow higher , but not lower , affinity TCR structures to recognize CD1d molecules presenting weaker lipid antigens but , crucially , without differential patterns of lipid antigen selectivity between iNKT TCRs of similar CD1d affinity . In other words , this model predicts that the inherent CDR3β sequence in a given human iNKT clone would determine its iNKT TCR's general ability to bind to diverse ligand-CD1d complexes . An important corollary of this would be a fixed hierarchy of high and low affinity iNKT clones . A prediction arising from this model would be that iNKT cells lack the ability to develop immunological memory to specific pathogens , which is a hallmark of adaptive immunity . Although iNKT TCRs clearly belong to the broader family of rearranged , and therefore “adaptive , ” TCRs and BCRs , their limited structural diversity and lack of antigen-selectivity , as proposed by this model , are strongly reminiscent of innate immune receptors . In order to test which of the two above models best explains the observed CDR3β-dependent variation in iNKT TCR binding to OCH-CD1d , we examined recognition of two β-linked glucosylceramides , βGC and LacCer , by a panel of iNKT TCRs . K7 and OCH are α-linked monosaccharide glycosylceramides and are not expressed in mammals , whereas βGC and LacCer are natural β-linked glycosylceramides of mammalian cell membranes . The different configurations of α- and β-anomeric glycolipids enforce substantial differences in the orientation of their glycosyl headgroups when presented by CD1d [30] , [31] . Therefore , if the substantial variation in iNKT TCR affinity to OCH-CD1d observed in our study was mainly a function of clonal variation in lipid antigen specificity , as predicted by the “adaptive” model , there should be no association between an individual iNKT TCR's affinity to OCH-CD1d and its affinity to either βGC-CD1d or LacCer-CD1d . However , the results of the present study strongly support the “innate” model: βGC-CD1d tetramer binding to human iNKT clones correlated in a linear fashion with OCH-CD1d tetramer binding , and our binding studies with several different soluble iNKT TCRs demonstrated that the CDR3β loop of human iNKT TCRs strongly modulated the overall binding affinity to different human ligand-CD1d complexes , independent of the bound ligand . CDR3β loop hypervariability of human iNKT TCRs therefore strongly impacts on overall affinity to CD1d but does not exert a relevant effect on antigen selectivity . The powerful effect of natural CDR3β variations on human iNKT TCR:CD1d affinity observed in our study was unexpected as previous iNKT TCR mutagenesis studies in mice have suggested only a weak impact of CDR3β structure on iNKT TCR binding affinity [24] , [25] , [26] . Indeed , hybridomata expressing mouse iNKT TCRs with randomized CDR3β regions only displayed moderate variability in binding to K7-CD1d tetramers , and only very few TCRs were capable of interacting with CD1d presenting endogenous lipids [25] . Furthermore , previously published iNKT TCR:CD1d co-crystal structures showed a negligible contribution of the CDR3β to the interaction [20] , [23] . The apparent discrepancies between these studies and the current findings could indicate relevant species differences , as the mutagenesis studies have concentrated on mouse iNKT binding or else might reflect differences in study design: the only crystal structure study of human iNKT TCR:CD1d binding was limited to a single iNKT TCR of unknown weak antigen-CD1d affinity while the current study systematically screened a large panel of naturally occurring human iNKT clones . Interestingly , while the iNKT TCR used for the human co-crystal structure study displayed very limited contacts between its CDR3β loop and CD1d , a modeling exercise of TCR Vβ11 docking onto CD1d in the same study [23] pointed to a significant degree of plasticity of the CDR3β conformation . In particular , the CDR3β loop of one of our previously published CD1d-restricted Vα24− Vβ11+ TCRs , TCR 5E [32] , could make significant contacts with the alpha-2 helix of human CD1d [23] . Consistent with this , a refolded hybrid TCR of the 5E Vβ11 chain and the invariant Vα24-Jα18 chain binds with high affinity to both CD1d/OCH and CD1d/βGC ( unpublished data ) . Therefore , certain CDR3β loop structures can potentially facilitate the recognition of human CD1d loaded with weak ligands by providing additional binding energy to the TCR-CD1d interaction . Sequence analysis of the CDR3β loops studied did not reveal any obvious correlations between CD1d binding affinity and either physicochemical properties of the loop as a whole or the position of specific residues within the sequence . This is not surprising , given the high degree of conformational flexibility of CDR loops . The above described considerable binding affinities of some human iNKT TCRs to naturally occurring beta-anomeric glycolipids , i . e . βGC and LacCer , have important implications for the clonal distribution of iNKT autoreactivity . CD1d-dependent autoreactivity of iNKT cells , i . e . their CD1d-mediated activation in the absence of exogenous antigens , is likely to play important biological roles , but the molecular mechanisms determining iNKT autoreactivity have been unresolved . CD1d-dependent autoreactivity is observed in approximately 30% of mouse iNKT hybridomas[19] , and studies in iNKT deficient and autoimmune prone mice have shown that autoreactive CD1d-recognition is required for iNKT selection and also iNKT-mediated immunological tolerance [15] , [18] , [33] , [34] . However , much less is known about the role of CD1d-dependent iNKT autoreactivity in humans . Neonatal human iNKT cells exhibit an activated memory phenotype , indicating their in vivo recognition of CD1d molecules in the absence of exogenous ligands [35] . An “adaptive” model has been proposed to explain autoreactive activation of iNKT cells in mouse models of bacterial infection , and it was postulated that autoreactive murine iNKT cells specifically recognize de novo synthesized antigens , such as isogloboside 3 [36] . Consistent with this model , mouse CD1d requires endosomal trafficking to elicit autoreactive activation of murine iNKT cells , which suggests that processing of the ligand-CD1d complex is essential [37] . However , in contrast to mouse iNKT cells , human iNKT cell autoreactivity is not dependent on CD1d trafficking or endosomal acidification [38] , again suggesting important species differences between mouse and human iNKT cell activation . The antigen-independent “innate-like” model discussed above offers a simpler explanation for the clonally distributed iNKT autoreactivity . iNKT clones with higher overall iNKT TCR:CD1d affinity would have an intrinsically greater autoreactive potential than low affinity clones , and these differences in autoreactive potential would be independent of de novo synthesized CD1d-bound ligands . Autoreactive activation of iNKT clones in this model would still be controlled by local conditions , such as TLR signaling [12] , CD1d expression [16] , or cytokine expression [39] . High affinity iNKT clones would be capable of exerting autoreactive functions under physiological conditions , while low affinity iNKT clones would only be recruited under more pro-inflammatory conditions , e . g . during bacterial infections . Our functional analyses of autoreactive activation of OCHHIGH and OCHLOW iNKT clones support the “innate-like” model . Firstly , autoreactive activation of several matched pairs of human iNKT clones was closely associated with their OCH-CD1d tetramer binding characteristics . Secondly , only iNKT TCR-tetramers generated from OCHHIGH iNKT clones were able to bind to CD1d-expressing antigen-presenting cells in the absence of exogenous lipid . The above data therefore underpin the “innate-like” model , whereby the hypervariable CDR3β loop balances TCR binding affinity to CD1d protein , and hence the autoreactive potential of an iNKT clone , independent of the bound ligand . The different activation thresholds of ex vivo sorted human OCHHIGH and OCHLOW iNKT clones shown herein suggest different in vivo functions of these subsets . For example , OCHHIGH and OCHLOW iNKT cells might differ in their ability to drive the formation of immature DCs and consequently in their capability to constitutively promote peripheral tolerance . Finally , it is intriguing to speculate that CDR3β-dependent asymmetrical activation of the human iNKT repertoire could , over time , skew the balance between OCHHIGH and OCHLOW iNKT clones , with ensuing consequences for iNKT-dependent functions in both host defense and immunological tolerance .
Peripheral blood mononuclear cells ( PBMC ) were isolated from human peripheral venous blood by density gradient centrifugation ( Ficoll-Hypaque; Amersham Pharmacia and Upjohn ) . The study was approved by the local ethics committee ( KEK , Bern , Switzerland ) . All donors gave informed consent . Human iNKT clones and lines were generated by FACSVantage sorting of Vα24+/Vβ11+ T cells into round-bottomed 96-well plates . Sorted cells were stimulated with 1 µg/ml phytohaemagglutinin ( Remel , USA ) in the presence of autologous γ-irradiated ( 35Gy ) PBMCs . Cells were grown in T cell growth medium ( RPMI 1640 , 2% human AB serum ( SRK , CH ) , 10% fetal bovine serum ( FBS ) , 0 . 1 mg/ml kanamycin , 1 mM sodium pyruvate , 1% non-essential amino acids , 1% L-glutamax , and 50 µM 2-mercaptoethanol ( all from Gibco Invitrogen ) and IL-2 ( Proleukin , Chiron ) 200 IU/ml ) . IL-2 concentration in the medium was gradually reduced to 20 IU/ml 3 wk after sorting . The following fluorescent reagents were used to analyze human iNKT cells: PE-conjugated human CD1d tetramers loaded with either K7 , OCH , βGC [40]; FITC-conjugated anti-human TCR Vβ11 , PE-anti-human TCR Vα24 , ( Serotec , UK ) ; PerCP-anti-CD3 , FITC-anti-CD3 , APC-anti-CD4 , APC-anti-CD8 , ( BD Pharmingen ) . After addition of staining reagents , cells were incubated at 4°C for 45 min , washed twice in ice-cold PBS/1% FBS , and acquired on a four-color FACSCalibur flow cytometer ( Becton Dickinson ) . Propidium iodide was used to exclude dead cells . Data were processed using CellQuest Pro software ( BD Biosciences , USA ) . Staining with PE-streptavidin conjugated iNKT-TCR tetramers ( 4C12 and 4C1369 ) were carried out in the same way as CD1d-tetramer stainings . Soluble TCR heterodimers were generated as previously described [41] . Briefly , the extracellular region of each TCR chain was individually cloned in the bacterial expression vector pGMT7 and expressed in Escherichia coli BL21-DE3 ( pLysS ) . Residues Thr48 and Ser57 , respectively , of the α- and β-chain TCR constant region domains were both mutated to cysteine . Expression , refolding , and purification of the resultant disulfide-linked iNKT TCR αβ heterodimers was carried out as previously described [32] . Streptavidin ( ∼5 , 000 RU ) was linked to a Biacore CM-5 chip ( BIAcore AB , UK ) using the amino-coupling kit according to manufacturer's instructions , and lipid-CD1d complexes or control proteins ( βGC-CD1b and HLA-A2*01-NY-Eso-1 ( 157-165 ) complex ) were flowed over individual flow cells at ∼50 µg/ml until the response measured ∼1 , 000 RU . Serial dilutions of recombinant iNKT TCRs were then flowed over the relevant flow cells at a rate of 5 µl/min ( for equilibrium binding measurements ) or 50 µl/min ( for kinetic measurements ) . Responses were recorded in real time on a Biacore 3000 machine at 25°C , and data were analyzed using BIAevaluation software ( Biacore , Sweden ) . Equilibrium dissociation constants ( KD values ) were determined assuming a 1∶1 interaction ( A+B ↔ AB ) by plotting specific equilibrium binding responses against protein concentrations followed by non-linear least squares fitting of the Langmuir binding equation , AB = B×ABmax/ ( KD+B ) , and were confirmed by linear Scatchard plot analysis using Origin 6 . 0 software ( Microcal , USA ) . Kinetic binding parameters ( kon and koff ) were determined using BIAevaluation software . Stable human CD1d-expressing T2-lymphoblast lines and clones ( T2-CD1d ) were generated by spin infection of T2 lymphoblasts with lentiviral particles encoding the human CD1d gene . VSV–G pseudotyped lentiviral particles were generated as previously described [42] . The following primers were used to clone full-length human CD1d into the lentiviral vector pHR'SIN18: 5′-AGCGGGATCCGCCGCCACCATGGGGTGCCTGCTGTTTCTGCTG-3′ ( forward ) , and 5′-GCGTCTCGAGTCACAGGACGCCCTGATAGGAAGTTTG-3′ ( reverse ) . In brief , HEK293T cells were co-transfected with 5 µg of pVSV-G [43] , 10 µg of the packaging plasmid pCMV δ8 . 91 [44] , and 15 µg of the human CD1d-encoding transfer vector pHR'SIN18-hCD1d by calcium phosphate method . Viral supernatants were harvested 48–60 h post-transfection , filtered , and concentrated by centrifugation at 25 , 000 rpm , 4°C for 90 min . Viral pellets were resuspended in 1 ml fresh RPMI 1640 for transduction . Transduced cells were maintained in growth medium for 10 d before sorting of human CD1d-expressing T2 single cells and lines on a FACSVantage SE apparatus ( Becton Dickinson , USA ) , using PE-conjugated anti-human-CD1d antibody CD1d42 ( Pharmingen , Switzerland ) . T2 lymphoblast cells ( T2- ) and CD1d-expressing T2 lymphoblast cells ( T2-CD1d ) were used as antigen presenting cells ( APC ) . 5×104 iNKT cells were plated in a 96-well round-bottom plate in triplicates with either medium alone , with 2 . 5×104 T2-CD1d , or with T2 lymphoblasts . Before use , T2-CD1d and T2 lymphoblasts were treated with 0 . 1 mg/ml mitomycin C for 1 h at 37°C and extensively washed with PBS . Lipid antigens ( K7 , OCH , and βGC ) were added at a final concentration of 100 ng/ml . Lipids were solubilized at 200 µg/ml by sonication in vehicle ( 0 . 5% Tween-20 ) , which was also used as a negative control . IL-2 was added to the culture medium at a final concentration of 10 IU/ml . Proliferation was measured during the last 18 h of a 96 h incubation by addition of 1 µCi [3H]-methyl-thymidine ( 1 Ci = 37 GBq , Amersham Pharmacia ) , followed by harvesting and scintillation counting ( Perkin Elmer beta counter ) . Levels of IL-4 , IL-5 , IL-10 , IL-13 , GM-CSF , IFN-γ , and TNF-α were measured in the cell supernatants , collected after 48 h of incubation , by Bio-Plex suspension array system ( Bio-Rad , USA ) , according to manufacturer's recommendations . T2 lymphoblasts and T2-CD1d were cultured for 16 h either in the presence of lipid antigens at 100 ng/ml concentration or an equivalent quantity of vehicle . They were then labeled with 100 µCi of 51Cr ( GE Healthcare , UK ) for 1 h at 37°C and washed 3 times with warm RPMI 1640 supplemented with 1% FBS . iNKT cells were added in duplicates at different effector-to-target cell ratios and cultured for 4 h . Maximal 51Cr release was determined from target cells lysed by hydrochloric acid . The percentage of specific lysis was calculated by the following formula: [ ( experimental cpm − spontaneous release cpm ) / ( maximum release cpm − spontaneous release cpm ) ] ×100% . Percentage of unspecific lysis was always <20% . Soluble iNKT-TCR heterodimers were biotinylated via an engineered BirA motif on the C-terminus of their TCR β-chain and then conjugated to PE-streptavidin ( Molecular Probes , USA ) . Multimeric complexes were purified by FPLC ( Pharmacia , Sweden ) on an SD200 column ( Pharmacia , Sweden ) and concentrated to 1 mg/ml using Vivaspin20 concentrators ( Vivascience , UK ) . | Our immune system uses randomly modified T-cell receptors ( TCRs ) to adapt its discriminative capacity to rapidly changing pathogens . The T-cell receptor ( TCR ) has six flexible , variable peptide loops that make contact with antigens presented to them on the surface of other cells . Invariant Natural Killer T-cells ( iNKT ) are regulatory T-cells with a unique type of TCR ( iNKT-TCR ) that recognizes lipid antigens presented by specific MHC-like molecules known as CD1d . In human iNKT-TCRs , only one of the six loops , CDR3beta , is variable . By comparing how different human iNKT clones bind and react to different CD1d-lipid complexes we uncover the existence of a hierarchical order of the human iNKT cell repertoire in which strongly CD1d-binding clones are autoreactive while weak CD1d-binding clones are non-autoreactive . Direct measurements of iNKT-TCR binding to CD1d using surface plasmon resonance recapitulated this hierarchy at the protein level . The data show that variation in the CDR3beta loop conveys dramatic differences in human iNKT TCR affinity that are independent of the CD1d bound ligand . Thus the CDR3beta loop provides the structural basis for the functional hierarchy of the human iNKT repertoire . We postulate that during the life-course , CDR3beta-dependent asymmetrical activation of different human iNKT clones leads to a bias in the iNKT repertoire , and this could result in age-dependent defects of iNKT-mediated immune regulation in later life . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods",
"and",
"Materials"
]
| [
"immunology/antigen",
"processing",
"and",
"recognition",
"immunology/autoimmunity",
"immunology/innate",
"immunity"
]
| 2010 | Innate-Like Control of Human iNKT Cell Autoreactivity via the Hypervariable CDR3β Loop |
European population genetic substructure was examined in a diverse set of >1 , 000 individuals of European descent , each genotyped with >300 K SNPs . Both STRUCTURE and principal component analyses ( PCA ) showed the largest division/principal component ( PC ) differentiated northern from southern European ancestry . A second PC further separated Italian , Spanish , and Greek individuals from those of Ashkenazi Jewish ancestry as well as distinguishing among northern European populations . In separate analyses of northern European participants other substructure relationships were discerned showing a west to east gradient . Application of this substructure information was critical in examining a real dataset in whole genome association ( WGA ) analyses for rheumatoid arthritis in European Americans to reduce false positive signals . In addition , two sets of European substructure ancestry informative markers ( ESAIMs ) were identified that provide substantial substructure information . The results provide further insight into European population genetic substructure and show that this information can be used for improving error rates in association testing of candidate genes and in replication studies of WGA scans .
Differences in population genetic structure and substructure between cases and controls can lead to false positive association tests [1–5] . Interest in this issue has accelerated with the application of whole genome association ( WGA ) screens for deciphering the genetics of complex diseases . The importance of recognizing and controlling for population structure is magnified when population controls are not closely matched to cases , a process that requires multiple demographic considerations and similar sample acquisition methods . These conditions are difficult and often not practical to fulfill completely . Since many studies focus on participants of European descent , the potential impact of European substructure on association testing has specifically engendered interest [6 , 7] . In fact , the current study was undertaken as part of an effort to effectively ascertain and adjust for differences in population substructure among cases and controls in our studies of the genetics of rheumatoid arthritis in a participant set that predominantly includes participants of European descent . Recent studies have addressed differences in population substructure and methods to control for these differences in association testing [8–13] . Population substructure can be explored and ascertained using a variety of algorithms that apply principal component analysis ( PCA ) or non-hierarchical cluster analysis based on allele frequencies in individuals and groups . Unlike other multi-locus adjustments ( e . g . genomic control methods [14] ) these newer approaches adjust for the fact that some SNPs have large frequency variations across different populations compared to other SNPs [11] . The ability of these methods to control for large differences in population substructure has been at least partially demonstrated by both real data and simulations [6 , 11 , 12] . However , the practical application of these methods and limitations requires more extensive exploration in a variety of real datasets . Recent studies by our group and others have led to the identification of SNP subsets that can provide European substructure information [6 , 7]; this is consistent with previous work suggesting distinct clines of genetic variation within Europe [15–20] . These European substructure ancestry informative markers ( ESAIMs ) may be particularly important in large replication studies in which independent sets of case and control genotypes are necessary to confirm and further define associations without the benefit of genome-wide SNP typing . Previous studies have been limited to initial SNP genotyping sets of less than 10 , 000 SNPs [6 , 7] . The current study uses 300K to 500K genome-wide SNP data to enhance the ability to define elements of European substructure that were not evident or poorly defined using smaller sets of SNPs .
A set of 952 self-identified participants of diverse European descent genotyped with >300K SNPs was used for the first phase of European population substructure analysis . This participant group predominantly included European Americans as well as smaller numbers of individuals from Italy and Spain ( see Methods ) . In order to reduce potential noise created by continental admixture this study included only those individuals who did not have evidence of non-European continental ancestry ( see Methods ) . The genotypes were examined using the principle component analysis ( PCA ) algorithm implemented in the EIGENSTRAT program [11] , a computational method that enables rapid analyses of very large datasets . Using multiple criteria including ANOVA , a split half reliability test ( see Methods ) and a test for normality of distribution , substructure was present in multiple principle components ( Table 1 ) . However , most of the variance among the populations was observed in the first principal component ( PC ) . This PC accounted for >5 fold the variance of the second PC . The clustering of individuals for PC1 and PC2 corresponded to self-reported regional and ethnic origins ( Figure 1A and 1B ) . This is best illustrated when considering only those participants with the same grandparental country of origin and those individuals that indicated Ashkenazi Jewish ancestry ( Figure 1B ) . Similar to our previous studies using smaller sets of SNPs , the clustering of individuals of Ashkenazi Jewish ancestry does not correspond to grandparental European country of origin , which was diverse [6] . The first PC showed a gradient that distinguished “southern” or Mediterranean origin from “northern” European ancestry ( Figure 1B ) . The mean +/− SD of the first PC scores for those individuals with the same ( or adjoining for Scandinavian ) 4 grandparent ( GP ) country of origin or 4GP Ashkenazi ancestry information were: Irish ( 51 individuals ) , mean −0 . 022 +/− 0 . 002; Scandinavian ( 3 individuals ) , mean , −0 . 022 +/− 0 . 002; United Kingdom ( 5 individuals ) , −0 . 020 +/− 0 . 002; German ( 11 individuals ) , −0 . 016 +/− 0 . 004; Spanish ( 14 individuals ) , 0 . 004 +/− 0 . 003; Italian ( 28 individuals ) , 0 . 015 +/− 0 . 006; Greek ( 9 individuals ) , 0 . 022 +/- 0 . 011; and Ashkenazi ( 38 individuals ) , 0 . 045 +/− 0 . 003 . For participants self-identified as of Ashkenazi heritage , but who lacked 4 GP information ( 234 individuals ) , the mean PC1 score value was 0 . 043 +/− 0 . 008 . The same dataset was also examined using a Bayesian clustering algorithm ( STRUCTURE ) [21] . For these analyses we examined three sets of >3500 SNPs that were selected randomly except for the criterion that the minimum inter-SNP distance was >500 Kb ( see Methods ) . This was done to both ensure genome-wide distribution and eliminate linkage disequilibrium between SNPs . This analysis similar to our previously reported studies was most consistent with two population groups ( K = 2 ) explaining the major substructure in this set of European individuals ( Figure 1C ) . The distribution of the individuals ( K = 2 ) was similar to that shown on the first axis of the PCA ( Figure 1D ) and the individual population contributions were highly correlated with the first PC scores ( r2 > 0 . 95 for each of the three random sets compared with the for the 500K SNP data analyzed by PCA ) . We also explored whether the PCA was affected by either inclusion or exclusion of specific population groups or the number of individuals in different population groups . Most prominently , a major difference in the relationships among the populations for the second PC was observed when either Ashkenazi Jewish individuals or Irish individuals were excluded ( Figure 2 ) . These results suggest some caution in interpretation of specific clines and particular relationships among different European groups ( see discussion ) . For many association tests including candidate genes and replication studies for candidate chromosomal regions it is useful to identify smaller numbers of SNPs that can distinguish European substructure . Previous studies including our own utilized genome-wide SNP sets of ≤10K SNPs . To identify a more robust set of SNPs that could distinguish the largest component of substructure observed in the current data we used the genotypic differences observed in >300K SNPs between two groups of individuals , 150 Ashkenazi Jewish and 125 Northern European individuals . The Ashkenazi Jewish individuals were chosen since 1 ) this individual group was most clearly distinguishable from the Northern European individuals , 2 ) might more closely represent an “older” population of Mediterranean origin and 3 ) we had substantial number of genotyped individuals to enable a good representation of this population . To select the most informative SNPs distinguishing between these groups we determined the informativeness ( In ) [22] for each of >300K SNPs . The 20 , 000 SNPs with the highest In values were then selected to capture the most informative SNPs . To ensure both a more uniform genome-wide distribution and minimize linkage disequilibrium the set of putative European substructure ancestry informative markers ( ESAIMS ) were chosen to obtain the markers with highest In with a minimum inter-SNP distance >500 Kb . This resulted in a set of 1441 SNPs ( Table S1 ) . The STRUCTURE results ( K = 2 ) from individuals with 4 grandparental data ( not used for ESAIM selection ) showed separation of most of the 220 self-identified individuals of Ashkenazi Jewish heritage ( mean 83% south; median , 87% ) from 37 individuals of Western , Northern or Central heritage belonging to the “northern” group ( mean 4% south; median , 3% ) , and 51 individuals of Greek , Italian , or Spanish origin were intermediate ( mean 41% south; median , 42% ) ( Figures 3 and S1 ) . These 1441 north/south-ESAIM showed small confidence limits in the assignments; of the total of 677 individual individuals not used in ESAIM selection the maximum 90% Bayesian confidence interval ( CI ) was 21 . 1% ( e . g . 13 . 7 % south , 90% CI 2 . 6 % – 23 . 0 % ) and the median CI was 13 . 9% . Smaller north/south-ESAIM sets showed strong correlations with the 1441 set e . g . 384 ESAIMs ( r2 = 0 . 970 ) ( Figure S2 ) . However , the smaller north/south-ESAIM sets showed somewhat broader confidence limits ( e . g . 384 north/south ESAIM set showed maximum CI = 38 . 9% and a median CI = 17 . 1% . However , these differences are unlikely to affect most studies . The larger number of north/south-ESAIMs may be useful if a very homogeneous set of individuals of a particular ethnic group is desired for a specific study . Although the STRUCTURE analysis was most consistent with two population groups explaining most of the substructure within Europe , the distribution of individuals from different countries of origin along the second axis in the PCA ( Table 2; Figure 1B ) suggested that further analysis of substructure was warranted . This substructure was examined using individuals of “northern” European ancestry in the context of a large dataset of rheumatoid arthritis cases and controls ( over 2000 total individuals ) that were recently genotyped with >500K SNPs as part of the NARAC studies ( see Methods ) . For these PCA we examined only those European individuals that showed >90% membership in the northern European group by STRUCTURE analysis using the 1441 north/south-ESAIMs . This criterion closely matched the individual distribution along the first principal component axis of this dataset ( Figure S3 ) . Controlling for this first vector in analysis of cases vs . controls decreased the inflation of the median chi-square distribution using the genomic controls parameter ( λgc ) from 1 . 43 to 1 . 15 . PCA of the “north” only subset showed substantial substructure differences in the distribution of North American Rheumatoid Arthritis ( NARAC ) cases and controls along the first PC ( Figure 4 ) . Importantly , we controlled for this difference in our genome-wide association scan and excluded SNPs that showed association based on this substructure difference [23] . The distribution of individuals in this PC showed a distinct pattern with respect to the context of country of origin information that was available for a subset of control individuals ( Figure 4B ) . Most notably , Irish individuals were distinguished from those of eastern , northern and central European descent . These relationships were further defined by inclusion of additional individuals with the same country of origin genotyped with the 300K SNP set ( Table 2 ) . Similar results were also observed using a STRUCTURE analysis of the same dataset ( Table 2 ) . The results suggest that the difference in numbers of individuals of Irish ancestry was primarily responsible for the major difference in substructure observed in the NARAC cases and controls [23] . Controlling for this aspect of substructure the λgc in this individual set decreased from 1 . 15 to 1 . 07 . Since the sample set had a disproportionately large contribution of participants of Irish ancestry we also examined a small set of individuals with nearly proportionate representation of Irish , German , Eastern European , and United Kingdom individuals . Similar to the results on the larger set of individuals , these PCA results showed a west-east gradient ( Figure S4 ) . Here however , there was no difference observed between the Irish and UK individuals . Thus , these results further indicate that the number of individuals from each individual group may partially alter relationships among individual groups . Inspection of the second axis of the Northern European subset ( see Figure 4A and 4B ) , also showed an unexpected grouping of individuals on the Y axis into three separate groups . When we ascertained informative SNPs between the top and bottom groups , all of the SNPs with In values >0 . 02 were found to be located in a 3 . 8 Mb segment of human Chromosome 8 ( 8 . 135 – 11 . 936 Mb ) . This region has been previously shown to contain a common inversion within European populations [24 , 25] . When only SNPs within this interval were used the distribution of the individuals formed the same grouping of three clusters as found using the entire 500K set ( data not shown ) . As expected two dominant haplotypes ( A and B ) were ascertained with twenty selected markers with very large Ins and described the same three individual groups ( AA , AB , and BB ) and were highly correlated ( r2 = 0 . 83 ) ( Figure 5 ) . Although the λgc in the entire NARAC case-control dataset is decreased from 1 . 073 to 1 . 048 by considering this axis , our analyses indicate that the position of individuals on this axis is almost completely due to this localized inversion . This region is presumably identified by PCA because of the long stretch of linkage disequilibrium caused by the chromosomal inversion . Another set of ESAIMs ( north-ESAIMs ) was ascertained using the results of the first PC scores of the “northern” European only analysis . We selected two disparate sets of individuals comprised of 93 and 132 individuals , by randomly selecting half of the individuals with PC scores one standard deviation above the mean and half of the individuals with PC scores one standard deviation below the mean . We used this procedure to provide both a distribution of allele frequencies in the disparate individual sets as well as maintain a well distributed set of individuals for evaluating the functional performance of the putative north-ESAIMs . These ESAIMs were then selected using the same method ( In values ) and criterion ( minimum inter-SNP distance = 500 kb ) . We initially examined the best 1250 north-ESAIMs with 1608 individuals that had not been included in any of the ESAIM selections . Initial evaluation of this north-ESAIM set showed a distortion of the PCA in which the individuals were divided in three groups diagonally across the first two axes . Deletion of markers within the Chromosome 8 inversion ( see above ) resulted in a set of 1211 SNPs that no longer showed this pattern . This north-ESAIM set distinguished “northern” European individuals in a pattern similar to that observed using the 500K SNP set along the first PC ( Figure 4C and see Table S2 for SNP list ) . The PC scores using these north-ESAIMs in the “northern” European only set correlated with the 500K first PC result; r2 = 0 . 46 ( p < 10−15 ) . Smaller north-ESAIM sets showed dramatically smaller correlations if the individual individual values compared to the 500K PCA ( data not shown ) . Larger panels of SNPs ( up to 5000 SNPs ) chosen using the same criteria showed similar results to the 1211 ESAIMs . To further explore European substructure we examined additional PCs in the initial all European participant set after excluding the SNPs within the Chromosome 8 inversion . The distribution of individuals in the first two vectors in the entire group did not change . However , the third vector now showed clusters corresponding to population affiliation ( Figure 6 ) . However , this PC accounts for only very small amount of the population variation within our different sample sets ( see Table 1 ) . Although PC4 showed marginal evidence for clustering by the ANOVA test there was little apparent correlation with self-identified ancestry . Additional PCs did not show evidence for substructure by ANOVA , or a significant split half reliability test . To examine whether ESAIMs could control for European population substructure in association testing the cases and controls from the NARAC RA studies were analyzed for selected SNPs . These analyses were performed using a set of individuals that did not include those used in the ESAIM selections . The SNPs for testing were selected based on our previous results to specifically address the effect of population substructure ( Table 3 ) . These included two gene associated SNPs that showed potential false positive results in our association tests , rs1446585 for lactase ( LCT ) , and rs12203592 for interferon regulatory factor 4 ( IRF4 ) . In addition , a PTPN22 , a TRAF1 and an MHC SNP were included in the testing as positive controls since these SNPs and gene loci are believed to be RA susceptibility genes based on our studies in European populations [23] . As a comparison to the ESAIMs , the same data was analyzed using the entire 500K SNP set with the EIGENSTRAT method . As expected when the EIGENSTRAT analysis was performed using the entire SNP set , this showed strong evidence for association of the PTPN22 , TRAF1 and the MHC SNPs . Similarly , when the ESAIMs were applied either using EIGENSTRAT or a method for structured association ( STRAT ) , evidence for the association of these SNPs remained after controlling for population substructure ( Table 3 ) . For the LCT SNP in strong linkage disequilibrium with lactose intolerance the evidence for association ( p < 5E−5 ) was no longer present when the entire 500K SNPs or ESAIMs were utilized in EIGENSTRAT ( p > 0 . 05 ) or the STRAT analysis ( p > 0 . 05 ) . For IRF4 , the association becomes stronger after correcting for the north/south difference ( Eigen statistic 1 ) . However , when the second PC is considered the signal is greatly diminished . With the combined ESAIMs ( north/south and north ) , the evidence for association is also greatly diminished by the EIGENSTRAT analysis and eliminated in the STRAT analysis . We examined different sets of ESAIMs including several different combinations of north/south ESAIMs and north-ESAIMs . The results were identical when 192 north/south ESAIMs or 384 north/south ESAIMs were used for the PC1 correction ( data not shown ) . However , as expected based on our PCA results , decreasing the 1211 north-ESAIMs led to poorer PC2 correction and less complete correction of the false positive IRF4 association ( see footnote Table 3 ) . Together these results suggest the potential application of ESAIMs in association studies of candidate genes or in replication studies ( see discussion ) .
The current study provides additional insight into European substructure and differences among different ethnic groups that may impact our understanding of the genetics of complex diseases . First , together with our recent report of a whole genome association study for RA in European Americans , this report emphasizes the importance of controlling for substructure in the ascertainment of putative susceptibility associated SNPs . Most notably without an analysis of substructure , IRF4 would appear as a very strong candidate for this disease . However , the large differences in allele frequency for this gene are largely due to the difference in allele frequency among different European subpopulation groups . Furthermore , this difference is accentuated when the northern population subgroup is examined . When only NYCP controls are considered an IRF4 SNP ( rs12203592 ) showed the largest allele frequency difference between the Irish individuals and those individuals of Northern , Central European and Eastern European descent ( δ = 0 . 40 , Fst = 0 . 27 ) . Using an algorithm based on the PCs , EIGENSTRAT , this SNP no longer appears significantly associated with RA . The difference in allele frequency for IRF4 within European populations has recently also been described by the Welcome Trust Case Control Consortium study [26] . The current study extends and complements other studies showing evidence of European substructure . Overall , the current results are consistent with a major north/south ( or northwest/southeastern ) gradient as the largest difference within European groups confirming both our previous studies and others using up to 10 , 000 markers and is generally consistent with much earlier studies using classic gene-frequency data [15 , 18] . The current results differ from previous studies in defining a northern European axis that was critically important in the case control analyses [23] . The relationship between the population groups was consistent when analysis was restricted to “northern” European population groups . As discussed further below , when more disparate populations are examined ( including different “southern” populations ) these relationships are not as clearly defined ( see Figure 2 ) . Thus , differences in these results compared to other studies can in part be attributed to both inclusion of different population groups and perhaps complex relationships reflecting different population origins that includes migration , admixture , and isolation . In addition , the much larger SNP set , 300K compared to maximum of 10K SNPs in previous studies , is also likely to have exposed aspects of substructure not evident in other studies . For PC's >1 , comparison of sets of <11K SNPs had much lower correlations with the full 300K SNP set than those random sets with >40K SNPs ( Table S3 ) . Our results also suggests the potential for further definition of more homogeneous population groups for genetic studies that may theoretically decrease both type 1 and type 2 error rates . Geneticists have long recognized that different population groups may provide enhanced opportunities to uncover susceptibility loci based on more limited genetic heterogeneity . For complex genetic diseases some specific studies may focus on particular population groups to enhance the power to find important gene variants . For example , the study of Crohn's disease [#266600] in Ashkenazi Jewish individuals has the advantage of examining a potentially more homogeneous population with a higher frequency of this particular disease than in a mixed European population . This approach is supported by our results suggesting that a very large proportion of this particular ethnic group can be distinguished by analysis of substructure . Moreover , our results provide the ability to further define and restrict this study population by allowing the identification and exclusion of subgroup outliers in association tests in studies of complex genetics in Ashkenazi Jewish populations . In addition , pre-genotyping of potential cases and controls with as few as several hundred north/south-ESAIMs could enable pre-identification of a more homogenous subgroup for WGA or be utilized in candidate SNP replication studies to reduce error rates . With respect to identification of population substructure there are several limitations in the current study . First , analyses are based on a diverse set of individuals of European descent with variable ancestral contributions from different European countries that is only partially defined . This limits certain conclusions with regards to specific aspects of substructure related to population subgroups . However , we believe that the concordant grouping of the majority of participants with grandparental information provides strong support for the major relationships and differences in these population groups . The overall strong correlation between results using principal components and those using a Bayesian clustering algorithm provide additional confidence in the general results . Second , the PCA is sensitive to differences in the inclusion or exclusion of specific population groups . When the second axis is considered for the entire European group , we observed changes in the country-of-origin order for the northern group with respect to the southern group in subset analyses ( Figure 2 ) . We speculate that this observation may reflect the difference in the origins of the additional substructure in the northern group compared to the other elements of substructure in the southern group . This result suggests that overall geographic suggestions of clines based on principal components must be cautiously interpreted . Third , the PCA can be dramatically affected by differences in relatively small genomic regions that may not reflect true population substructure . This is illustrated by our finding that the second axis in the “northern” European analysis ( also observed for the third axis in the entire European set ) is dependent solely on a <4 Mb segment of Chromosome 8 that carries a common inversion . The effect of such an inversion on PCA is presumably due to a long stretch of linkage disequilibrium that is a result of non-recombination between the inverted and non-inverted chromosomal segments . The genomic distribution of particularly informative SNPs for each PC axis provides one method to inspect whether the apparent differences in substructure are due to a single or very limited number of genomic intervals . For the first two axes of the PCA the particularly informative SNPs , ESAIMs , are widely distributed ( Tables S1 and S2 ) . Deletion of subsets of particularly informative markers ( e . g . SNPs in lactase and MHC regions ) did not change the patterns observed using these ESAIMs for either PC1 or PC2 . Since we observed that the Chromosome 8 inversion affected the PCA , we also examined the common European inversion on chromosome 17 [27] . Here , deletion of this chromosomal interval had no effect on the first 10 PCs presumably due to the smaller size of this inversion , 900 kb compared to ∼4Mb for the Chromosome 8 inversion . An interesting observation in this study is that within the “northern” European population group , individuals of Irish descent showed substantial differences in substructure compared to participants of Scandinavian , Central , and Eastern Europe descent . It also appears that United Kingdom individuals were intermediate between the other non-Irish groups and those of Irish descent further supporting an east/west gradient ( Table 2 ) . However , the later observation is based on small numbers of individuals ( six 4GP United Kingdom individuals ) . It is unclear whether these relationships may reflect remnants of early populations including differences in Mesolithic or Neolithic contributions to the Irish population 5 , 000–6 , 000 years ago [28] , or later Celtic contributions . An extensive Neolithic contribution from the Iberian peninsular is consistent with Irish archeological information but it is unknown whether this population group survived [28 , 29] . As discussed above , it is difficult to determine the relationship between certain population groups and the suggestion of a cline extending from the Spanish to Irish population is tenuous based on the current data . However , we note that there is modest support for such a cline in both PC2 and PC3 ( Figures 2C and 6 ) The current study identifies SNPs that are particularly informative for European population substructure ( Tables S1 and S2 ) . This includes two SNP sets: one that distinguishes substructure along the “north/south” gradient and the other that distinguishes substructure along a west-east gradient among northern European groups tested . Together these ESAIMs appear to provide good control for subpopulation differences in the NYCP individuals as demonstrated by testing a real dataset using both EIGENSTRAT and structured association methods . Additional studies will be necessary to further optimize ESAIM sets and in particular to determine their efficacy in additional European and European American sample groups that may have different ancestral representation . Finally , it is worth noting that particularly informative ESAIMs may correspond to population selection events and hence also be linked to important biologic processes . The most informative locus for the “north/south” distinction , a lactase gene associated SNP , has been previously noted in this regard [6] . Another strong candidate for selection includes the IRF4 gene that is an important immunologic response regulator [30–33] , and ongoing studies are examining these and other genes for evidence of positive selection in different subgroups . | Ancestry differences corresponding to ethnic groups may be important in determining disease risk factors and optimizing treatment . Our study further defines ancestry relationship among different European ethnic groups by examining over 300 thousand variations in DNA , in over 2 , 000 individuals . This study allowed a clearer ascertainment of differences that could not be discerned in smaller studies using more limited numbers of DNA variations . We show clear differences among European American participants of different self-identified ethnic affiliation . The analyses showed multiple components of variation . The components showing the largest variations generally corresponded to the grandparental country or region of origin within Europe . We also show the importance of applying this information in determining genetic risk factors for complex diseases . Moreover , the results have enabled a better selection of smaller numbers of DNA variations that can be used in future disease studies to identify more homogenous participant groups and minimize false positive and false negative results in assessing genetic risk factors for disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion"
]
| [
"homo",
"(human)",
"genetics",
"and",
"genomics"
]
| 2008 | Analysis and Application of European Genetic Substructure Using 300 K SNP Information |
In response to DNA damage , the eukaryotic genome surveillance system activates a checkpoint kinase cascade . In the fission yeast Schizosaccharomyces pombe , checkpoint protein Crb2 is essential for DNA damage-induced activation of downstream effector kinase Chk1 . The mechanism by which Crb2 mediates Chk1 activation is unknown . Here , we show that Crb2 recruits Chk1 to double-strand breaks ( DSBs ) through a direct physical interaction . A pair of conserved SQ/TQ motifs in Crb2 , which are consensus phosphorylation sites of upstream kinase Rad3 , is required for Chk1 recruitment and activation . Mutating both of these motifs renders Crb2 defective in activating Chk1 . Tethering Crb2 and Chk1 together can rescue the SQ/TQ mutations , suggesting that the main function of these phosphorylation sites is promoting interactions between Crb2 and Chk1 . A 19-amino-acid peptide containing these SQ/TQ motifs is sufficient for Chk1 binding in vitro when one of the motifs is phosphorylated . Remarkably , the same peptide , when tethered to DSBs by fusing with either recombination protein Rad22/Rad52 or multi-functional scaffolding protein Rad4/Cut5 , can rescue the checkpoint defect of crb2Δ . The Rad22 fusion can even bypass the need for Rad9-Rad1-Hus1 ( 9-1-1 ) complex in checkpoint activation . These results suggest that the main role of Crb2 and 9-1-1 in DNA damage checkpoint signaling is recruiting Chk1 to sites of DNA lesions .
Maintaining genome integrity requires the DNA damage checkpoint signaling pathways , which typically involve a protein kinase cascade . In vertebrates , residing at the top of the signaling pathways are two members of the phosphatidylinositol 3-kinase-like protein kinase ( PIKK ) family , ATM ( ataxia-telangiectasia , mutated ) and ATR ( ATM and Rad3 related protein kinase ) [1] , [2] . The activation of ATM and ATR upon DNA damage leads to the phosphorylation and activation of downstream effector kinases Chk2 and Chk1 , which then regulate a myriad of cellular processes including cell cycle progression [3] . In the fission yeast Schizosaccharomyces pombe , the ATM ortholog Tel1 plays a minor role in checkpoint signaling at DSBs except when DNA end resection is inhibited [4] . The major checkpoint signaling activity is provided by the ATR ortholog Rad3 , which is essential for the signaling through both Chk1 and Chk2 ( called Cds1 in fission yeast ) [5]–[7] . The activation of Cds1 requires the checkpoint mediator protein Mrc1 [8] , [9] . The role of Mrc1 in checkpoint signaling is to recruit Cds1 to stalled replication forks , and this recruitment function relies on multiple Rad3-depedent phosphorylation sites on Mrc1 , which serve as docking sites for Cds1 [10] . Acting in between Rad3 and Chk1 is the checkpoint mediator protein Crb2 [11] . Loss of Crb2 completely abolishes Chk1 activation but does not affect Cds1 activation . The requirement for a checkpoint mediator acting upstream of Chk1 is universally conserved . In the budding yeast Saccharomyces cerevisiae , the ortholog of Crb2 , scRad9 , is also vital for the activation of Chk1 [12] , [13] . Intriguingly , in vertebrates , the ortholog of Crb2 and scRad9 , 53BP1 , is dispensable for Chk1 activation . Instead , vertebrate Chk1 activation involves a different mediator , Claspin , which shares no significant homology with Crb2 or scRad9 [14] . The spatial regulation of Crb2 in response to DSBs has served as an instructive model for understanding the influence of chromatin modification on checkpoint mediator localization . When DSBs are generated by ionizing radiation ( IR ) , Crb2 rapidly forms microscopically detectable IR-induced foci ( IRIF ) , and these foci overlap with DSB markers [15] . IRIF formation by Crb2 requires two types of histone modifications: one is histone H2A C-terminal tail phosphorylation catalyzed redundantly by Rad3 and Tel1 kinases [16]; the other is H4-K20 methylation catalyzed by Set9 methyltransferase [17] . Both types of modifications promote histone-Crb2 interactions . Phosphorylated H2A binds to the tandem BRCT domains in the C-terminus of Crb2 , which also mediate Crb2 dimerization [18] , [19] . K20-methylated H4 interacts with the tandem Tudor domains located N-terminally to the BRCTs [20] . Disrupting either histone-Crb2 interaction dramatically diminishes Crb2 IRIF but only partially impairs checkpoint signaling , indicating the existence of an alternative pathway of recruiting Crb2 to DSBs that does not involve large-scale interactions with chromatin . Indeed , by examining the relocalization of Crb2 to a persistent DSB induced by the HO endonuclease , we previously identified a histone modification-independent Crb2 recruitment pathway , which requires an interaction between Crb2 and Rad4/Cut5 [21] . Rad4/Cut5 is a multi-BRCT domain protein with dual functions in DNA replication and DNA damage checkpoint signaling [22] . The N-terminal tandem BRCT domains in Rad4/Cut5 mediate the interaction with Crb2 [11] , and this interaction requires the Crb2-T215 residue , a cyclin-dependent kinase ( CDK ) phosphorylation site [21] , [23] . The second pair of BRCT domains in Rad4/Cut5 mediates an interaction with Rad9 ( unrelated to scRad9 ) , which is a subunit of the 9-1-1 checkpoint clamp complex [24] . As the 9-1-1 complex can directly associate with DNA at sites of DSBs [25] , the interactions between Rad9 and Rad4/Cut5 , and between Rad4/Cut5 and Crb2 , provide a means to recruit Crb2 independently of histone-Crb2 interactions . Compared to our knowledge on how Crb2 is targeted to sites of DNA damage , much less is known about the molecular mechanism by which Crb2 mediates Chk1 activation . Crb2 is hyperphosphorylated upon DNA damage in a Rad3-dependent manner [11] , but the Rad3-dependent phosphorylation sites on Crb2 have not been identified . It has also been shown by yeast two-hybrid assay and co-immunoprecipitation under overexpression conditions that Crb2 can interact with Chk1 [11] , [26] , but the functional significance of such interactions remains unknown . In S . cerevisiae , the role of scRad9 in Chk1 activation is also not well understood , except that an N-terminal region of scRad9 is required [13] . In vertebrates , phosphorylated Claspin interacts with the kinase domain of Chk1 to facilitate its activation by ATR [27]–[31] . Due to the lack of homology between Claspin and Crb2/scRad9 , it is uncertain whether a common mechanism exists for mediators acting upstream of Chk1 . In this study , we show that Crb2 directly interacts with Chk1 in a phosphorylation dependent manner . Two neighboring SQ/TQ motifs in Crb2 , which are consensus sites for ATM/ATR kinases , are critical for Crb2-Chk1 interactions , Chk1 relocalization to DSBs , and DNA damage-induced checkpoint activation . Tethering a Chk1-binding Crb2 peptide to sites of DSBs can bypass endogenous Crb2 and 9-1-1 complex for checkpoint activation , suggesting that the main function of these proteins in DNA damage checkpoint activation is recruiting Chk1 .
When DSBs occur , the checkpoint mediator Crb2 accumulates at sites of DSBs , forming nuclear foci detectable by fluorescence microscopy [15] , [21] . However , it is unclear to what extent its downstream effector Chk1 is concentrated at DSBs . To test if Chk1 relocalizes to DSBs , we examined the subcellular distribution of C-terminally GFP-tagged Chk1 ( Chk1-GFP ) by live cell imaging . Expressed from the endogenous promoter , Chk1-GFP is fully functional , as demonstrated by its ability to confer wild-type levels of DNA damage resistance ( Figure S1A ) . Chk1-GFP displayed diffuse nuclear signal in the absence of DNA damage ( Figure S1B ) . IR induced the formation of Chk1-GFP nuclear foci in asynchronized cells ( 15% of nuclei contained foci ) , and these foci significantly overlapped with CFP-Crb2 foci in the same cells ( Figure 1A ) . Synchronizing cells to early S-phase with the DNA synthesis inhibitor hydroxyurea ( HU ) followed by IR exposure and release from HU block ( S-phase IR treatment ) led to higher frequency ( 79% of nuclei contained foci ) and stronger intensity of Chk1 foci compared to IR treatment of asynchronized cells ( Figure 1A and 1B ) . S-phase IR treatment is expected to result in irreparable DSBs , due to the lack of homologous recombination templates if DSBs occur in unreplicated genomic regions . Such persistent DSBs are known to cause high-level recruitment of Crb2 in a Rad4/Cut5-dependent , histone modification-independent manner [21] . Besides IR , we also used the HO endonuclease to generate a specific DSB in the genome [15] . In response to HO expression , Chk1-GFP formed a single distinct nuclear focus in the majority of the cells , and HO-induced Chk1 foci were completely colocalized with foci formed by a DSB repair protein , Rad22 ( homolog of budding yeast Rad52 ) ( Figure 1A and 1B ) . Similar to the S-phase IR treatment , HO endonuclease cleavage also induces irreparable DSBs , as both sister chromatids are cut by the enzyme . Due to prolonged G2 arrest , both S-phase IR treatment and HO expression led to significant cell elongation , which serves as a useful readout for checkpoint activation ( Figure 1A ) . Together , our observations suggest that Chk1 is recruited to DSBs , and the level of recruitment is enhanced at persistent DSBs . Chk1 activation requires the upstream kinase Rad3-Rad26 , the 9-1-1 checkpoint clamp , the clamp loader Rad17 , Rad4/Cut5 , and the mediator protein Crb2 [5] , [11] . To examine the genetic requirement of DSB-induced Chk1 focus formation , the genes encoding upstream checkpoint factors were individually deleted in strains expressing Chk1-GFP . IR-induced Chk1 foci were not observed in rad3Δ , rad9Δ and crb2Δ cells ( Figure S2 and Figure 1C ) , suggesting that Chk1 relocalization is regulated by the same upstream factors controlling its activation . We only examined asynchronized cells for rad3Δ and rad9Δ mutants , because S-phase synchronization by HU cannot be performed due to the lack of replication checkpoint . To understand how Crb2 facilitates Chk1 relocalization , we examined Chk1 focus formation in crb2Δ cells expressing truncated forms of Crb2 , Crb2 ( 1–358 ) -LZ or Crb2 ( 276–778 ) . Crb2 ( 1–358 ) -LZ , which lacks the histone-binding Tudor domains and BRCT domains , but is supplemented with a heterologous leucine zipper ( LZ ) dimerization motif , forms nuclear foci at persistent DSBs by binding to Rad4/Cut5 [21] . Crb2 ( 276–778 ) , on the other hand , forms transient IRIF in a histone modification-dependent manner [21] . crb2 ( 1–358 ) -LZ cells challenged with S-phase IR treatment formed Chk1 foci with dynamics similar to wild type ( Figure 1C ) . In contrast , no Chk1 foci were observed in crb2 ( 276–778 ) cells ( Figure 1C ) . As previously reported [21] , Crb2 ( 1–358 ) -LZ was sufficient for checkpoint activation that led to G2 arrest and cell elongation , whereas Crb2 ( 276–778 ) failed to mediate a checkpoint response and cells entered mitosis with unrepaired DSBs ( Figure 1C ) . Thus , the structure-function relationship for the role of Crb2 in Chk1 relocalization parallels that for the checkpoint function of Crb2 , namely , the first 275 amino acids of Crb2 is necessary and Crb2 ( 1–358 ) -LZ is sufficient . To identify the sequence elements important for Chk1 recruitment and activation in Crb2 ( 1–358 ) -LZ , sequences of S . pombe Crb2 and its homologs from three other fission yeast species were inspected . The N-terminal region of Crb2 lacks significant homology even among these closely related homologs , except for some short stretches of conserved amino acids . One of these short conserved stretches contains an invariant LTQLFE motif followed by an SQ or TQ motif two amino acids downstream ( Figure 2A ) . Previous studies have suggested that clustered SQ/TQ motifs , often phosphorylated by ATM/ATR kinases upon DNA damage , can bridge protein-protein interactions and thereby play important roles in DNA damage response [32] . Thus , we hypothesized that these two conserved SQ/TQ motifs may be involved in the checkpoint function of Crb2 . To test the functional importance of these two SQ/TQ motifs , we substituted one or both of the phosphorylatable residues in these motifs , threonine 73 ( T73 ) and serine 80 ( S80 ) , with alanine ( s ) . The two single-residue mutants , denoted as crb2-T73A and crb2-S80A , displayed mild sensitivity to various types of DNA damage ( Figure 2B ) . They were significantly more sensitive than wild type when treated with higher doses of UV , HU , and CPT , but were much more resistant than either chk1Δ or crb2Δ at all doses tested . The strain with both T73 and S80 mutated , denoted as crb2-2AQ , on the other hand , showed much stronger sensitivity than the single-residue mutants . It appeared to be as sensitive to HU and CPT as chk1Δ , and only slightly more resistant to UV and IR than chk1Δ ( Figure 2B and Figure S3A ) . The strong synergistic effect of combining the two mutations suggests that these two SQ/TQ motifs may play partially redundant roles in the checkpoint function of Crb2 . In a cdc25-22 block-and-release assay , irradiated crb2-2AQ cells entered mitosis as soon as crb2Δ cells upon releasing from a G2 block , suggesting a strong defect in checkpoint arrest ( Figure S4A ) . In contrast , both crb2-T73A and crb2-S80A delayed the mitotic entry significantly , although not as long as the wild type ( Figure S4A ) . To analyze Chk1 phosphorylation and activation , we then examined the DNA damage-induced mobility shift of Chk1 on SDS-PAGE [5] . Chk1 extracted from DNA-damage-treated wild-type cells showed two bands , the upper one corresponding to the phosphorylated form of Chk1 and the lower one corresponding to the unmodified form ( Figure 2C and Figure S3B ) . Only the lower band was observed in either crb2Δ or crb2-2AQ ( Figure 2C and Figure S3B ) . Consistent with the milder sensitivity and checkpoint defect of single-residue mutants , Chk1 phosphorylation in crb2-T73A or crb2-S80A was still detectable but weaker than wild type ( Figure 2C and Figure S3B ) . Together , these results suggest that this conserved stretch of residues with two SQ/TQ motifs , which we will thereafter refer to as the SQ/TQ cluster , plays a critical role in Chk1 activation . To understand how the SQ/TQ cluster contributes to Chk1 activation , we examined whether the mutations at the SQ/TQ cluster affect the DNA damage-induced relocalization of Chk1-GFP . To simultaneously monitor the localization of Crb2 in the same cells , we used strains expressing CFP-tagged Crb2 as the only version of Crb2 . Upon IR treatment of S-phase cells , Chk1-GFP formed distinct nuclear foci in cells expressing wild-type CFP-Crb2 , and these foci completely overlapped with Crb2 foci ( Figure 2D ) . The frequencies of detecting Chk1 foci dramatically decreased in cells expressing Crb2-T73A or Crb2-S80A , and only in a small minority of these cells ( about 3% ) could we see very faint Chk1 foci , which were also colocalized with Crb2 foci . No Chk1 foci could be detected in cells expressing Crb2-2AQ . In contrast to the strong reduction of Chk1 focus formation , the three mutant forms of Crb2 themselves showed robust focus formation like wild-type Crb2 ( Figure 2D ) . To rule out the possibility that an effect on Crb2 recruitment was masked by the redundancy between the two Crb2 recruitment pathways , we examined the localization of Crb2 ( 1–358 ) -LZ and found that its focus formation was also unaffected by the 2AQ mutations ( Figure S5 ) . Thus , the Crb2 SQ/TQ cluster is not important for the relocalization of Crb2 itself , but rather specifically controls the accumulation of Chk1 at DSBs . In agreement with the checkpoint defect detected by the cdc25-22 block-and-release assay and the inability to support Chk1 phosphorylation , crb2-2AQ cells did not elongate after the S-phase IR treatment and displayed the “cut” ( cell untimely torn ) phenotype ( Figure 2D , Figure S4B and S4C ) , indicating a severe defect in G2 arrest . In contrast , cells expressing Crb2-T73A or Crb2-S80A became substantially elongated , consistent with their partial proficiency in Chk1 phosphorylation . We speculate that Chk1 molecules were recruited to DSBs in crb2-T73A or crb2-S80A cells at a level high enough for partial checkpoint activation but too low to be clearly distinguished from the diffuse nucleoplasmic Chk1-GFP signal . Crb2 is known to undergo DNA damage-induced hyperphosphorylation , which manifests as mobility shift on SDS-PAGE [11] , [18] , [26] . To assess whether the SQ/TQ cluster contributes to Crb2 phosphorylation , we examined the DNA damage-induced mobility shift of Crb2 . The 2AQ mutations significantly reduced but did not abolish the mobility shift of Crb2 induced by IR or camptothecin ( CPT ) ( Figure 3A ) . We hypothesized that other SQ/TQ motifs may contribute to the residual shift in 2AQ mutant , as there are a total of 11 SQ/TQ motifs in Crb2 ( Figure 3B ) . Thus , we mutated all remaining SQ/TQ motifs except S666 , which plays a critical structural role at the Crb2 dimer interface and is unlikely to be a phosphorylation site [19] , [26] . The resulting 8AQ mutant did not show any DNA damage sensitivity ( Figure S6 ) , suggesting that T73 and S80 are the only functionally important ATM/ATR consensus sites . Compared to wild-type Crb2 , the 8AQ mutant displayed a less pronounced IR-induced shift , which could be further reduced by the mutations at T73 and/or S80 ( Figure 3C ) . Residual mobility shift was still observed with the 10AQ ( 8AQ+2AQ ) mutant , indicating that DNA damage-induced phosphorylation may also occur on non-SQ/TQ sites . The contribution of T73 and S80 to Crb2 mobility shift suggests that they are phosphorylated in vivo after DNA damage . Because Crb2 mobility shift is dependent on Rad3 [11] , Rad3 is most likely the kinase phosphorylating these sites . To obtain more direct evidence on the phosphorylation of the SQ/TQ cluster , we attempted to create phosphorylation-specific antibodies as well as tried to use an anti-phospho-SQ/TQ antibody . However , we failed to detect Crb2 phosphorylation with these reagents , presumably because the antibodies did not have sufficiently high titers . We then resorted to mass spectrometry for detecting phosphorylation sites on Crb2 . TAP-tagged Crb2 purified from cells exposed to IR was digested by three different proteases to increase the peptide coverage . Using a 12-step multidimensional protein identification technology ( MudPIT ) procedure [33] , we identified the phosphorylation of S80 residue ( Figure 3D ) . The failure to detect T73 phosphorylation by mass spectrometry could be due to the difficulty of generating suitable peptides covering that site . As T73 and S80 on Crb2 are critical for Chk1 recruitment to DSBs , we hypothesized that their phosphorylation may facilitate a direct interaction between Crb2 and Chk1 . To test this possibility , Crb2 ( 67–85 ) , a 19-amino-acid peptide containing the two functionally important SQ/TQ motifs , together with its mono- and di-phosphorylated forms were synthesized and used in in vitro Chk1 pull-down assays ( the amino acid sequence of the peptide is as depicted in Figure 2A ) . Full-length Chk1 cannot be easily expressed in bacterial cells [34] . Therefore , as input , we used YFP-Flag-His6 ( YFH ) -tagged Chk1 affinity-purified from S . pombe cells with anti-Flag beads [35] . By Coomassie staining of proteins in a polyacrylamide gel , we found that more than 5% of the Chk1 in the input was pulled down by the di-phosphorylated peptide , Crb2 ( 67–85 ) pT73pS80 ( Figure 4A ) . The peptide mono-phosphorylated at T73 , Crb2 ( 67–85 ) pT73 , pulled down slightly lower amount of Chk1 than the di-phosphorylated peptide . Crb2 ( 67–85 ) pS80 showed significantly weaker affinity but still pulled down clearly visible amount of Chk1 . In contrast , unphosphorylated Crb2 ( 67–85 ) did not pull down a detectable amount of Chk1 . As Chk1 was the only dominant band in the phosphopeptide pull-down lanes on the Coomassie-stained gel , we surmise that Chk1 probably bound to the phosphopeptides directly , and if there were other proteins bridging the interactions , they had to act in a highly substoichiometric manner . The abilities of both mono- and di-phosphorylated forms of Crb2 peptides to bind Chk1 in vitro are consistent with the data that mutating either T73 or S80 only partially affected Chk1 activation in vivo . To better quantitate the affinity difference between the phosphorylated and unphosphorylated peptides , we repeated the pull-down assay and used the more sensitive immunoblotting method to estimate the levels of peptide-bound Chk1 ( Figure 4B ) . Using serial dilutions of input as standards , we determined that Crb2 ( 67–85 ) pT73pS80 and Crb2 ( 67–85 ) pT73 pulled down about 7% of the input , whereas Crb2 ( 67–85 ) pS80 pulled down about 1% of the input . Again , we were not able to detect any Chk1 signal in the eluate from the unphosphorylated peptide , but could only estimate that if there was any Chk1 , the amount had to be lower than 0 . 08% of the input ( Figure 4B ) . The phosphopeptide binding by Chk1 not only requires a phosphate group on the peptide but also is sequence context dependent , as a phosphorylated histone H2A peptide can pull down Crb2 but not Chk1 ( Figure S7 ) . Together , these results suggest that phosphorylation of the SQ/TQ cluster on Crb2 promotes a direct and specific interaction between Crb2 and Chk1 . If the only defect of Crb2-2AQ mutant is its inability to engage a physical interaction with Chk1 , we predicted that by artificially tethering Crb2 and Chk1 together , we might be able to rescue this defect . To test this possibility , we constructed a strain expressing a Chk1-Crb2 fusion protein as the only version of Crb2 in the cells . This strain was nearly as resistant as wild type to a wide range of genotoxins , suggesting that fusing Crb2 with Chk1 did not significantly attenuate Crb2 functions or otherwise grossly perturb checkpoint signaling ( Figure 4C ) . Remarkably , in the same genetic background , when we mutated both T73 and S80 , the resulting strain , chk1-crb2-2AQ , behaved exactly like the strain expressing the wild-type fusion protein ( Figure 4C ) , indicating that the defect caused by the 2AQ mutations was completely rescued by the enforced interaction between Crb2 and Chk1 . Together with the in vitro binding data , these results suggest that the only essential role of the Crb2 SQ/TQ cluster is to promote a phosphorylation-dependent interaction between Crb2 and Chk1 . It has been shown in mammalian cells that checkpoint effector kinases Chk2 and Chk1 are phosphorylated and activated at sites of DNA damage [36]–[38] . Thus , a parsimonious model for the action of a checkpoint mediator like Crb2 calls for two , and only two , essential functions: first , it needs to recognize the DNA lesions by binding to DNA damage sensors or other upstream signaling components; second , it should be able to interact with the downstream effector kinase and bring it to sites of DNA damage . Such a model has not been formally demonstrated for any checkpoint mediators because it is not yet clear whether these two functions are imparted by separable parts of a mediator . Our previous study has established that Crb2 relocalization to DSBs requires sequence features outside of the SQ/TQ cluster , such as the T215 residue and the C-terminal histone-binding domains [21] . Here we show that the Crb2 SQ/TQ cluster is dispensable for Crb2 relocalization , but is essential for the Crb2-Chk1 interaction . Thus , we postulated that Crb2 may conform to a modular organization and has domains separately responsible for the DSB targeting function and the effector recruitment function . As the Crb2 ( 67–85 ) phosphopeptide is sufficient for Chk1 binding in vitro , we envisioned that by artificially tethering this peptide to DSBs , where the Rad3-mediated phosphorylation of this peptide presumably can happen , we may be able to bypass the need for the remaining portion of Crb2 that provides the DSB targeting function . To test this hypothesis , we fused Crb2 ( 67–85 ) to Rad22 protein . As a DNA repair protein recruited to DSBs independently of checkpoint factors , Rad22 can form IR-triggered foci in crb2Δ cells ( Figure 5A ) . Chk1 distribution after IR treatment was monitored in crb2Δ cells expressing the Rad22 fusion protein . Remarkably , in crb2Δ rad22-crb2 ( 67–85 ) cells , Chk1-GFP formed distinct nuclear foci , which were colocalized with the foci formed by the mCherry-tagged Rad22 fusion protein . In fact , the Chk1 foci in these cells were much brighter than the foci we saw in similarly IR-treated wild-type cells , probably due to the higher local concentration of Rad22 protein at DSBs than that of Crb2 . In contrast , crb2Δ rad22-crb2 ( 67–85 ) -2AQ cells behaved exactly like crb2Δ in that no Chk1 foci were detected ( Figure 5A ) . Thus , the SQ/TQ cluster represented by the 19-amino-acid peptide Crb2 ( 67–85 ) , when targeted to DSBs , is sufficient for mediating Chk1 relocalization in a manner that maintains the need for the two SQ/TQ motifs , presumably due to the phosphorylation-dependent nature of the Crb2-Chk1 interaction . Even though we were not able to detect SQ/TQ cluster phosphorylation on endogenous Crb2 using phospho-specific antibodies , the strong Chk1 foci in rad22-crb2 ( 67–85 ) cells prompted us to attempt this approach again on the SQ/TQ cluster peptide fused to Rad22 . In an immunoblot analysis , a commercially available anti-pSQ/TQ antibody reacted with Rad22-Crb2 ( 67–85 ) immunoprecipitated from cells treated with CPT , but did not react with Rad22-Crb2 ( 67–85 ) from untreated cells , nor with Rad22 alone ( Figure 5B ) , suggesting that T73 and/or S80 residues were phosphorylated in response to DNA damage . The DNA damage-inducible nature of the SQ/TQ cluster phosphorylation is consistent with our preposition that T73 and S80 are substrate sites of Rad3 kinase , the only ATM/ATR family kinase essential for checkpoint signaling in fission yeast . To further verify this hypothesis , we examined the phosphorylation of Rad22-Crb2 ( 67–85 ) in rad3Δ mutant . As predicted , the phosphorylation-specific immunoblot signal was abolished in rad3Δ cells ( Figure 5C ) . Another prediction we can make is that Rad3 should be required for Rad22 fusion-mediated Chk1 accumulation at DSBs . Indeed , we found that rad3Δ abolished Chk1 foci in crb2Δ rad22-crb2 ( 67–85 ) cells ( Figure 5A ) . We and others have not been able to detect the physical interactions between endogenous Chk1 and Crb2 , most likely due to the transient nature of the interactions [26] . However , in accordance with the strong Chk1-GFP foci we observed in rad22-crb2 ( 67–85 ) cells , we found that Chk1 could be co-immunoprecipitated with Flag-tagged Rad22-Crb2 ( 67–85 ) , in a manner dependent on the SQ/TQ motifs and Rad3 kinase ( Figure 5D ) . To assess the functional consequences of Chk1 relocalization mediated by Rad22-Crb2 ( 67–85 ) , we analyzed the DNA damage sensitivity of cells expressing Rad22-Crb2 ( 67–85 ) . In crb2+ background , expressing this fusion protein as the only version of Rad22 did not significantly enhance the sensitivity , suggesting that the DNA repair function of Rad22 was not grossly compromised by the fusion ( Figure 5E ) . In crb2Δ background , cells expressing Rad22-Crb2 ( 67–85 ) showed stronger resistance to UV , IR , and CPT treatment compared to cells expressing Rad22 not fused with Crb2 peptide . We note that this rescuing effect was incomplete , as the cells were still more sensitive than the crb2+ strain . This partial rescue requires the SQ/TQ motifs , as the crb2Δ cells expressing Rad22-Crb2 ( 67–85 ) -2AQ did not show improved genotoxin resistance ( Figure 5E ) . crb2Δ cells expressing Rad22-Crb2 ( 67–85 ) appeared to be capable of checkpoint arrest as they became significantly elongated after DNA damage treatment ( Figure 5A ) . To more directly monitor checkpoint arrest , we performed a cdc25-22 block-and-release assay . Cells synchronized in G2 by the temperature-sensitive cdc25-22 mutation were irradiated with IR and then released to permissive temperature to allow mitotic entry . crb2Δ cells rapidly entered mitosis after the release , whereas wild type cells showed a checkpoint response as their mitotic entry was delayed for 2 h compared to crb2Δ cells ( Figure 5F ) . Strikingly , crb2Δ cells expressing Rad22-Crb2 ( 67–85 ) did not enter mitosis during the observation period of more than 3 h , suggesting that they were capable of a robust checkpoint arrest . The prolonged arrest could be due to slower DNA repair , or defective checkpoint recovery , or a combination of both . This checkpoint arrest is entirely dependent on Chk1 , as chk1Δ completely abolished the mitotic delay ( Figure 5F ) . Mutating the two SQ/TQ motifs in the Rad22 fusion protein also rendered the cells completely defective in checkpoint arrest ( Figure 5F ) . Together , our observations suggest that artificially tethering Crb2 ( 67–85 ) to DSBs by a Rad22 fusion is sufficient for recruiting Chk1 to DSBs and activating a Chk1-dependent checkpoint response in the absence of endogenous Crb2 . The PCNA clamp-like 9-1-1 complex , composed of Rad9 , Rad1 , and Hus1 , and its clamp loader , the Rad17-RFC complex , are essential for DNA damage-induced Chk1 phosphorylation and activation by Rad3 [5] . The roles of 9-1-1 complex in Chk1 activation are not entirely clear . On one hand , it may act as a recruitment platform for downstream factors at DSBs by interacting with Rad4/Cut5 [24] , which in turn binds to Crb2 [11] , [21] , and eventually brings Chk1 to the proximity of Rad3 at DSBs through the Crb2-Chk1 interaction we report here . Consistent with this model , we found that Chk1 focus formation requires Rad9 ( Figure S2 ) , and deletion of Rad9 or Rad17 abolished Rad4/Cut5 accumulation at DSBs ( Figure S8 ) . However , on the other hand , it has also been shown that the orthologs of Rad9 and Rad4/Cut5 in budding yeast and the ortholog of Rad4/Cut5 in vertebrates possess the so-called ATR-activating domains , which are sufficient for activating ATR kinase in vitro and important for checkpoint activation in vivo [39]–[45] . Thus , it is possible that in fission yeast , 9-1-1 complex enhances Rad3 activity either directly or indirectly through recruiting Rad4/Cut5 . Our finding that artificially tethering a Crb2 ( 67–85 ) peptide to DSBs is sufficient for Chk1 relocalization and activation provides a means to assess the functional importance of the two proposed roles of 9-1-1 in checkpoint signaling . We anticipated that Rad22-Crb2 ( 67–85 ) may be able to bypass the Crb2-recruitment role of 9-1-1 . However , if the putative ATR-activating role of Rad9 and Rad4/Cut5 is essential for the phosphorylation of either Crb2 SQ/TQ cluster or Chk1 by Rad3 , Rad22-Crb2 ( 67–85 ) would not be sufficient for Chk1 activation in the absence of 9-1-1 . As a readout for Rad3-dependent phosphorylation of Crb2 SQ/TQ cluster , we examined Chk1 localization in rad9Δ crb2Δ cells expressing Rad22-Crb2 ( 67–85 ) . Remarkably , Chk1 formed nuclear foci colocalizing with the Rad22 fusion protein ( Figure 5A ) , suggesting that Rad3 is able to phosphorylate Crb2 SQ/TQ cluster in the absence of 9-1-1 if we bypass the recruitment function of 9-1-1 . We went on to test whether checkpoint can be activated under such a circumstance . We found that , in a cdc25-22 block-and-release experiment , rad9Δ crb2Δ rad22-crb2 ( 67–85 ) cells arrested cell cycle progression in response to IR treatment ( Figure 5F ) . This result indicates that abolishing 9-1-1 functions did not block checkpoint activation when Chk1 is recruited to DSBs by Rad22-Crb2 ( 67–85 ) . Thus , the main DNA damage checkpoint function of 9-1-1 appears to be recruiting Crb2 and , in turn , Chk1 . Rad22-Crb2 ( 67–85 ) fusion did not fully rescue crb2Δ , possibly due to the differences in DSB recruitment kinetics and local microenvironment compared to native Crb2 . As Rad4/Cut5 recruits Crb2 to DSBs , we reasoned that fusing Crb2 ( 67–85 ) to Rad4/Cut5 may allow a better bypass of endogenous Crb2 . As we expected , expressing Cut5-Crb2 ( 67–85 ) in crb2Δ cells rescued the Chk1 focus formation defect and restored the checkpoint signaling as the cells arrested in a mononucleated , elongated state after DNA damage treatment ( Figure 6A ) . In contrast , crb2Δ cut5-crb2 ( 67–85 ) -2AQ cells behaved exactly like crb2Δ in that no Chk1 foci were detected and the cells failed to arrest in response to DNA damage ( Figure 6A ) . In a cdc25-22 block-and-release assay , we found that crb2Δ cut5-crb2 ( 67–85 ) cells delayed mitosis after IR treatment to the same extent as wild type , whereas crb2Δ cut5-crb2 ( 67–85 ) -2AQ cells resumed cell cycle progression as quickly as crb2Δ cells ( Figure 6B ) . Consistent with the rescuing of the checkpoint defect , hypersensitivities of crb2Δ to several genotoxins were fully rescued by expressing Cut5-Crb2 ( 67–85 ) ( Figure 6C ) . Furthermore , Chk1 phosphorylation defect of crb2Δ was substantially alleviated by expressing Cut5-Crb2 ( 67–85 ) ( Figure 6D ) . Together , these data suggest that the DSB targeting function fulfilled by Crb2 sequence outside of amino acids 67–85 can be completely bypassed by a Rad4/Cut5 fusion , and the Crb2 ( 67–85 ) peptide , when properly targeted to DSBs , can carry out all the checkpoint functions of full-length Crb2 .
Multiple lines of evidence suggest that T73 and S80 residues in Crb2 are phosphorylated in response to DNA damage . First , the DNA damage-induced Crb2 mobility shift was significantly diminished by 2AQ mutations in both wild-type and 8AQ mutant context . Second , anti-phospho-SQ/TQ antibody specifically recognized the Crb2 ( 67–85 ) peptide fused to Rad22 after DNA damage in a Rad3-dependent manner . Third , the requirement of these residues for the co-immunoprecipitation of Rad22-Crb2 ( 67–85 ) and Chk1 , and the rescue of the 2AQ mutations by a Chk1-Crb2 fusion strongly suggest that these residues mediate a Crb2-Chk1 interaction in vivo , and correspondingly , the in vitro interaction between the Crb2 ( 67–85 ) peptide and Chk1 requires the phosphorylation of at least one of these residues . Fourth , mass spectrometry analysis showed that the S80 residue is phosphorylated in vivo . Even though we did not obtain direct evidence that T73 residue is phosphorylated in vivo , there are good reasons to believe this is the case . First , T73 is in a conserved LxLTQLFE motif , which fits the preference of ATR kinase for hydrophobic residues at the −1 and −3 positions of it substrate sites [46] . Second , the Crb2 ( 67–85 ) peptide singly phosphorylated at the T73 residue showed robust binding to Chk1 in vitro . To obtain additional corroborating evidence , we have attempted to create phospho-mimetic mutants , but substituting both of these residues to either glutamate or aspartate resulted in the same phenotypes as the 2AQ mutant ( our unpublished observations ) , suggesting that proper checkpoint mediator function of Crb2 needs phosphorylation and not simply negatively charged side chains at these positions . Neither T73A nor S80A mutation alone strongly affected the checkpoint mediator function of Crb2 , whereas the 2AQ mutations completely abolished Chk1 recruitment and activation , indicating that these two phosphorylation sites play redundant roles . Correspondingly , the Crb2 ( 67–85 ) peptide phosphorylated on either T73 or S80 is able to pull down Chk1 . The weaker in vitro binding affinity of S80-phosphorylated peptide suggests that once the binding strength is above a certain minimal threshold , Crb2 is able to fulfill its role in recruiting Chk1 to DSBs . Alternatively , our in vitro binding assay conditions may have not faithfully mimicked the in vivo environment and underestimated the true Chk1-binding ability of S80-phosphorylated Crb2 . The conservation of Crb2 SQ/TQ cluster may not be restricted to the fission yeast species . A pair of neighboring SQ/TQ sites in a similar sequence context also exists in Crb2 orthologs in many other Ascomycota fungi species , such as Neurospora crassa and Aspergillus nidulans ( Figure S9 ) , suggesting that the mechanism we describe here may represent an ancient and conserved mode of Chk1 activation by its mediator . We failed to detect similar sequence motifs in budding yeast scRad9 , and a previous study had assigned the Chk1 activation function to the 40–200 amino acid region of scRad9 , which does not contain any SQ/TQ sites [13] . Thus , scRad9 may have evolved a different way of binding to and activating Chk1 , or alternatively , the ATR-like Mec1 kinase may phosphorylate the 40–200 amino acid region of scRad9 on non-SQ/TQ sites , as has been shown for the Mec1-mediated phosphorylation of Rad53 [47] . In metazoans , Claspin mediates the activation of Chk1 [14] , [48] . It has been suggested that Claspin is related by sequence homology to the replication checkpoint mediator Mrc1 in yeasts [8] , [9] . Thus , it is unlikely that Claspin and Crb2 share evolutionary ancestry . Despite this , our findings have revealed mechanistic similarities between the ways Claspin and Crb2 mediate Chk1 activation , namely , both Claspin and Crb2 undergo ATR/Rad3-dependent phosphorylation on multiple sites , and these phosphorylation events promote interactions with Chk1 kinase [28] , [30] . There is also a notable difference . The Chk1-binding region in Crb2 is phosphorylated on SQ/TQ motifs , probably by Rad3 , whereas the phosphorylation sites in the Chk1-binding region of Claspin are SG motifs directly phosphorylated by casein kinase 1 gamma 1 [31] . The Chk2 family effector kinases harbor one or two FHA domains , which are phosphopeptide-binding modules and can interact directly with their respective checkpoint mediators in a phosphorylation-dependent manner [10] , [49]–[52] . In contrast , Chk1 family kinases do not have any known phosphopeptide-binding domain . There are two conserved domains in Chk1 , the N-terminal kinase domain and the C-terminal regulatory domain . Vertebrate Chk1 appears to use its kinase domain to interact with phosphorylated Claspin [27] . However , in S . cerevisiae , conserved sequence motifs in the C-terminal domain of Chk1 were shown to be required for a yeast two-hybrid interaction between Chk1 and scRad9 [53] . We have attempted to use Crb2 peptide pull-down to identify the region of Chk1 involved in Crb2-Chk1 interaction . Neither the kinase domain nor the C-terminal domain is sufficient for binding with a phosphorylated Crb2 ( 67–85 ) peptide ( our unpublished observations ) , suggesting that both domains of Chk1 contribute to Crb2-Chk1 interaction . Previous studies have established the functional importance of phosphorylation-dependent interactions for the mediator-effector pairs of Claspin-Chk1 in vertebrates [28] , [30] , scRad9-Rad53 in budding yeast [51] , and Mrc1-Cds1 in fission yeast [10] , [52] . We show here that an SQ/TQ cluster-mediated Crb2-Chk1 interaction is critical for Chk1 activation in fission yeast . Thus , it appears that a common feature of the checkpoint signaling pathways is an essential direct interaction between checkpoint mediator and its downstream effector kinase . How such an interaction facilitates the activation of effector kinase is not entirely clear . As the ATR-mediated phosphorylation of aforementioned effector kinases is necessary for their activation , a current consensus of the field is that mediator-effector interactions increase the efficiency of ATR-catalyzed phosphorylation of effectors [10] , [29] , [47] . Two models , not exclusive to each other , can account for the impact of mediator-effector interactions on effector phosphorylation . The first model postulates that mediators directly participate in the phosphorylation reactions , either by increasing the enzyme-substrate affinity through simultaneously interacting with ATR and the effector , or by alternating the conformation of the effector to make it a better substrate for ATR kinase . Evidence supporting this model came from cell-free or reconstituted in vitro systems , showing that the presence of a mediator boosts the phosphorylation of its corresponding effector but not a generic ATR substrate [29] , [47] , [54] , [55] . As the spatial organization of cellular components was not maintained in these in vitro systems , the roles of spatial regulation could not be assessed . The other model suggests that a DNA damage-induced mediator-effector interaction alters the spatial distribution of an effector and brings it to DNA lesions , where ATR kinase also accumulates . As a consequence , the effector phosphorylation is enhanced due to heightened local concentrations of both enzyme and substrate . Consistent with this model is the fact that all mediators are capable of relocalizing to the proximity of aberrant DNA structures that trigger checkpoint signaling . For example , the DNA damage checkpoint mediators Crb2 and scRad9 form nuclear foci at DSBs [15] , [56] , and the replication checkpoint mediators Mrc1 and Claspin can bind to branched DNA structures in vitro , which may form at stalled replication forks in vivo [57] , [58] . In the case of Crb2 , the ability to relocalize to sites of DNA damage is essential for its checkpoint mediator function [21] . Our data here lend further support to the second model , as the fusion of Crb2 ( 67–85 ) peptide to either Rad22 or Rad4/Cut5 can largely bypass the need of the remaining portion of Crb2 for Chk1 activation . It is unlikely that the same peptide can simultaneously bind to Chk1 and Rad3 . Furthermore , as the SQ/TQ motifs , or even the 19-amino-acid Crb2 ( 67–85 ) peptide sequence , became dispensable when Crb2 and Chk1 were fused together ( Figure 4C and Figure S10 ) , the function of this peptide probably does not include altering Chk1 conformation or otherwise directly participating in the phosphorylation of Chk1 by Rad3 . Thus , the main checkpoint function of Crb2 appears to be recruiting Chk1 to DNA lesions where Rad3 also resides . In fission yeast , Rad3-mediated phosphorylation of the T412 residue on the Rad9 subunit of the 9-1-1 complex is essential for DNA damage-induced Chk1 activation [24] . This phosphorylation event promotes an interaction between Rad9 and the second pair of BRCT domains in Rad4/Cut5 . Consistent with these data , we show here that the formation of Rad4/Cut5 foci at DSBs requires Rad9 ( Figure S8 ) . Furthermore , the N-terminal tandem BRCT domains of Rad4/Cut5 mediate an interaction between Rad4/Cut5 and Crb2 [11] , [21] . Thus , a recruitment cascade composed of Rad9 , Rad4/Cut5 , and Crb2 can be envisioned , which eventually leads to the targeting of Crb2 to DSBs ( Figure 7 ) . A similar set of phosphorylation-dependent binary interactions between the orthologs of these proteins in budding yeast have been described , suggesting that such a checkpoint mediator recruitment pathway may be conserved [45] , [59] , [60] . Vertebrate and budding yeast homologs of Rad4/Cut5 , as well as the budding yeast homolog of Rad9 , possess in vitro ATR-activating activities . Such activities require sequence motifs containing a critical aromatic amino acid . Similar motifs have been found in fission yeast Rad4/Cut5 and Rad9 [43] , [44] . Whether Rad4/Cut5 and Rad9 are capable of activating Rad3 awaits verification by biochemical assays . However , we show here that Rad22-Crb2 ( 67–85 ) fusion can mediate checkpoint signaling in the absence of Rad9 . Thus , the Rad3 kinase activities towards Crb2 ( 67–85 ) and Chk1 do not absolutely need 9-1-1 , and the main function of 9-1-1 complex during DNA damage checkpoint signaling appears to be recruiting Crb2 to DSBs . This conclusion is consistent with recent reports showing that abolishing the ATR-activating activities of both ATR activators in budding yeast does not cause as strong defects as the loss of ATR ortholog Mec1 [44] , [45] . It is also consistent with studies showing that activation of Chk1 by Tel1 ( ATM ) , as revealed in strains lacking Ctp1 DSB resection factor , also requires the 9-1-1 complex [4] .
Strains used in this study are listed in Table S1 . Cells were maintained in logarithmic phase in EMM minimal media at 30°C . Microscopy was performed using a DeltaVision personalDV system equipped with a CFP/YFP/mCherry filter set ( Chroma 89006 set ) and a Photometrics CoolSNAP HQ2 camera . Images were acquired with a 100× , 1 . 4-NA objective . Four Z-sections at 0 . 5-µm intervals were merged into one image using the maximum intensity projection method with the softWoRx software . Whole cell extracts were prepared by boiling 10 OD600 units of cells with 100 µl SDS loading buffer following a 0 . 35 M NaOH treatment . To assess the mobility shift of Myc-tagged Chk1 , samples were run on 10% SDS-PAGE ( Bis-acrylamide∶acrylamide ratio of 1∶100 ) and immunoblotted with a polyclonal anti-Myc antibody ( Santa Cruz , sc-789 ) . To detect Crb2 mobility shift , samples were run on 6% SDS-PAGE and immunoblotted with a polyclonal anti-Crb2 antibody ( Du et al . , 2003 ) . Proteins were extracted from about 1000 OD600 units of cells by glass bead beating ( FastPrep-24 ) in lysis buffer ( 50 mM Hepes , pH 7 . 5 , 100 mM NaCl , 1 mM EDTA , 10% glycerol , 0 . 05% NP-40 , 1 mM PMSF , 1 . 5 mM DTT , 1× Protease Inhibitor Cocktail ( Roche ) , 1× PhosSTOP ( Roche ) ) . Anti-Flag M2 affinity gel ( Sigma , A2220 ) was applied to immunoprecipitate YFH-tagged Chk1 . After binding , beads were briefly washed with lysis buffer and eluted with 3×Flag peptide . Eluted Chk1 was incubated with 2 µg biotin-labeled Crb2 ( 67–85 ) peptides at 4°C for 2 hours . Then 30 µl pre-washed Dynabeads M-280 Streptavidin ( Invitrogen ) was added to pull down peptides . Beads were briefly washed with lysis buffer and eluted with SDS loading buffer . Chk1 was detected by Coomassie staining or immunoblotting with an anti-GFP antibody ( Roche #11814460001 ) . Protein extraction and Flag-IP were performed as above except eluting from anti-Flag M2 affinity gel by boiling with SDS loading buffer . Immunoblotting was performed with a polyclonal anti-Flag antibody ( Sigma , F7425 ) and an anti-pSQ/TQ antibody ( Cell Signaling #2851 ) . TAP-tagged Crb2 was purified from IR-treated cells using IgG Sepharose beads and eluted by TEV protease cleavage . The eluate was dissolved in 8 M urea , 100 mM Tris , pH 8 . 5 , reduced with 5 mM TCEP for 20 min , and alkylated with 10 mM iodoacetamide for 15 min in the dark , all at the room temperature . Then the sample was split into three aliquots , digested separately overnight at 37°C with trypsin ( in 2 M urea , 1 mM CaCl2 , 100 mM Tris , pH 8 . 5 ) , elastase ( in 2 M urea , 100 mM Tris , pH 8 . 5 ) , or subtilysin ( in 6 M urea , 100 mM Tris , pH 8 . 5 ) . The digestions were stopped with 5% formic acid ( final concentration ) . Peptides from three digestions were combined and loaded onto a desalting column ( 250-µm i . d . fused silica capillary column with 2 cm Aqua C18 resin ( Phenomenex ) with a 2-µm filtered union ) . After desalting , a 100-µm i . d . column packed with 10 cm of Aqua C18 resin and 2 cm of Partisphere SCX resin ( Whatman ) was connected to the desalting column through the filtered union . MS analysis was performed on LCQ Deca mass spectrometer ( Thermo-Finnigan ) using a 12-step MudPIT method described previously [33] . The MS/MS spectra were searched with SEQUEST [61] with or without an addition of 80 on S , T , or Y ( phosphorylation ) against an S . pombe protein database . The search results were combined and filtered with DTASelect [62] . | To preserve the integrity of genomic DNA , eukaryotic cells use a genome surveillance system to detect DNA damage and send a signal to prevent cell division before DNA repair has been completed . This signal transduction mechanism is called checkpoint signaling and is conserved from yeasts to humans . A key checkpoint component is the protein kinase Chk1 , which is activated in response to DNA damage . In fission yeast , the activation of Chk1 requires a number of upstream checkpoint proteins including Crb2 , but their exact roles and mechanisms of action are unclear . Here we show that Crb2 interacts with Chk1 when Crb2 is phosphorylated at two conserved N-terminal positions . This phosphorylation-dependent interaction is critical for Chk1 relocalization to damaged DNA and Chk1 activation . By tethering a Crb2 peptide sufficient for Chk1 binding to DNA lesions , we demonstrated that the main role of Crb2 and 9-1-1 complex in checkpoint signaling is to bring Chk1 to sites of DNA damage , where it can be activated . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"signal",
"transduction",
"signaling",
"in",
"cellular",
"processes",
"biology",
"molecular",
"cell",
"biology",
"molecular",
"biology"
]
| 2012 | Phosphorylation-Dependent Interactions between Crb2 and Chk1 Are Essential for DNA Damage Checkpoint |
Genomic disorders are often caused by recurrent copy number variations ( CNVs ) , with nonallelic homologous recombination ( NAHR ) as the underlying mechanism . Recently , several microhomology-mediated repair mechanisms—such as microhomology-mediated end-joining ( MMEJ ) , fork stalling and template switching ( FoSTeS ) , microhomology-mediated break-induced replication ( MMBIR ) , serial replication slippage ( SRS ) , and break-induced SRS ( BISRS ) —were described in the etiology of non-recurrent CNVs in human disease . In addition , their formation may be stimulated by genomic architectural features . It is , however , largely unexplored to what extent these mechanisms contribute to rare , locus-specific pathogenic CNVs . Here , fine-mapping of 42 microdeletions of the FOXL2 locus , encompassing FOXL2 ( 32 ) or its regulatory domain ( 10 ) , serves as a model for rare , locus-specific CNVs implicated in genetic disease . These deletions lead to blepharophimosis syndrome ( BPES ) , a developmental condition affecting the eyelids and the ovary . For breakpoint mapping we used targeted array-based comparative genomic hybridization ( aCGH ) , quantitative PCR ( qPCR ) , long-range PCR , and Sanger sequencing of the junction products . Microhomology , ranging from 1 bp to 66 bp , was found in 91 . 7% of 24 characterized breakpoint junctions , being significantly enriched in comparison with a random control sample . Our results show that microhomology-mediated repair mechanisms underlie at least 50% of these microdeletions . Moreover , genomic architectural features , like sequence motifs , non-B DNA conformations , and repetitive elements , were found in all breakpoint regions . In conclusion , the majority of these microdeletions result from microhomology-mediated mechanisms like MMEJ , FoSTeS , MMBIR , SRS , or BISRS . Moreover , we hypothesize that the genomic architecture might drive their formation by increasing the susceptibility for DNA breakage or promote replication fork stalling . Finally , our locus-centered study , elucidating the etiology of a large set of rare microdeletions involved in a monogenic disorder , can serve as a model for other clustered , non-recurrent microdeletions in genetic disease .
Copy number variations ( CNVs ) are defined as DNA segments that are present at a variable copy number in comparison with a reference genome such as a deletions , duplications or insertions [1] , [2] . In recent years it has become clear that CNVs are a major source of genetic diversity , competing with the single nucleotide variants ( SNVs ) as the main source of genetic variation between individuals . With the use of several technologies such as array-based comparative genomic hybridization ( aCGH ) , single nucleotide polymorphism ( SNP ) genotyping and next-generation sequencing , numerous CNVs have been identified during the last decade [3]–[11] . Many of the identified CNVs represent benign polymorphic variants; however , CNVs can lead to a genetic disease when for instance a dosage-sensitive gene is affected . Such genetic diseases caused by genomic rearrangements are defined as genomic disorders [12]–[15] . The genomic rearrangements causing these disorders can be recurrent sharing a common interval and size , and having clustered breakpoints in multiple different subjects . These rearrangements are mostly the result of nonallelic homologous recombination ( NAHR ) between low-copy repeats ( LCRs ) or segmental duplications ( SDs ) , a recombination-based mechanism [16] . In contrast , non-recurrent , locus-specific rearrangements can vary in size and have scattered breakpoints , thus suggesting the absence of a recombination hotspot . Only recently , several mechanisms causing non-recurrent genomic rearrangements have been proposed such as ( i ) non-replicative repair mechanisms: non-homologous end-joining ( NHEJ ) [17] , microhomology mediated end-joining ( MMEJ ) [18] and NAHR between repetitive elements ( for example , Alu or L1 ) [19] , [20]; and ( ii ) replicative-based repair mechanisms: fork stalling and template switching ( FoSTeS ) [21] , microhomology-mediated break-induced replication ( MMBIR ) [22] , serial replication slippage ( SRS ) [23] and break-induced SRS ( BISRS ) [24] . Interestingly , as genomic rearrangements are assumed not to be random events , it has been proposed that the local genomic architecture other than LCRs or SDs stimulates these mechanisms by predisposing to CNV formation [25] . Indeed , several studies have revealed repetitive elements , sequence motifs or non-B DNA conformations overlapping with or located in the vicinity of CNV breakpoints . Another genomic characteristic frequently observed at the junctions is microhomology . These studies confirm that the majority of non-recurrent , locus-specific , pathogenic CNVs are not caused by NAHR , but rather by a diverse range of mechanisms [26]–[35] . The conclusions of these studies are however mostly based on a small number of sequenced junctions . Therefore , it was our aim to investigate which mechanisms underlie a large , unique set of locus-specific non-recurrent genomic rearrangements causing the rare developmental disorder blepharophimosis-ptosis-epicanthus inversus syndrome ( BPES ) [MIM #110100] . This disorder is characterized by a complex eyelid malformation with or without ovarian dysfunction [36] , [37] . BPES is an autosomal dominant disorder caused by genetic defects of the FOXL2 locus [38]–[44] . Even though intragenic mutations are most prevalent ( 81% ) , an important fraction of BPES cases is caused by heterozygous deletions . These deletions can encompass the FOXL2 gene ( 12% ) or can be located outside the FOXL2 transcription unit removing potential regulatory elements such as conserved non-coding sequences ( CNCs ) and the long non-coding RNA ( lncRNA ) PISRT1 , necessary for the correct transcription of FOXL2 ( 5% ) [41]–[44] . Here , we study 32 FOXL2 encompassing and 10 regulatory deletions , respectively . As the observed deletions range from 1 . 4 kb to 5 . 51 Mb and the breakpoint locations are heterogeneous , a common deletion mechanism such as NAHR mediated by LCRs can be excluded . In order to unravel the underlying deletion mechanisms , we analyzed the extent of microhomology at the characterized breakpoints and explored the presence of repetitive elements , non-B DNA conformations and sequence motifs as well . We found that microhomology was present in 91 . 7% of 24 delineated breakpoint junctions . Moreover , particular genomic architectural features were found in all breakpoint regions . In conclusion , we propose that the majority of these deletions are caused by microhomology-mediated mechanisms such as MMEJ or the replicative-based repair mechanisms FoSTeS , MMBIR , SRS and BISRS . Finally , the genomic architecture might stimulate the formation of these rare deletions by increasing the susceptibility for DNA breakage or promote replication fork stalling .
Two of the 42 deletions were already delineated at base-pair resolution in previous studies [42] , [43] . For the delineation of the remaining 40 deletions a strategy was followed as described in Figure 1 . In short , a combination of aCGH , qPCR , long-range PCR and Sanger sequencing was applied . Based on the aCGH and qPCR analyses , long-range PCR was performed for 35 deletions of which 22 resulted in a specific junction product . The inability to obtain a product for the remaining 13 deletions may relate to the complexity of the genomic sequence at these junctions . To overcome this , several primer combinations were used however this was without success . The 22 specific junction products underwent Sanger sequencing to determine the exact physical location of the breakpoints . The FOXL2 encompassing deletions ranged from 1 . 4 kb to 5 . 51 Mb while the regulatory deletions ranged from 7 . 4 kb to 3 . 02 Mb , including one complex deletion consisting of two deletions interspersed with a segment without copy number variation ( namely deletion F , Figure S1 ) . Overall , we were able to characterize the exact breakpoints of 16 FOXL2 encompassing ( 1–16 ) and 8 regulatory deletions ( A–H ) using this strategy ( Figure 2 ) . The breakpoints of the locus-specific , non-recurrent deletions were subjected to an extensive bioinformatic analysis to explore underlying mechanisms and to assess the contribution of the genomic architecture . To this end , we analyzed the extent of microhomology at the breakpoints and investigated the presence of repetitive elements , sequence motifs and non-B DNA conformations . An overview of the output of the different bioinformatic analyses can be found in Table 1 . Visual representations of the breakpoint regions with the observed local genomic architecture of 5 selected deletions are shown in Figure 3 and of the remaining deletions in Figure S2 . Microhomology is defined as one or more base pairs ( bp ) of perfectly matching sequence shared between the proximal and distal reference sequences surrounding the breakpoints . Also , it is an important hallmark of several mechanisms [14] . The extent of microhomology was evaluated using multiple sequence alignments ( Figure 4 , Figure S3 ) . Of the 24 deletion junctions analyzed , 22 ( 91 . 7% ) displayed microhomology between their breakpoints , ranging from 1 bp up to 66 bp . Only two deletions ( deletion A and 6 ) showed a perfect transition at their junction of which one ( deletion 6 ) was accompanied by a deletion of one bp . To exclude whether the observed microhomology at the breakpoints did just occur by chance , we compared our results against a random control population of 500 human genomic sequences representing artificial breakpoint regions . Using a Fisher's exact test we observed that microhomology is significantly enriched ( p = 2 . 28×10−08 ) at our studied breakpoints . In addition , using a Wilcoxon rank sum test we observed that the distribution of microhomology in our breakpoints significantly differed ( p = 2 . 21×10−12 ) from the random control population ( Figure 5 ) . This distribution pattern is in accordance with the ones observed by previous studies [45] , [46] . The Repeat Masker track in the UCSC genome browser was used to analyze the presence of known repetitive elements intersecting the breakpoints . A repetitive element was found at 31 of 48 breakpoints ( 64 . 6% ) ( Table 1 ) . In the random control population a repetitive element was observed to intersect with 236 of 500 breakpoints ( 47 . 2% ) . Using a Fisher's exact test , we could conclude that our breakpoints are indeed significantly enriched with repetitive elements ( p = 2 . 4×10−2 ) . Interestingly , Alu elements were observed about three times more at our breakpoints in comparison with the control population ( 29 . 2% versus 10 . 6% ) . Indeed , when performing a Fisher's exact test with Bonferroni correction , we observed a significant enrichment of Alu elements at our breakpoints ( p = 0 . 001 ) . The frequency of L1-elements does not significantly differ from the control population ( 25% versus 16 . 2%; p = 0 . 156 ) . In 13 of 24 deletions ( 54 . 2% ) , a repetitive element was observed at both breakpoints . Of these , 9 had repetitive elements belonging to the same class consisting of 6 Alu-Alu and 3 L1PA-L1PA combinations . In these cases , a Blast2 analysis was performed to determine the percentage of sequence identity between the repetitive elements . The highest percentage of sequence identity was observed between two L1PA3 elements in deletion 16 ( 96% ) . The lowest percentage of sequence identity was observed between an AluSx3 and an AluSz6 in deletion 14 ( 77% ) . The percentages for the other 7 deletions can be found in Table 1 . The well-known capacity of sequence motifs to predispose to DNA breakage led us to analyze the nucleotide context of the breakpoint regions for the presence of 40 known sequence motifs [47] . An overview of the results can be found in Table S2 . This analysis was also performed for the random control population . In total , 26 of 40 sequence motifs were present in one or more breakpoint regions . Only the proximal breakpoint region of deletion 4 did not contain a sequence motif . In comparison with the random control population , we observed that none of the motifs was significantly overrepresented in our breakpoint regions . In addition to individual motifs , we also analyzed if the overall density of sequence motifs might be increased . For this purpose , we counted the number of motifs present in each breakpoint region for the studied deletions and the random control population . In our deletions we observed a mean of 9 . 69 motifs per breakpoint region while a mean of 7 . 86 was observed for the random control population . However , the overall density of sequence motifs does not differ significantly ( Wilcoxon rank sum test , p = 0 . 207 ) . No new sequence motifs could be found in our deletion cohort . Different bioinformatic tools were applied to determine the presence of sequences capable of forming non-B DNA conformations . Of note , genomic architecture resulting from DNA conformational changes , but not the primary sequence information , is crucial in these processes [48] . In total , a sequence capable of forming a non-B DNA structure could be identified in 14 of the 48 breakpoints ( 29 . 2% ) . Such sequences were identified in 107 of the 500 ( 21 . 4% ) breakpoint regions of the random control population indicating that the frequency of sequences capable of forming a non-B DNA structure does not differ significantly between both populations ( Fisher's exact test , p = 0 . 208 ) . The comparison with the random control population was made for the individual non-B DNA conformations as well . The frequency of slipped hairpin structures and left-handed Z-DNA does not differ significantly from those observed in the control population ( Fisher's exact test , p>0 . 05 ) . However , for the tetraplex structures a significant overrepresentation could be observed ( Fisher's exact test , p = 0 . 006 ) . Notably , four deletions have sequences capable of forming non-B DNA conformations present in both breakpoint regions ( Table S3 ) . Even more remarkable is that the non-B DNA conformations are from the same class in these deletions . Deletion 14 has a direct repeat in both breakpoint regions , while an oligo ( G ) n tract is observed in both breakpoint regions of deletions 1 , 2 and 7 respectively . Interestingly , of the 14 breakpoint regions harboring a sequence capable of forming non-B DNA conformations , only 1 breakpoint region belonged to a regulatory deletion ( deletion H ) . This means that such sequences are significantly overrepresented in the breakpoint regions of the FOXL2 encompassing deletions ( Fisher's exact test , p = 0 . 018 ) .
Non-recurrent CNVs can be caused by a large spectrum of different mechanisms which can be grossly classified as non-replicative - ( NAHR , NHEJ and MMEJ ) or replicative-based repair mechanisms ( FoSTeS , SRS , BISRS and MMBIR ) . If successful , the only reminder of a rearrangement is a unique breakpoint signature which can be used as the key to unraveling the underlying mechanism . NAHR causes rearrangements by misalignment and subsequent unequal cross-over between nonallelic sequences in meiosis or mitosis . For NAHR to occur , segments of a minimal length sharing extremely high similarity or sequence identity - named minimal efficient processing segments ( MEPS ) - between the homologous recombination substrates are required . These are mostly LCRs but can also be L1s , Alu elements or pseudogenes [49] . Breakpoints of rearrangements inferred by NAHR should therefore be intersected by these elements . NHEJ is utilized by human cells to repair two-ended , double stranded DNA breaks . NHEJ is characterized by two main features . First , NHEJ does not require the presence of substrates with extended homology but can be facilitated by the presence of microhomology ( 1–4 bp ) . Second , NHEJ can leave an ‘information scar’ at the joint point comprising of the loss or insertion of several random nucleotides [17] . An alternative pathway of NHEJ is called MMEJ . The difference between these two is that while the presence of microhomology is optional in NHEJ , it is a requirement for MMEJ to occur . Also , MMEJ uses longer stretches of microhomology ( 5–25 bp ) than those used in NHEJ [50] . Two similar models , FoSTeS and SRS , were proposed to explain the sequence complexity sometimes seen at breakpoints . According to these models , the DNA replication fork can stall; the lagging strand consequently disengages from the original template , switches to another replication fork and then restarts DNA synthesis on the new fork by priming it via the microhomology between the switched template site and the original fork . Switching to a downstream replication fork would therefore result in a deletion , while upstream switching results in a duplication [21] , [23] . Although both models share the same hypothesis of fork template switching , a difference can be observed . While the SRS model assumes that replication slippage occurs on closely adjacent sites and causes DNA rearrangements of small sizes , the FoSTeS model emphasizes that the template switch can occur over long distances ( even 100 kb or megabase size ) and therefore cause DNA rearrangements on a much larger scale [49] . Further molecular details of FoSTeS and SRS were extended in two more generalized models , namely MMBIR and BISRS . The major feature distinguishing these generalized models is that they are initiated by a single-end , double strand DNA break generated by a collapsed fork to expose a 3′ end that can be used to prime synthesis at a distant fork [22] , [24] . All of these replicative-based repair mechanisms do not only cause complex rearrangements but can also form simple rearrangements where the evidence for sequence complexity has been removed during the rearrangement process . In addition , these mechanisms may be stimulated by the local genomic architecture . Consequently , the only option to elucidate the mechanism behind a CNV , is to delineate it at base-pair resolution and examine the sequence context of the breakpoints . Of our deletions of the FOXL2 locus , 24 could be delineated at the base-pair level . Using several bioinformatics tools , we could examine the sequence context of these deletions , define their breakpoint signature and deduce the most likely underlying mechanism . Remarkably , no major differences were observed between the mechanisms underlying FOXL2 encompassing and regulatory deletions . Based on the observed breakpoint signatures , the deletions could be classified in three different groups . The first small group contains only two deletions ( deletion A and 6 ) both of which have a perfect transition at the junction . Additionally , the loss of a T nucleotide at the junction of deletion 6 represents an information scar pointing to NHEJ as potential mechanism . The 9 deletions of the second group are characterized by the presence of repetitive elements of the same family at both breakpoints ( deletion G , H , 7 , 8 , 9 , 11 , 13 , 14 and 16 ) which could indicate that NAHR has caused these deletions like observed in other studies [28] , [30] , [32] , [34] , [35] . An Alu-Alu-mediated NAHR might have resulted in 6 deletions while the other three deletions probably result from a L1-L1-mediated NAHR . However , the level of sequence identity is probably too low in most deletions for NAHR to occur . Three deletions do have a high percentage of sequence identity over a long length between L1 elements ( Table 1 ) . These L1 elements could therefore provide the MEPS required for efficient NAHR . On the other hand , microhomology ranging from 5 bp to 66 bp is observed at the junctions of these 9 deletions , suggesting that a replicative-based repair mechanism may have formed these deletions instead of NAHR [51] . It has also been suggested that repetitive elements may represent more difficult sequences to replicate leading to an increased chance of replication fork stalling or collapsing [46] . Alternatively , formation of secondary structures within or between repetitive elements may contribute to generate DSBs and further contribute to genomic instability involving those elements . So the presence of a repetitive element may initiate a replicative-based repair mechanism while the observed microhomology then facilitates the template switching and serves as the priming site in the second replication fork . The above assumptions are purely speculative though and further experimental evidence is needed to substantiate them . Another possible mechanism underlying these deletions is MMEJ which requires microhomology of 5 bp or more . It is however currently impossible to distinguish between replicative-based repair mechanisms and MMEJ , as they all share the breakpoint signature , namely microhomology at the junction . Nonetheless , because none of the junctions have an information scar , replicative-based repair mechanisms are favored . The 13 deletions of the third group ( deletion B , C , D , E , F , 1 , 2 , 3 , 4 , 5 , 10 , 12 and 15 ) also have microhomology at their junction but as opposed to the second group they only have a repetitive element at one of their breakpoints or in case both breakpoints intersect with a repetitive element , the elements are from a different family . The microhomology in this third group ranges from 1 bp to 5 bp . Like the deletions of the second group , these 13 deletions also could have resulted from NHEJ , MMEJ or replicative-based repair mechanisms but again favoring the latter because no information scar was present at the junctions . Nonetheless , NHEJ or MMEJ could still have occurred , where a distinction can be made between both based on the length of microhomology . Microhomology of 1–4 bp may facilitate NHEJ ( deletions C , D , E , F , 2 , 3 , 4 , 5 , 10 , 12 and 15 ) [17] while longer microhomology stretches of 5 bp or more are used by MMEJ ( deletions B and 1 ) [50] . Interestingly , a substitution of one and two nucleotides was observed near the junctions of deletion 3 and E respectively . None of these substitutions are described as a known SNP and they originate most likely as a side-effect of the underlying mechanism . The majority of these mechanisms are based on the occurrence of DSBs and the subsequent repair of these breaks for the formation of genomic rearrangements . It has been described that the repair polymerases recruited for these processes , are more prone to errors and thus may incorporate wrong bases during DNA synthesis [52] , [53] . These mutations are referred to as break-repair-induced mutations [54] . In conclusion , in this set of junctions of non-recurrent , locus-specific deletions involving the FOXL2 locus , we propose that the majority of deletions are caused by the microhomology-mediated mechanisms MMEJ , FoSTeS , MMBIR , SRS or BISRS . This conclusion is in accordance with the observations of the most recent similar locus-specific studies [31]–[35] . Moreover , microhomology is observed at the majority of sequenced junctions in both locus-specific and genome-wide benign or pathological CNVs supporting the role of replicative-based repair mechanisms in CNV formation [55] . Less recent studies conversely suggest NHEJ to be the major mechanism in causing non-recurrent deletions . These studies were however performed before replicative-based repair mechanisms were reported [26]–[30] . Interestingly , when revisiting the data of these studies , microhomology is observed at more than half of these junctions indicating that replicative-based repair mechanisms could potentially also occur ( Table S4 ) . Furthermore , based on our results we hypothesize that other unique , non-recurrent , clustered microdeletion cohorts [56]–[60] are potentially also caused by a variety of microhomology-mediated mechanisms such as MMEJ , FoSTeS , MMBIR , SRS and BISRS . The role of genomic architectural features in the formation of recurrent CNVs is well established as flanking LCRs or SDs act as homologous recombination substrates for an NAHR or ectopic recombination event mediated by these homologous sequence substrates . However , the role of genome architecture in non-recurrent rearrangements is currently still unclear . Studies like ours therefore contribute to the elucidation of a potential role of the genomic architecture and help delineate what those potential features may be . The presence of repetitive elements , sequences forming non-B DNA conformations and sequence motifs may lead to genomic instability and subsequently genomic rearrangements by promoting the formation of DSBs or by stalling the replication [48] , [61]–[64] . Such genomic architectural features were observed in all breakpoint regions but only repetitive elements within particular Alu elements were found to be significantly enriched . To investigate whether this enrichment was not a bias , we compared the fraction of Alu elements in the CNV region with that in chromosome 3 and in the entire genome . Indeed , the fraction of sequence length occupied by Alu elements in the region containing the deletions ( chr3:129230494–148645311 , hg19 ) is only 8 . 32% which is comparable to the fraction found for chromosome 3 ( 8 . 84% ) and the human genome 10 . 6% [65] . Overall , this indicates that Alu elements do occur more frequently at the breakpoints compared to the genome average . Although this observation is in accordance with a similar study by Vissers et al . [46] , the mechanistic significance of this is currently unknown . Oligo ( G ) n tracts capable of forming tetraplex structures also displayed a significant overrepresentation in the breakpoint regions . Interestingly , both breakpoint regions of deletions 1 , 2 and 7 display an oligo ( G ) n tract while deletion 14 has direct repeats in both breakpoint regions which could indicate that 2 DSBs have occurred in these deletions , favoring NHEJ or MMEJ . Conversely , the presence of the non-B DNA conformations in these and the other deletions can cause collapsing of the replication fork . Replicative-based repair mechanisms can therefore not be ruled out . Interestingly , sequences capable of forming non-B DNA conformations were observed more frequently in the breakpoints of the FOXL2 encompassing deletions than in those of the regulatory deletions suggesting that the genomic architecture differs between both types of deletions . This might explain the higher prevalence of deletions encompassing FOXL2 . We propose that the majority of non-recurrent deletions of the FOXL2 locus are caused by microhomology-mediated mechanisms like MMEJ , FoSTeS , MMBIR , SRS or BISRS . Finally , the genomic architecture might drive the formation of these rare , locus-specific deletions by increasing the susceptibility for DNA breakage or promote DNA replication fork stalling . The insights from our locus-centered study investigating a large set of breakpoint sequences from non-recurrent , gene encompassing and regulatory microdeletions causing monogenic disease , can therefore serve as a paradigm for other clustered , non-recurrent microdeletions involved in genetic disease .
This study was conducted following the tenets of Helsinki and approved by the institutional review board ( 99/250 ) . Forty-two consenting BPES patients with a FOXL2 encompassing ( 32 ) or regulatory deletion ( 10 ) were enrolled in this study . All patients were clinically diagnosed with BPES based on the presence of minimal three out of the four typical BPES features . Patients can be subdivided based on the genetic center where they were molecularly diagnosed . The largest group of deletions was diagnosed at the Center for Medical Genetics at Ghent University ( CMGG ) in Belgium . This group contains 25 FOXL2 encompassing deletions and 10 regulatory deletions . The second group of 7 FOX2 encompassing deletions was diagnosed at the Instituto de Genética Médica y Molecular ( INGEMM ) at the Hospital Universitario La Paz in Spain . Molecular diagnosis of all FOXL2 encompassing deletions was performed using a commercially available multiplex ligation-dependent probe amplification ( MLPA ) mix ( P054 , MRC-Holland , Amsterdam , the Netherlands ) according to the manufacturer's instructions . The regulatory deletions located outside the FOXL2 transcription unit were identified using a combined approach of microsatellite analysis and a custom-made quantitative PCR assay in the FOXL2 region ( qPCR-3q23 ) as previously described [42] , [43] . Two different array-based methods were used: ( i ) custom high-resolution 8×60 K Agilent microarrays at the CMGG , and ( ii ) genome-wide Illumina Human610-Quad BeadChip arrays at the INGEMM . The custom high-resolution 8×60 K Agilent microarray was designed using the online design tool eArray ( Agilent Technologies ) , targeting a region of 10 Mb around FOXL2 ( chr3:133517310–143517310; UCSC , Human Genome Browser , hg19 ) consisting of 52 , 800 probes spaced at an average density of 200 bp . Hybridizations were performed according to manufacturer's instructions with minor modifications [66] . The results were subsequently visualized in arrayCGHbase [67] . The genome-wide Illumina Human610-Quad BeadChip arrays contain 620 , 901 tag SNPs and have an average resolution of 4 . 7 kb . Hybridization and subsequent data-analysis was performed as previously described [44] . The proximal and distal breakpoint regions were defined as the regions between the last proximal normal and first deleted probe proximally , and the last deleted and first distal normal probe , respectively . If the sum of the breakpoint regions outsized the predefined , arbitrary threshold of 15 kb , qPCR was used to reduce the breakpoint regions , resulting in more suitable fragments for long-range PCR . Primers were designed equally throughout the breakpoint regions and subjected to a stringent in silico and in vitro validation according to previously described parameters . The qPCR primers that qualified were used in a qPCR-based copy number analysis as previously described [68] . In short , 7 . 5 µl qPCR reactions contained 3 . 75 µl 2× master mix ( qPCR core kit for SYBR Green I , Eurogentec ) , 0 . 375 µl of each primer ( 5 µM working solution ) , 1 µl nuclease-free water and 2 µl template ( 10 ng/µl ) . The reactions were carried out on the LightCycler 480 Instrument II ( Roche ) using the following qPCR protocol: 10 min pre-incubation at 95°C followed by 45 cycles of 95°C for 10 s , 60°C for 45 s and 72°C for 1 s , next a dissociation run from 60 to 95°C and ending with a cooling step . Data-analysis was performed with qBasePlus software [69] . Two reference genes were used for normalization of the relative quantities and two positives controls with known copy number were used as a reference to calculate the copy numbers [68] . For the delineation of the deletions at nucleotide level , specific junction products need to be obtained . Therefore , inward-facing PCR primers were designed in the normal regions flanking the breakpoint regions . Long-range PCR reactions were performed in a total volume of 20 µl containing 1× iProof HF buffer , 200 µM of each dNTP , 0 . 5 µM of each primer , 0 . 4 units of iProof DNA-polymerase ( Bio-Rad ) and 100 ng of template DNA . The standard PCR protocol is defined as follows: 94°C for 2 min , 35 cycles of ( 94°C for 30 sec , Ta for 30 sec , 68°C for 1 min/kb ) , and a final extension of 72°C for 10 min with an optimized annealing temperature and extension time for each junction product . To evaluate the specificity of a junction product , a control sample of a healthy individual accompanied the deletion samples . After amplification , the PCR products were visualized using the LabChip GX with the DNA 5K assay kit ( Caliper Life Sciences ) if junction products are assumed to be smaller than 5 kb or using gel electrophoresis . Next , specific junction products were sequenced using internal primers with the BigDye Terminator v . 3 . 1 Cycle Sequencing Kit ( Applied Biosystems ) . Sequencing reactions were then loaded on an Applied Biosystems Prism 3130 or 3730 genetic Analyzer . The sequences generated from the internal primers were first aligned to the reference sequence ( obtained from UCSC , hg19 ) with SeqScape v1 . 1 ( Applied Biosystems ) to visualize the junction . To determine the exact genomic location of the breakpoints , the proximal and distal sequences flanking the junction were loaded into the Blat tool provided by the UCSC browser [70] . If microhomology was present at the junction , the genomic location of the proximal breakpoint was defined as the last nucleotide adjacent to the microhomology-stretch and the genomic location of the distal breakpoint was defined as the first nucleotide adjacent to the microhomology-stretch . Breakpoints , breakpoint regions and junction fragments were subjected to an extensive bioinformatic analysis , with breakpoint region defined as a 150 bp fragment surrounding a breakpoint and junction fragment as a 150 bp fragment surrounding the junction , to assess the involvement of the genomic architecture in the origin of the deletions . First , the presence of microhomology at the breakpoints was analyzed with a multiple sequence alignment between the proximal and distal breakpoint regions , and the junction fragment using ClustalW [71] . Second , the presence of known repetitive elements intersecting the breakpoints was investigated using the Repeat Masker track in the UCSC genome browser [72] . In cases where both breakpoints of a deletion overlap with a repetitive element , BLAST2 was used to determine the percentage of sequence identity between the elements [73] . Third , the presence of DNA sequences leading to non-B DNA conformations in the breakpoint regions was examined with several different tools: GT-repeats ( forming left-handed Z-DNA ) with Zhunt online [74]; direct , inverted and mirror repeats capable ( forming slipped hairpin , cruciform and triplex structures , respectively ) with RepeatAround [75]; oligo ( G ) n tracts ( forming tetraplex structures ) with QGRS [76] . Non-B DNA conformations were only included if both counterparts flanked the breakpoint . And fourth , the presence of previously described sequence motifs [47] was analyzed with Fuzznuc [77] . These results were compared against a random control population representing the human genome as described by Vissers et al . [46] and Hannes et al . [78] , to assess the statistical significance of the presence of genomic architecture . This random control population consists of 500 human genomic sequences of 150 bp each , randomly extracted from Ensembl using an in-house developed script . These sequences represent artificial breakpoint regions with the breakpoint between nucleotides 75 and 76 . The same bioinformatic analyses were performed on these 500 sequences . The nucleotides surrounding the artificial breakpoint were evaluated for the presence of microhomology and the artificial breakpoints were analyzed for the possible presence of intersecting repetitive elements . Finally , the entire breakpoint regions were evaluated for the presence of motifs or sequences capable of forming non-B DNA conformations . Fisher's exact tests were performed to verify if the presence of a genomic element in the deletion population differed significantly in comparison with the control population . | Genomic disorder is a general term describing conditions caused by genomic aberrations leading to a copy number change of one or more genes . Copy number changes with the same length and clustered breakpoints for a group of patients with the same disorder are named recurrent rearrangements . These originate mostly from a well-studied mechanism , namely nonallelic homologous recombination ( NAHR ) . In contrast , non-recurrent rearrangements vary in size , have scattered breakpoints , and can originate from several different mechanisms that are not fully understood . Here we tried to gain further insight into the extent to which these mechanisms contribute to non-recurrent rearrangements and into the possible role of the surrounding genomic architecture . To this end , we investigated a unique group of patients with non-recurrent deletions of the FOXL2 region causing blepharophimosis syndrome . We observed that the majority of these deletions can result from several mechanisms mediated by microhomology . Furthermore , our data suggest that rare pathogenic microdeletions do not occur at random genome sequences , but are possibly guided by the surrounding genomic architecture . Finally , our study , elucidating the etiology of a unique cohort of locus-specific microdeletions implicated in genetic disease , can serve as a model for the formation of genomic aberrations in other genetic disorders . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"genetics",
"molecular",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"and",
"genomics"
]
| 2013 | Microhomology-Mediated Mechanisms Underlie Non-Recurrent Disease-Causing Microdeletions of the FOXL2 Gene or Its Regulatory Domain |
Plexins are cell surface receptors for the semaphorin family of cell guidance cues . The cytoplasmic region comprises a Ras GTPase-activating protein ( GAP ) domain and a RhoGTPase binding domain . Concomitant binding of extracellular semaphorin and intracellular RhoGTPase triggers GAP activity and signal transduction . The mechanism of this intricate regulation remains elusive . We present two crystal structures of the human Plexin-B1 cytoplasmic region in complex with a constitutively active RhoGTPase , Rac1 . The structure of truncated Plexin-B1-Rac1 complex provides no mechanism for coupling RhoGTPase and Ras binding sites . On inclusion of the juxtamembrane helix , a trimeric structure of Plexin-B1-Rac1 complexes is stabilised by a second , novel , RhoGTPase binding site adjacent to the Ras site . Site-directed mutagenesis combined with cellular and biophysical assays demonstrate that this new binding site is essential for signalling . Our findings are consistent with a model in which extracellular and intracellular plexin clustering events combine into a single signalling output .
Plexins constitute a large family of semaphorin receptors that mediate the repulsive chemotactic response necessary for axon guidance in the developing nervous system . They also play an important role in regulating cell migration , angiogenesis , and immune responses [1] , [2] . Mutations in plexin receptors have been found in cancers from a variety of tissues [3] , [4] . There are four classes of Plexins ( A , B , C , and D ) [1] . Their architecture is conserved across the family with a large extracellular region including the ligand binding sema domain , a single transmembrane spanning helix , and an intracellular region that transduces signals to a number of downstream pathways [1] , [2] , [5] . Recently , truncated ectodomain structures of plexins from different classes in complex with their cognate semaphorin ligands have been solved [6]–[8] . They revealed a common architecture in which two plexin monomers bind one semaphorin dimer . This bivalency has been shown to be crucial for the function of the plexin-semaphorin complex [6] . Plexins are transmembrane receptors distinguished by their ability to interact directly with small GTPases of the Ras and Rho family through their intracellular region [9] , [10] . They consist of two domains , the GTPase activating protein ( GAP ) domain , first identified by sequence similarity to RasGAP , and the RhoGTPase binding domain ( RBD ) [11]–[13] . Recent structural studies of the intracellular region of human Plexin-B1 and mouse Plexin-A3 revealed that the GAP domain is an integral structural unit , with the RBD forming a domain insertion into one of the exposed GAP domain loops [14] , . Importantly , the catalytic machinery remained identical , with catalytic arginines found in the same positions in RasGAP and both Plexin-B1 and Plexin-A3 [14]–[17] . Within the plexin family , the human Plexin-B1 signalling pathway is the most extensively characterized to date; two members of the Ras superfamily have been identified as targets of the Plexin GAP activity so far , R-Ras and M-Ras [9] , [18] . Inactivation of R-Ras by Plexin-B1 GAP leads to suppression of integrin activation and cell migration , ultimately leading to repulsive axonal guidance [19] , [20] . Downregulation of M-Ras leads to reduced dendritic outgrowth and branching [18] . The Plexin-B1 RBD has been shown to bind to the Rho GTPases Rnd1 , Rac1 , and RhoD exclusively in their active , GTP-bound form [21]–[23] . Small GTPases of the Rho family are key players in remodelling of the actin cytoskeleton and are involved in a plethora of processes initiated by extracellular stimuli [24] , [25] . Both Rac1 and Rnd1 are important for the ligand-induced activation of the plexin GAP activity and Rac1 has been found to increase semaphorin binding to Plexin-B1 [19] , [26]–[28] . Simultaneous binding of semaphorin on the extracellular side and a RhoGTPase on the intracellular side is a prerequisite for plexin GAP activity [27] , [29] . Bivalent semaphorin binding can be mimicked by extracellular , antibody-induced , clustering of the intracellular domain and activation is observed in the presence but not in the absence of Rnd1 [9] , [29] . This suggests that semaphorins have a crucial role in bringing together plexin receptors as a step towards activation . Despite a number of structural studies on the plexin RBD and its complex with Rnd1 [15] , [30] , [31] it remains unclear how RhoGTPases modulate plexins and how the concomitant binding of ligands on the extracellular and the intracellular side of the receptor is integrated into a single signalling output , inactivation of Ras . To address this question we characterized the complex between the intracellular region of Plexin-B1 and a constitutively active form of the RhoGTPase Rac1 both structurally and functionally .
Several constructs of the intracellular domain of human Plexin-B1 were designed , of which three , Plexin-B1cyto , Plexin-B1Δ1 , and Plexin-B1Δ2 , could be solubly expressed in insect cells ( Figure 1a ) . Rac1 was rendered constitutively active by introducing a Gln61Leu mutation [32] in addition to loading with the non-hydrolyzable GTP analogue GppNHp . This Rac1 mutant , expressed in E . coli , was used in all subsequent experiments and is named Rac1* hereafter . We have determined the crystal structure of Plexin-B1Δ1 in complex with Rac1* to a resolution of 3 . 2 Å and refined it to a crystallographic R-factor of 20 . 7% ( Rfree = 23 . 8% , Figure 1b , Table 1 , Figure S1 ) . The overall structures of Plexin-B1Δ1 and Rac1* in the complex are very similar to their apo-structures [15] , [33] with rmsd values of 1 . 5 Å and 0 . 6 Å , respectively . However , there is some flexibility between the Plexin-B1 GAP and the RBD with the RBD being rotated by ∼6° compared to the apo-structure ( Figure S2 ) . Rac1* binds exclusively to the RBD and does not form any contacts with the GAP domain . The interface between Rac1* and the Plexin-B1 RBD covers a buried surface area of 707 Å2 and is dominated by hydrophobic interactions . Plexin-B1 residues Trp1807Plex , Leu1815Plex , Thr1823Plex , and Tyr1839Plex form a continuous hydrophobic patch that is complemented by Rac1 residues Phe37Rac , Val36Rac , Leu67Rac , and Leu70Rac ( Figure 1b ) . All of these residues are almost completely buried within the interface ( at least 80% of the solvent accessible surface area ) with the exception of Val36Rac ( 38% ) . Thr1823Plex and Tyr1839Plex are part of a potential hydrogen bonding network involving Asp1821Plex , Ser1824Plex , Asn1834Plex , and His1838Plex that is likely to be crucial for the structural integrity of the domain . The hydrophobic interaction between Plexin-B1Δ1 and Rac1* is extended by two potential hydrogen bonds formed between the sidechain of Asp38Rac and the backbone amides of Val1811Plex and Ala1812Plex . Remarkably , all of the Plexin-B1 residues described above are conserved across A- and B-class plexins ( Figure S3 ) , therefore most likely preserving this mode of recognition . On Rac1* , all residues mentioned above map onto the switch I or switch II region [11] ( Figure S4 ) whose conformation resembles that of active Rac1 in other Rac1-effector complexes [34] . Since these regions undergo large conformational changes upon GTP binding , this explains why Plexin-B1 is highly specific for active , GTP-bound Rac1 [21] . Recently , the structure of the RBD fragment of Plexin-B1 in complex with the constitutively active RhoGTPase , Rnd1 , has been reported [15] . Structural superposition of the RBD-RhoGTPase complexes gives an rmsd of 0 . 96 Å ( Figure S5 ) . Despite a sequence identity of only 32% between Rac1 and Rnd1 , the Plexin-B1 RBD-Rnd1 complex interface is very similar to the one described here . All hydrophobic interactions as well as the two potential hydrogen bonds are conserved in both structures . To corroborate our structural findings we studied the affinity between Plexin-B1cyto and Rac1* , as well as Rnd1 , using surface plasmon resonance ( SPR ) . Rnd1 is constitutively active due to its lack of GTPase activity [35] . Plexin-B1cyto binds to Rac1 and Rnd1 with an affinity of 18 . 9 µM and 22 . 9 µM , respectively ( Figure 1c–e , Figure S6 ) , which is in agreement with recently published affinities determined by isothermal titration calorimetry [15] . We found that a series of Plexin-B1 mutations in the hydrophobic interface , Trp1807GluPlex , Leu1815ProPlex ( previously linked to prostate cancer [4] ) , and Leu1815GluPlex , completely abolished its interactions with Rac1* and Rnd1 ( Figure 1d–e , Figure S6 ) . To validate these effects on binding in a functional context , we performed COS cell-based collapse assays with the full-length transmembrane receptor , testing for Plexin-B1 activity in vivo [36] . Surprisingly , none of the mutants shown to abolish Rac1* or Rnd1 binding had an effect on the collapse response of the cells ( Figure 1f–h ) . We explored this finding further in an independent experimental assay to monitor directly Ras GTPase activity in vivo . In agreement with our results from the collapse assay , none of the interface mutants had an effect on the GAP activity of Plexin-B1 towards R-Ras in this COS cell-based pull-down ( Figure S7 ) . Since the necessity of RhoGTPase binding for plexin function is well established [9] , [27] , it was unclear how to correlate the biophysical and cellular results . The relative position of Rac1* in regard to the putative Ras binding site revealed no mechanism for the direct regulation of the catalytic activity of Plexin-B1 by the small RhoGTPase . To address whether the N-terminal residues missing in the Plexin-B1Δ1 construct might harbour an important site for RhoGTPase mediated plexin activity , we solved the crystal structure of the entire cytoplasmic domain of Plexin-B1 ( Plexin-B1cyto ) in complex with Rac1* ( Figure 2a ) . The 4 . 4 Å model is of high-quality for this resolution range , reflected by the crystallographic R-factor of 23 . 4% ( Rfree = 26 . 4% , Table 1 , Figure S8 ) . The asymmetric unit contains a trimeric arrangement comprising three copies of the Plexin-B1cyto-Rac1* unit , with each Rac1* molecule contacting two Plexin-B1cyto molecules ( Figure 2a–b ) . This arrangement is not the result of crystallographic symmetry but does show near perfect 3-fold geometry ( 120° between pairs of Rac1* molecules and 117° , 119° , and 124° between the copies of Plexin-B1cyto ) . Moreover , the interfaces between Plexin-B1cyto and Rac1* are essentially identical across the three copies in the asymmetric unit and are not found in any of the crystallographic symmetry generated interfaces . These observations strongly suggest that this 3-fold complex is not purely a product of crystal lattice formation . We were not able to show a 3-fold complex with the soluble constructs in analytical ultracentrifugation at a plexin concentration of 250 µM ( unpublished data ) . This suggests that high local concentrations in the crystal or indeed at the plasma membrane are necessary for this arrangement and that the 3-fold interaction might be too weak to be detected in solution [37] . There are no major conformational changes within the three Plexin-B1cyto-Rac1* units when compared to the one-to-one complex . This 3-fold arrangement is mediated by a previously unidentified second binding site for the Rho-GTPase on Plexin-B1cyto ( site B , the previously observed Plexin-B1 RBD-RhoGTPase interface hereafter called site A ) . Site B involves the N-terminal region of α-helix 11 and the loop preceding it ( residues 1913Plex–1923Plex ) plus α-helix 16 ( residues 2036Plex–2039Plex ) and is in close proximity to the putative Ras binding site ( Figure 2a , c , Figure S3 ) . It covers a total buried surface area of ∼570 Å2 , therefore significantly extending the interface for Rac1 binding , and is dominated by hydrophobic interactions . On Plexin-B1cyto the majority of contacts ( 60% of buried surface area ) are made by the loop residues 1913Plex–1918Plex ( Figure 2c ) . Interestingly , these residues are disordered in the apo-structures of Plexin-B1 and Plexin-A3 . The site B interface on Rac1* is predominantly formed by residues that precede the switch I region ( residues 24–33 , Figure 2c ) . The conformation of these residues is known not to depend on the activation state of the RhoGTPase [11]; thus , the specificity of Plexin-B1 for active RhoGTPases appears to result exclusively from interactions with site A . The 3-fold complex is further stabilized by contacts between two adjacent Plexin-B1cyto molecules , on the one side mainly involving a loop comprising residues 1808Plex–1813Plex , and on the other side a surface directly adjacent to site B ( residues 1919Plex–1938Plex and residues 2036Plex–2044Plex , Figure 2c ) . However , this plexin-plexin interaction is unlikely to be stable without the addition of the bridging Rac1* since it only contributes a total buried surface area of ∼310 Å2 . In order to assess the potential functional significance of site B , we designed three Plexin-B1 mutants , Thr1920GluPlex , Arg1921AlaPlex , and Leu2036ArgPlex . We first studied the binding affinity of these mutants to Rac1* and Rnd1 using SPR . None of the site B mutations had a significant effect on the affinity towards either of the RhoGTPases , suggesting that site A alone is sufficient for Rac1* and Rnd1 binding ( Figure 3a , Figure S9 ) . However , these mutations as well as additional ones at these and other site B residues ( Ile1917Plex , Leu1923Plex , and Ala2039Plex ) completely abolished the typical collapse response in the COS-cell assay ( Figure 3b ) . Every site B mutation tested was detrimental to Plexin-B1 activity and led to a complete loss of function . All mutant proteins showed a similar expression level , as judged by immunofluorescence , and were present in the plasma membrane , indicating their structural integrity ( unpublished data ) . In the same background , the Plexin-B1 site B mutants also did not show any GAP activity towards R-Ras , thus further validating the findings of the collapse assay ( Figure S7 ) . These results indicate that although site B is not essential for binding of the RhoGTPase , it is crucial for Plexin-B1 activity , suggesting that the 3-fold complex seen in the crystal has functional significance . In accordance with this putative functional role , all of the residues in site B are conserved across all species and classes of plexins with the exception of Ala1913Plex and Pro1915Plex , whose sidechains do not participate in the Rac1-site B interaction ( Figure 3c ) . The 3-fold complex revealed by the crystal structure and the cell collapse assays suggest that GTPase binding at site B contributes to plexin function . However , the SPR experiments reveal no direct evidence for GTPase binding at this site . We therefore sought an explanation for this lack of binding . In both the one-to-one and 3-fold complex structures , we were unable to trace the N-terminal helix ( residues 1511Plex–1562Plex , Figure S3 ) due to a lack of well-ordered electron density . Interestingly , there is a similar absence of electron density in this region for the high resolution apo-structure of Plexin-B1 [15] . This suggests that the N-terminal helix of Plexin-B1 has some internal flexibility , likely around the hinge region adjacent to Ile1563 . In agreement with this , the three residues preceding Ile1563 are Gly1562 , Ser1561 , and Gly1560 , which may allow large conformational freedom of the N-terminal helix even in a trimeric arrangement . In contrast , in the apo-structure of mouse Plexin-A3 [14] , this region was well-defined . Superposition of the Plexin-A3 structure with the 3-fold complex reveals that this helix would block site B , therefore preventing its interaction with Rac1* ( Figure 4a ) . This steric hindrance model predicts that shortening of the N-terminal helix will remove this block and allow Rac1* and Rnd1 to bind to site B . To test this model we generated mutant constructs lacking the N-terminal helix ( Plexin-B1Δ2 ) and assayed for RhoGTPase binding in SPR . Indeed , RhoGTPase binding to the site A Plexin-B1Δ2 mutant Leu1815GluPlex was now observed , suggesting that truncation of the N-terminal helix has exposed site B ( Figure 4b , Figure S10 ) . However , binding can only be observed with high coupling densities of the Plexin-B1Δ2 on the SPR chip ( Figure 4b ) . This is consistent with a bivalency effect in which two adjacent plexin molecules bind the same RhoGTPase molecule , implying that the mutated site A is still competent to contribute to an avidity effect [38] . At low coupling densities the Plexin-B1Δ2 molecules are too far apart from each other to allow a bivalent interaction with Rac1* or Rnd1 . We did not observe an increase in affinity for wild-type N-terminal truncated Plexin-B1Δ2 compared to the full-length Plexin-B1cyto even at high coupling densities ( unpublished data ) . Binding studies on the isolated Plexin-B1 RBD show similar affinities [21] to those we determined for the full-length cytoplasmic region . Thus these observations suggest that , if intact , site A dominates binding to the RhoGTPase . Interestingly , sequence analysis of the N-terminal cytoplasmic segment of Plexin-B1 ( residues 1511–1539 ) predicts a trimeric coiled-coil ( Figure S11 ) and similar regions in other plexins from all classes are also predicted to adopt a coiled-coil conformation . In accordance with this , Ile1563Plex in Plexin-B1cyto , the first residue visible in the electron density map , points towards the inside of the 3-fold complex locating the three N-terminal segments in close proximity to each other ( Figure 2a , right panel ) . This proximity suggests an explanation for the observation of the higher oligomeric state 3-fold complex in the crystal . Although the N-terminal helix is not well-ordered in the structure , it could form a trimeric coiled-coil , albeit containing significant flexibility .
Plexin-semaphorin signalling is dependent on signals from both the extra- and intracellular side . Several studies have shown that both semaphorin binding on the outside and RhoGTPase binding on the inside of the cell are required for plexin activity to occur [9] , [27] . The nature of these signals and how they are integrated into a single output , namely RasGAP activity , has been a critical question in this field and several models have been proposed [6]–[8] , [14] , [15] , [39] . Recently , several structures of truncated plexin ectodomains in complex with their cognate semaphorins have been reported [6]–[8] . Despite ranging across three different classes , all of these ectodomain complexes share the same overall architecture with one semaphorin dimer bringing together two plexin monomers . In combination with a detailed biophysical and cellular characterization , these structural data have led to the proposal that the bivalency effect is a prerequisite for plexin signalling [6] , [8] . For the cytoplasmic region , our structures of Plexin-B1 in complex with Rac1* do not show major structural rearrangements when compared to the apo-structure of Plexin-B1 [15] . For the one-to-one complex , Rac1* is positioned distant from the Ras binding site on the Plexin-B1Δ1 molecule . This excludes the possibility of a direct interaction or regulation of RasGAP activity by the RhoGTPase . Instead , the 3-fold complex reveals an additional binding site on a neighbouring Plexin-B1cyto molecule that is in close proximity to the predicted Ras binding site . This interaction may result in a small conformational change in the Ras binding region , although a detailed analysis of these changes cannot be made due to the low resolution of our data . It is , however , noteworthy that allosteric regulation of Ras binding by RhoGTPase binding has been proposed by He et al . based on a homology model of the Plexin-A3-Rnd1 complex [14] . We cannot exclude the possibility that within the protein crystal the trimeric arrangement is favoured over other site B mediated oligomeric states due to the lattice contacts . Indeed , the 3-fold arrangement constitutes the asymmetric unit in the crystal and therefore accommodates slight variations between the three Plexin-B1-Rac1 units ( see Results section ) . The occurrence of a 3-fold arrangement in crystals of Plexin-B1-Rac1 complexes appears to be dependent on the juxtamembrane , N-terminal helix . Physiologically , this region connects the intracellular domain with the transmembrane and extracellular region . This suggests a mechanism by which both semaphorin binding on the outside and RhoGTPase binding on the inside are connected to result in RasGAP activity ( Figure 5 ) . The first step in this model is binding of the RhoGTPase to binding site A of the intracellular domain . Although RhoGTPase binding has been shown to be a prerequisite for Ras binding , it is not sufficient to trigger signalling [9] . Semaphorin binding on the outside of the cell may result in clustering of the receptors [6] either from an autoinhibited , monomeric , or dimeric state [14] , [15] , [39] . Such extracellular rearrangement could be transmitted to the intracellular N-terminal helix . The rearrangement of this juxtamembrane helix would free up binding site B , allowing the RhoGTPase to bridge two plexin molecules and stabilize the 3-fold arrangement . Formation of a trimeric cluster could result in the proper positioning of the catalytic machinery allowing RasGAP activity to occur , since it has been shown that clustering of the intracellular domain is crucial for this activity [9] . In summary , we propose that receptor clusters nucleated by the dimeric complex on the extracellular side and the trimeric complex on the intracellular side will integrate both RhoGTPase and Semaphorin binding into a single signalling output .
A series of constructs of the intracellular domain of human Plexin-B1 ( GenBank ID: NP_001123554 ) lacking both C- and N-terminal regions as well as the RBD were designed and cloned into pBacPAK9 with a C-terminal His6-Tag for purification . Of these constructs three could be solubly expressed via baculovirus infection in Sf9 cells ( Plexin-B1Δ1 , residues 1533–2135; Plexin-B1Δ2 , residues 1543–2135; and Plexin-B1cyto , residues 1511–2135 ) . Cells were harvested at 2 , 000×g for 15 min , resuspended in binding buffer ( 20 mM phosphate , pH , 7 . 4 , 500 mM NaCl , 0 . 5 mM β-mercaptoethanol ) , sonicated , and then centrifuged at 46 , 000×g for 1 h at 4°C . The supernatant was collected and the protein was purified by ion metal affinity chromatography followed by size exclusion chromatography in 10 mM Hepes , pH 7 . 5 , 150 mM NaCl , 2 mM TCEP [40] . Mutations were generated by a two-step overlapping PCR using Pyrobest Polymerase ( Takara ) . Mutant plexin constructs used for SPR studies were expressed in human HEK 293T cells essentially as described [41] . Three days after transfection the cells were harvested and purified following the protocol used for the wild-type proteins . All mutant proteins had similar expression level compared to Plexin-B1cyto as determined by SDS-PAGE . Rac1 Gln61Leu ( residues 1–176 , GenBank ID: CAB53579 ) and Rnd1 ( residues 5–200 , GenBank ID: BAB17851 ) were cloned into the expression vector pET22b , expressed in E . coli BL21 Star ( Invitrogen ) , and purified following an established protocol described elsewhere [40] . After purification Rac1 was incubated with 10 mM EDTA , pH 8 . 0 , and calf intestine alkaline phosphatase ( NEB ) to degrade any bound nucleotide . Subsequently the protein was loaded with the non-hydrolyzable GTP analogue GppNHp and purified by size exclusion chromatography in 10 mM Hepes , pH 7 . 5 , 150 mM NaCl , 2 mM MgCl2 , 2 mM TCEP . SEMA4Decto ( residues 22–677 ) was expressed in CHO lecR cells as previously described [6] . The Ras binding domain of Raf-1 ( residues 51–131 ) was fused to GST ( GST-RBD ) , expressed in E . coli BL21 Star ( Invitrogen ) , and purified following an established protocol described elsewhere [40] . Prior to crystallization all proteins were concentrated by ultrafiltration to 10 mg/ml and complexes were formed by mixing Plexin-B1 and RhoGTPase in a 1∶1 . 2 molar ratio . Nano-litre crystallization trials were set-up using a Cartesian Technologies robot ( 100 nl protein solution plus 100 nl reservoir solution ) in 96-well Greiner plates [42] , placed in a TAP Homebase storage vault maintained at 295 K , and imaged via a Veeco visualization system [43] . The PlexinB1cyto-Rac1* complex crystallized in 1 M Li2SO4 , 0 . 5 M ammonium sulphate , 0 . 1 M citrate , pH 5 . 6 , and Plexin-B1Δ1-Rac1* complex crystallized in 20% PEG 3350 , 0 . 2 M KSCN , 0 . 1 M Bis-Tris Propane , pH 6 . 5 . Diffraction data were collected at 100 K with the crystals being flash-cooled in a cryo N2 gas stream . Prior to flash-freezing , crystals were treated with a cryo protectant solution consisting of 25% ( v/v ) glycerol in mother liquor . The Plexin-B1Δ1-Rac1* crystals crystallized as thin needles and data were collected at the microfocus beamline ID23-2 at the European Synchrotron Radiation Facility , France , following a helical data collection strategy . Plexin-B1cyto-Rac1* crystals crystallized as thin squares and data were collected at beamline I03 at Diamond Light Source , UK . X-ray data were processed and scaled with the HKL suite [44] . Data collection statistics are shown in Table 1 . Both structures were solved by molecular replacement using PHASER [45] with the structure of human Plexin-B1 ( PDB ID: 3HM6 [15] ) and active Rac1 ( PDB ID: 1MH1 [33] ) as search model . The solution was manually adjusted using COOT [46] and refined using autoBUSTER [47] . Refinement statistics are given in Table 1; all data within the indicated resolution range were included . The 4 . 2 Å structure was refined using 3-fold NCS as implemented in autoBUSTER [47] and tight geometric restraints to minimize the introduction of any model bias . Stereochemical properties were assessed by MOLPROBITY [48] . Ramachandran statistics are as follows ( favoured/disallowed ( % ) ) : Plexin-B1cyto-Rac1* 91 . 7/0 . 2 , Plexin-B1Δ1-Rac1* 95 . 5/0 . 2 ( pre-proline residue Leu1981 is in a disallowed region in both structures ) . Superpositions were calculated using SHP [49] . Buried surface areas of protein-protein interactions were calculated using the PISA webserver ( http://www . ebi . ac . uk/msd-srv/prot_int/pistart . html ) . SPR experiments were performed using a Biacore T100 machine ( GE Healthcare ) at 25°C in standard buffer supplemented with 0 . 05% ( v/v ) Tween 20 . Protein concentrations were determined from the absorbance at 280 nm using calculated molar extinction coefficients . All plexin constructs for surface attachment were enzymatically biotinylated within an engineered C-terminal tag . These proteins were then attached to surfaces on which 5 , 000 RU of streptavidin were coupled via primary amines [50] yielding a density of 500–5 , 000 response units ( RU ) of biotinylated protein . All experiments were done in duplicates with independently purified proteins . The signal from experimental flow cells was corrected by subtraction of a blank and reference signal from a mock or irrelevant protein coupled flow cell . In all experiments analyzed , the experimental trace returned to baseline after each injection and the data fitted to a simple 1∶1 Langmuir model of binding . Kd values were obtained by nonlinear curve fitting of the Langmuir binding isotherm ( bound = C*max/ ( Kd+C ) , where C is analyte concentration and max is the maximum analyte binding ) evaluated using the Biacore Evaluation software ( GE Healthcare ) . Cellular collapse assays were performed essentially as described [36] . Briefly , COS-7 cells were seeded on glass coverslips and transfected with full-length human Plexin B1 carrying an N-terminal Flag-tag essentially as described [42] . Two days after transfection , cells were treated with medium containing secreted SEMA4Decto and incubated for 30 min at 37°C . Finally , the cells were fixed and stained with anti-Flag primary antibody ( Sigma ) and Alexa 488-labelled secondary antibody ( Invitrogen ) . Cell nuclei were counterstained with DAPI ( Invitrogen ) and cells were visualized with a TE2000U fluorescence microscope ( Nikon ) equipped with an Orca CCD camera ( Hamamatsu ) . Plexin B1-expressing cells were classified as collapsed or non-collapsed on the basis of reduced surface area . Each experiment was repeated twice and 2×200 cells were counted each time . Results are shown as mean with error bars representing standard error of the mean . Pull-down assays were performed essentially as described [51] . COS-7 cells were seeded in 6-well dishes and transfected with full-length human Plexin-B1 and its mutants , respectively , and R-Ras . Two days after transfection , cells were treated with medium containing secreted SEMA4Decto and incubated for 10 min at 37°C . Cells were washed twice with ice-cold phosphate-buffered saline and then lysed with lysis buffer ( 50 mM Tris-HCl , pH 7 . 5 , 200 mM NaCl , 5 mM MgCl2 , 10% glycerol , 1% Non-ident P-40 substitute , 2 mM β-mercaptoethanol ) . Cell lysates were incubated with GST-RBD pre-coupled to glutathione-agarose beads ( GE Healthcare ) for 45 min at 4°C . After three wash steps with lysis buffer the beads were collected in Laemmli sample buffer and analyzed by SDS-PAGE and immunoblotting with R-Ras- and GST-specific antibodies , respectively . | Axon guidance is fundamental to the development of the central nervous system . The growing axon is guided to its correct location by a plethora of extracellular signals . One of the most important extracellular signals is semaphorin , which binds to plexin receptors on the axon . Usually , this kind of extracellular ligand binding is sufficient to transmit the extracellular signal to the intracellular space to trigger changes in the cell , like axon growth . However , activation of plexin receptors requires a “dual” ligand binding: semaphorin on the extracellular side , and a RhoGTPase on the intracellular side . Signal transduction can only occur if both ligands are present . How this intricate regulation mechanism is organized and how concomitant ligand binding can be integrated into a single signalling output within the cell has remained largely unclear . Here , we present crystal structures of one plexin receptor , Plexin-B1 , in complex with an intracellular RhoGTPase ligand ( Rac1 ) and show that binding of Rac1 brings together three Plexin-B1 molecules . In this trimeric arrangement each plexin molecule interacts with two Rac1 ligand molecules . This leads to a previously unidentified plexin-Rac1 ligand interface that is crucial for its function . Further biophysical and cellular analysis in combination with previous findings on the extracellular plexin-semaphorin complex allow us to propose a model for how ligand-induced clustering events on the extra- as well as intracellular side are combined to trigger signal transduction . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"molecular",
"neuroscience",
"protein",
"interactions",
"mechanisms",
"of",
"signal",
"transduction",
"signaling",
"in",
"selected",
"disciplines",
"neuroscience",
"gtpase",
"signaling",
"developmental",
"signaling",
"cell",
"movement",
"signaling",
"protein",
"structure",
"signaling",
"pathways",
"signaling",
"in",
"cellular",
"processes",
"proteins",
"ras",
"signaling",
"signal",
"initiation",
"biology",
"recombinant",
"proteins",
"molecular",
"biology",
"biochemistry",
"signal",
"transduction",
"molecular",
"cell",
"biology"
]
| 2011 | A Dual Binding Mode for RhoGTPases in Plexin Signalling |
The microRNA ( miRNA ) let-7 is an important miRNA identified in Caenorhabditis elegans and has been shown to be involved in the control of innate immunity . The underlying molecular mechanisms for let-7 regulation of innate immunity remain largely unclear . In this study , we investigated the molecular basis for intestinal let-7 in the regulation of innate immunity . Infection with Pseudomonas aeruginosa PA14 decreased let-7::GFP expression . Intestine- or neuron-specific activity of let-7 was required for its function in the regulation of innate immunity . During the control of innate immune response to P . aeruginosa PA14 infection , SDZ-24 was identified as a direct target for intestinal let-7 . SDZ-24 was found to be predominantly expressed in the intestine , and P . aeruginosa PA14 infection increased SDZ-24::GFP expression . Intestinal let-7 regulated innate immune response to P . aeruginosa PA14 infection by suppressing both the expression and the function of SDZ-24 . Knockout or RNA interference knockdown of sdz-24 dampened the resistance of let-7 mutant to P . aeruginosa PA14 infection . Intestinal overexpression of sdz-24 lacking 3’-UTR inhibited the susceptibility of nematodes overexpressing intestinal let-7 to P . aeruginosa PA14 infection . In contrast , we could observed the effects of intestinal let-7 on innate immunity in P . aeruginosa PA14 infected transgenic strain overexpressing sdz-24 containing 3’-UTR . In the intestine , certain SDZ-24-mediated signaling cascades were formed for nematodes against the P . aeruginosa PA14 infection . Our results highlight the crucial role of intestinal miRNAs in the regulation of the innate immune response to pathogenic infection .
The free-living nematode Caenorhabditis elegans lives in soil , and potentially composts rich in the microorganisms , including those that are human microbial pathogens [1–2] . In the laboratory , once pathogenic bacteria can be deposited in the gut , they will invade the host cells , and even kill the nematodes through infectious processes . C . elegans usually responds to pathogen exposure by avoiding pathogens or activating an inducible innate immune system [3] . C . elegans has been considered as a useful model for the study of innate immunity at least based on the identification of virulence-related microbial genes and immune-based host genes [4] . Moreover , studies in C . elegans may potentially provide evolutionary and mechanistic insights into the signal transduction and the physiology of innate immunity [3] . Both genetic and functional genomic approaches in C . elegans have identified some conserved and important signaling pathways required for the control of innate immunity . Some of these identified conserved signal pathways involve p38 mitogen-activated protein kinase ( MAPK ) , insulin , and TGF-β signaling pathways [5–8] . microRNAs ( miRNAs ) are a class of evolutionarily conserved non-coding RNAs with 19–22 nucleotides , and function to negatively regulate the gene expression [9] . miRNAs are encoded within the genome , and mature miRNAs post-transcriptionally regulate the gene expression by imperfectly binding to multiple target mRNAs [10] . miRNAs have been shown to be involved in the control of diverse fundamental biological processes , such as development , cell differentiation , apoptosis , and immune response [9 , 11–13] . C . elegans is a powerful in vivo model to study how miRNAs regulate the gene expression and regulate various biological processes during the development [11 , 14] . Recently , some miRNAs such as mir-84 , mir-241 , mir-233 , mir-251 , mir-252 , and mir-360 have been shown to be involved in the control of innate immunity in C . elegans [15–17] . let-7 is one of the founding members of miRNA family firstly identified in C . elegans via forward genetic screen [18] . Some studies have suggested that let-7 can act as a developmental switch to control the transition from larval to adult [18–19] . More recently , it has been shown that let-7 was also involved in the control of innate immunity [20] . However , the underlying molecular mechanism for let-7 in the regulation of innate immunity is still largely unclear . In the present study , we aimed to examine the molecular basis for intestinal let-7 in the regulation of innate immunity . We identified SDZ-24 as a potential target of intestinal let-7 in the regulation of innate immunity . In C . elegans , sdz-24 gene encodes an ortholog of human replication protein A1 ( RPA1 ) [21] . SKN-1 , a homolog to Nrf transcription factors , plays an important role in pathogen resistance [22] . Our results suggest that intestinal let-7 could directly target the SDZ-24 , and regulated the innate immune response to Pseudomonas aeruginosa PA14 infection by suppressing the function of SDZ-24-SKN-1 signaling cascade . Our findings will not only aid our understanding the molecular basis for intestinal let-7 in the regulation of innate immunity , but will also be helpful for further understanding the complex biological functions of let-7 in animals .
In C . elegans , let-7 is expressed in almost all tissues except the germline [23] . Using the transgenic strain of zaEx5[let-7::GFP] , we observed that exposure to P . aeruginosa PA14 at least significantly decreased the expression of let-7::GFP in the pharynx , the neurons , and the intestine of infected transgenic strain compared with exposure to Escherichia coli OP50 ( Fig 1A ) . Therefore , P . aeruginosa PA14 infection may suppress the let-7 expression in the body of nematodes . A previous study has demonstrated that the loss-of-function mutant of let-7 ( mg279 ) was resistant to P . aeruginosa PA14 infection as indicated by the increased survival compared with that in infected wild-type nematodes [20] . Meanwhile , we found that the loss-of-function mutation of let-7 significantly decreased the colony-forming unit ( CFU ) of P . aeruginosa PA14 in the body of nematodes ( Fig 1B ) . The normal accumulation of PA14::GFP in the lumen of pharynx of let-7 ( mg279 ) mutant implies that mutation of let-7 did not cause the deficit in the feeding of P . aeruginosa PA14 ( S1 Fig ) . Moreover , we investigated the effect of loss-of-function mutation of let-7 on the expression of antimicrobial genes . The examined antimicrobial genes were lys-1 , lys-8 , clec-85 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 . P . aeruginosa PA14 could induce the significant increase in transcriptional expression of these antimicrobial genes in C . elegans [24] . We observed the noticeable increase in transcriptional expression of some antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) in P . aeruginosa PA14 infected let-7 ( mg279 ) mutant compared with P . aeruginosa PA14 infected wild-type nematodes ( Fig 1C ) . On plates fed with E . coli OP50 , the expression levels of these antimicrobial genes in let-7 ( mg279 ) mutant were similar to those in wild-type nematodes ( Fig 1C ) . The let-7 ( mg279 ) mutant showed the similar brood size to wild-type ( S2 Fig ) , suggesting that the observed resistance to PA14 infection in let-7 ( mg279 ) mutant is not the potential effect of let-7 mutation on fecundity . These results suggest that the loss-of-function mutation of let-7 may suppress the P . aeruginosa PA14 colonization in the body of nematodes , and induce an elevated immune competence . In this study , we next investigated the intestine- , neuron- , pharynx- , muscle- , or hypodermis-specific activity of let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection . After P . aeruginosa PA14 infection , expression of let-7 in the pharynx , the muscle , or the hypodermis did not significantly affect the survival and the CFU of P . aeruginosa PA14 in let-7 ( mg279 ) mutant ( Fig 2A and 2B ) . In contrast , after P . aeruginosa PA14 infection , expression of let-7 in the intestine or the neurons significantly reduced the survival , increased the CFU of P . aeruginosa PA14 , and decreased the expressions of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) in let-7 ( mg279 ) mutant ( Fig 2 ) . Additionally , the survival , the CFU of P . aeruginosa PA14 , or the expression patterns of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) in P . aeruginosa PA14 infected transgenic strain of let-7 ( mg279 ) Ex ( Pges-1-let-7 ) or let-7 ( mg279 ) Ex ( Punc-14-let-7 ) were similar to those in P . aeruginosa PA14 infected wild-type nematodes ( Fig 2B ) . Therefore , let-7 may act in the intestine or the neurons to regulate the innate immune response to P . aeruginosa PA14 infection . To confirm the function of let-7 in the regulation of innate immunity , we investigated the effects of let-7 overexpression in the intestine or the neurons on innate immune response to P . aeruginosa PA14 infection . After exposure to P . aeruginosa PA14 , nematodes overexpressing let-7 in the intestine or the neurons exhibited the significant decrease in survival compared with wild-type nematodes ( S3A Fig ) . After exposure to P . aeruginosa PA14 , overexpression of let-7 in the intestine or the neurons also resulted in a significant increase in CFU of P . aeruginosa PA14 in the body compared with wild-type nematodes ( S3B Fig ) . Moreover , after exposure to P . aeruginosa PA14 , nematodes overexpressing let-7 in the intestine or the neurons showed the decreased expression in antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) compared with wild-type nematodes ( S3C Fig ) . These results suggest that the overexpression of let-7 in the intestine or the neurons may induce a susceptible property of nematodes to P . aeruginosa PA14 infection . To identify the potential targets for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection , we first used the TargetScan software to predict the potential targets for let-7 by searching for the presence of conserved sites that match the seed region of let-7 , and found 351 potential targets for let-7 in C . elegans ( Fig 3A ) . In C . elegans , a previous study has suggested that 697 genes could be dysregulated by P . aeruginosa PA14 infection [25] . Among the predicted 351 targeted genes , 7 genes ( tag-38 , nhr-43 , mtl-1 , nex-4 , sdz-24 , Y95B8A . 6 , and K01A2 . 10 ) were also dysregulated by P . aeruginosa PA14 infection ( Fig 3A ) . Using the corresponding mutants or RNA interference ( RNAi ) knockdown animals for these 7 genes , we found that only loss-of-mutation of nhr-43 gene or RNAi knockdown of sdz-24 gene significantly influenced the survival of infected nematodes after P . aeruginosa PA14 infection ( Fig 3B ) . After P . aeruginosa PA14 infection , loss-of-mutation of nhr-43 gene or RNAi knockdown of sdz-24 gene significantly reduced the survival in nematodes ( Fig 3B ) . In C . elegans , NHR-43 is a nuclear hormone receptor ( NHR ) , and SDZ-24 is a SKN-1-dependent zygotic protein . Meanwhile , loss-of-mutation of nhr-43 gene or RNAi knockdown of sdz-24 gene significantly enhanced the colony formation of P . aeruginosa PA14 in the body of nematodes ( Fig 3C ) . Moreover , loss-of-mutation of nhr-43 gene caused the decreased expression in antimicrobial genes ( lys-1 , lys-8 , clec-85 , and F55G11 . 7 ) , and RNAi knockdown of sdz-24 gene resulted in the decreased expression in antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) ( Fig 3D ) . To confirm the role of nhr-43 or sdz-24 as the targeted gene of let-7 , we constructed the double mutants of nhr-43 ( tm1381 ) ;let-7 ( mg279 ) and sdz-24 ( yd101 ) ;let-7 ( mg279 ) . We assumed that the nhr-43 or sdz-24 mutation would suppress the phenotype in nematodes with let-7 mutation , if nhr-43 or sdz-24 is the targeted gene of let-7 in nematodes . To confirmed the function of SDZ-24 in the regulation of innate immunity to P . aeruginosa PA14 infection , we generated the knockout strain for sdz-24 ( sdz-24 ( yd101 ) ) using the clustered , regularly interspersed , short palindromic repeats ( CRISPR ) RNA-guided Cas9 technique [26] . The related deletion information was shown in S4 Fig . The survival , CFU of P . aeruginosa PA14 , and expression patterns of antimicrobial genes in P . aeruginosa PA14 infected sdz-24 ( yd101 ) were similar to those in P . aeruginosa PA14 infected sdz-24 ( RNAi ) nematodes ( Fig 4 ) . The sdz-24 ( yd101 ) mutant also had the normal accumulation of PA14::GFP in the lumen of pharynx , which was similar to that in wild-type nematodes ( S1 Fig ) . After P . aeruginosa PA14 infection , we found that mutation of sdz-24 gene significantly reduced the survival in let-7 ( mg279 ) mutant; however , mutation of nhr-43 gene did not obviously affect the survival in let-7 ( mg279 ) mutant ( Fig 4A ) . These results suggest that the SDZ-24 , not the NHR-43 , is the direct target for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection . The assays on CFU of P . aeruginosa PA14 and expression pattern of antimicrobial genes further confirmed the role of SDZ-24 as the direct target for let-7 in the regulation of innate immunity . Mutation of sdz-24 gene significantly increased the CFU of P . aeruginosa PA14 in the body of let-7 ( mg279 ) mutant ( Fig 4B ) . Moreover , mutation of sdz-24 gene significantly decreased the expression of antimicrobial genes ( ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in let-7 ( mg279 ) mutant ( Fig 4C ) . In C . elegans , SDZ-24::GFP is primarily expressed in the posterior of intestine ( S5 Fig ) . After P . aeruginosa PA14 infection , the expression of SDZ-24::GFP was significantly increased ( S5 Fig ) , which is different from the effect of the P . aeruginosa PA14 infection on expression of let-7::GFP as indicated above . Moreover , loss-of-function mutation of let-7 significantly increased the SDZ-24::GFP expression under both the E . coli OP50 exposure condition and the P . aeruginosa PA14 exposure condition ( S5 Fig ) , suggesting that let-7 may potentially suppress the expression of SDZ-24 . Considering the fact that the sdz-24 gene is pronounced expressed in the intestine , we investigated the effects of intestinal RNAi of sdz-24 gene on innate immune response to P . aeruginosa PA14 infection using the VP303 strain as the intestinal RNAi tool [27] . After P . aeruginosa PA14 infection , intestinal RNAi of sdz-24 gene significantly reduced the survival , enhanced the CFU of P . aeruginosa PA14 in the body of nematodes , and decreased the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) ( Fig 5A–5C ) . We also employed another intestinal RNAi tool of sid-1 ( qt9 ) to perform the RNAi assay [28] . We obtained the similar results ( S6 Fig ) . Therefore , intestinal RNAi of sdz-24 gene may induce a susceptible property of nematodes to P . aeruginosa PA14 infection . Different from the effects of intestinal RNAi of sdz-24 gene on innate immune response to P . aeruginosa PA14 infection , we observed that overexpression of sdz-24 gene in the intestine led to significant increase in the survival , decrease in the CFU of P . aeruginosa PA14 in the body of nematodes , and enhancement of the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) after P . aeruginosa PA14 infection ( Fig 5D–5F ) . These results suggest that overexpression of sdz-24 gene in the intestine may induce a resistance for nematodes against the toxic effects from P . aeruginosa PA14 infection , which further confirms the important function of SDZ-24 in the intestine in the regulation of innate immunity . To further determine the genetic interaction between let-7 and SDZ-24 in the intestine in the regulation of innate immunity , we introduced the sdz-24 lacking 3’-UTR driven by ges-1 promoter , an intestine-specific promoter , into the nematodes overexpressing intestinal let-7 . After P . aeruginosa PA14 infection , the transgenic strain of Ex ( Pges-1-sdz-24-3’-UTR ) ;Is ( Pges-1-let-7 ) exhibited the similar survival , CFU of P . aeruginosa PA14 , and expression patterns of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) to those in the transgenic strain of Ex ( Pges-1-sdz-24-3’-UTR ) ( Fig 6A–6C ) . Meanwhile , the survival , CFU of P . aeruginosa PA14 , and expression patterns of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in P . aeruginosa PA14 infected transgenic strain of Ex ( Pges-1-sdz-24-3’-UTR ) ;Is ( Pges-1-let-7 ) were significantly different from those in P . aeruginosa PA14 infected transgenic strain of Is ( Pges-1-let-7 ) ( Fig 6A–6C ) . These results suggest that the overexpression of sdz-24 gene lacking 3’-UTR in the intestine may effectively suppress the susceptible property of nematodes overexpressing intestinal let-7 . Very different from these , we found that intestinal overexpression of let-7 could significantly reduce the survival , enhance the CFU of P . aeruginosa PA14 , and decrease the expressions of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in P . aeruginosa PA14 infected transgenic strain of Ex ( Pges-1-sdz-24+3’-UTR ) ;Is ( Pges-1-let-7 ) ( Fig 6D–6F ) , suggesting the effects of intestinal expression of let-7 on innate immunity . To further confirm whether SDZ-24 is a direct target of let-7 , we constructed a GFP vector driven by ges-1 promoter , which contained the 3’-UTR of the sdz-24 ( Pges-1::GFP-3’-UTR ( sdz-24 wt ) or Pges-1::GFP-3’-UTR ( sdz-24 mut ) ) . Because let-7 can not bind to the tag-192 3’-UTR , a Pges-1::mCherry-3’-UTR ( tag-192 ) construct was used as an internal control . After P . aeruginosa PA14 infection , the GFP expression was noticeably reduced in wild-type nematodes; however , mutagenesis of the putative binding site for let-7 in sdz-24 3’-UTR abolished this inhibition of GFP expression in wild-type nematodes ( Fig 7 ) . Moreover , after P . aeruginosa PA14 infection , the GFP expression was much higher in let-7 ( mg279 ) mutant than that in wild-type nematodes ( Fig 7 ) . These results further confirmed that let-7 may suppress the function of SDZ-24 through binding to its 3’-UTR and inhibiting its translation in P . aeruginosa PA14 infected nematodes . To determine the intestinal SDZ-24-mediated signaling pathway in the control of innate immunity , we investigated the genetic interaction between SDZ-24 and some known signaling pathways involved in the control of innate immune response to P . aeruginosa PA14 infection . In C . elegans , daf-16 gene encodes a FOXO transcription factor in insulin signaling pathway , and dbl-1 gene encodes a TGF-β ligand in TGF-β signaling pathway . In the p38 MAPK signaling pathway , pmk-1 gene encoded a MAPK , sek-1 gene encoded a MAPK kinase ( MAPKK ) , and nsy-1 gene encoded a MAPK kinase kinase ( MAPKKK ) . After P . aeruginosa PA14 infection , RNAi knockdown of daf-16 or dbl-1 gene did not significantly affect the survival in transgenic nematodes overexpressing intestinal sdz-24; however , RNAi knockdown of pmk-1 , sek-1 , or nsy-1 gene significantly suppressed the survival in transgenic nematodes overexpressing intestinal sdz-24 ( Fig 8A and 8B ) , suggesting that RNAi knockdown of genes encoding the p38 MAPK signaling pathway may potentially inhibit the resistant property of transgenic nematodes overexpressing intestinal sdz-24 to P . aeruginosa PA14 infection . RNAi knockdown of pmk-1 , sek-1 , or nsy-1 gene also significantly increased the CFU of P . aeruginosa PA14 in the body of P . aeruginosa PA14 infected transgenic nematodes overexpressing intestinal sdz-24 ( Fig 8C ) . Moreover , RNAi knockdown of pmk-1 , sek-1 , or nsy-1 gene significantly decreased the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in P . aeruginosa PA14 infected transgenic nematodes overexpressing intestinal sdz-24 ( Fig 8D ) . To confirm the interaction between SDZ-24 and p38 MAPK signaling pathway in the regulation of innate immunity to P . aeruginosa PA14 infection , we further examined the effect of sdz-24 mutation on expression of some immune effectors ( T24B8 . 5 , F08G5 . 6 , and F35E12 . 5 ) for the p38 MAPK signaling pathway [29–30] . However , we found that mutation of sdz-24 did not significantly affect the expression of these immune effectors for the p38 MAPK signaling pathway ( S7A Fig ) . Moreover , mutation of sdz-24 also did not significantly affect the expression of phosphorylated PMK-1 compared with wild-type ( S7B Fig ) . Therefore , SDZ-24 may actually do not act upstream of the p38 MAPK signaling pathway to regulate innate immunity in P . aeruginosa PA14 infected nematodes . Considering the fact that SDZ-24 is a SKN-1-dependent protein , we next investigated the genetic interaction between SDZ-24 and SKN-1 in the regulation of innate immune response to P . aeruginosa PA14 infection . After P . aeruginosa PA14 infection , RNAi knockdown of skn-1 gene significantly reduced the survival in transgenic nematodes overexpressing intestinal sdz-24 ( Fig 9A ) , suggesting that RNAi knockdown of skn-1 gene may suppress the resistant property of nematodes overexpressing intestinal sdz-24 to P . aeruginosa PA14 infection . RNAi knockdown of skn-1 gene further significantly increased the CFU of P . aeruginosa PA14 in the body of infected nematodes overexpressing intestinal sdz-24 ( Fig 9B ) . Moreover , RNAi knockdown of skn-1 gene significantly inhibited the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in P . aeruginosa PA14 infected nematodes overexpressing intestinal sdz-24 ( Fig 9C ) . Therefore , the function of SDZ-24 in the regulation of innate immunity may be dependent on the SKN-1 . Previous study has demonstrated that SKN-1 is involved in the control of innate immunity [22] . We further found that , after P . aeruginosa PA14 infection , intestinal RNAi of skn-1 gene significantly decreased the survival , enhanced the CFU of P . aeruginosa PA14 in the body , and suppressed the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) compared with VP303 ( Fig 9D–9F ) , suggesting that skn-1 gene can act in the intestine to regulate the innate immune response to P . aeruginosa PA14 infection . The transcriptional factor of SKN-1/Nrf usually acts downstream of DAF-2 , the insulin receptor in the insulin signaling pathway , to regulate the biological processes such as longevity [31] . We further asked whether SDZ-24 can function through the insulin signaling to regulate the innate immunity . We found that mutation of sdz-24 gene significantly inhibited the resistant property of daf-2 ( e1370 ) mutant to P . aeruginosa PA14 infection ( Fig 10A ) . Mutation of sdz-24 gene also significantly increased the CFU of P . aeruginosa PA14 in the body of P . aeruginosa PA14 infected daf-2 ( e1370 ) mutant nematodes ( Fig 10B ) . Moreover , mutation of sdz-24 gene significantly suppressed the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in P . aeruginosa PA14 infected daf-2 ( e1370 ) mutant nematodes ( Fig 10C ) . Therefore , SDZ-24 may act downstream of DAF-2 to regulate the innate immunity .
In C . elegans , the miRNA of let-7 may have multiple biological functions by suppressing the functions of different targets . It has been shown that let-7 is involved in the control of developmental timing [18–19] . In this study , we observed that P . aeruginosa PA14 infection could significantly decrease the expression of let-7::GFP ( Fig 1 ) . Meanwhile , the loss-of-function mutation of let-7 could induce a resistant property to P . aeruginosa PA14 infection , suppress the P . aeruginosa PA14 colonization , and lead to an elevated innate immune response ( Fig 1B and 1C ) [20] . Therefore , let-7 may encode a protection mechanism for nematodes against the toxic effects of P . aeruginosa PA14 infection . An emerging body of data has demonstrated the immune response induction as a source of cellular stress for nematodes [32–33] . Our data suggest the important physiological role of let-7 in suppressing the immune response genes . Previous study has implied that LIN-41 and HBL-1 , two proteins required for the control of developmental timing , may act as the potential targets for let-7 in the regulation of innate immunity [20] . In this study , we further identified the other potential targets for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection . Based on the information from TargetScan prediction and dysregulated mRNA profiling induced by P . aeruginosa PA14 infection , we identified 7 candidate molecular targets for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection ( Fig 3A ) . Among these 7 candidate targets , SDZ-24 was identified as a direct target for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection based on the assays of survival , CFU of P . aeruginosa PA14 , expression pattern of antimicrobial genes , genetic interaction , and in vivo sdz-24 3’-UTR analysis ( Figs 3B–3D , 4 and 7 ) . These results imply that let-7 could directly target both the developmental timing proteins and the non-developmental timing proteins , such as the SDZ-24 , to regulate the innate immune response . In C . elegans , mutation of sdz-24 did not affect the longevity ( S8 Fig ) . Nevertheless , we did not found that P . aeruginosa PA14 infection could significantly alter the expression levels of lin-41 and hbl-1 [25] . This implies that the altered expression of let-7 may be not enough to affect the expression of its targets of LIN-41 and HBL-1 in nematodes infected with P . aeruginosa PA14 . Or , besides the let-7 , the function of LIN-41 and HBL-1 in the regulation of innate immunity may be also under the control of other miRNAs or signaling pathways . In this study , we also found NHR-43 did not act as a direct target for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection ( Fig 4A ) , although the nhr-43 gene was involved in the control of innate immunity ( Fig 3B–3D ) , implying that other still unidentified miRNAs may target the NHR-43 to regulate the innate immune response to P . aeruginosa PA14 infection . The tissue-specific activity assay indicated that let-7 could function in intestine to regulate the innate immune response to P . aeruginosa PA14 infection . The expression of let-7 in the intestine could reduce the survival , increase the CFU of P . aeruginosa PA14 , and decrease the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) in let-7 ( mg279 ) after P . aeruginosa PA14 infection ( Fig 2 ) . Overexpression of let-7 in the intestine could reduce the survival , increase the CFU of P . aeruginosa PA14 , and decrease the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) in P . aeruginosa PA14 infected wild-type nematodes ( S3 Fig ) . In this study , we provide several lines of evidence to prove that SDZ-24 may act as a direct intestinal target for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection . Firstly , after P . aeruginosa PA14 infection , loss-of-function of let-7 significantly increased the expression of intestinal SDZ-24::GFP ( S5 Fig ) . Secondly , intestinal RNAi of sdz-24 gene could reduce the survival , increased the CFU of P . aeruginosa PA14 , and suppressed the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) after P . aeruginosa PA14 infection ( Fig 5A–5C ) . Thirdly , after P . aeruginosa PA14 infection , overexpression of sdz-24 gene in the intestine caused significant increase in the survival , decrease in the CFU of P . aeruginosa PA14 , and enhancement of the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) ( Fig 5D–5F ) . More importantly , we found that the intestinal overexpression of sdz-24 lacking 3’-UTR could effectively suppress the susceptible property of nematodes overexpressing intestinal let-7 to P . aeruginosa PA14 infection ( Fig 6A–6C ) . Different from this , we could observe the effects of intestinal expression of let-7 on innate immunity in P . aeruginosa PA14 infected transgenic strain overexpressing sdz-24 containing its 3’-UTR ( Fig 6D–6F ) . Moreover , our tissue-specific activity assay also indicated that let-7 may further act in the neurons to regulate the innate immune response to P . aeruginosa PA14 infection . The expression of let-7 in the neurons reduced the survival , increased the CFU of P . aeruginosa PA14 , and decreased the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) in let-7 ( mg279 ) after P . aeruginosa PA14 infection ( Fig 2 ) . Additionally , overexpression of the let-7 in the neurons also reduced the survival , increased the CFU of P . aeruginosa PA14 , and decreased the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , F55G11 . 7 , and F55G11 . 4 ) in P . aeruginosa PA14 infect wild-type nematodes ( S3 Fig ) . These results imply that let-7 may further mediate certain neuronal signaling pathway ( s ) to regulate the innate immunity . However , so far , we still do not know the possible direct neuronal target for let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection . For the intestinal SDZ-24-mediated signaling pathways , we found that the intestinal SDZ-24 may act upstream of SKN-1/Nrf to regulate the innate immunity , because RNAi knockdown the skn-1 gene could significantly decrease the survival , increase the CFU of P . aeruginosa PA14 , and inhibit the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in nematodes overexpressing intestinal sdz-24 ( Fig 9A–9C ) . It was reported that P . aeruginosa PA14 infection led to the intestinal SKN-1 accumulation [22] . In C . elegans , mir-84 and mir-241 , the other members in let-7 family , could even regulate the innate immune response to P . aeruginosa PA14 infection by directly targeting SKN-1 [15] . It was reported that PMK-1 in the p38 MAPK signaling pathway acted upstream of SDZ-24 in the regulation of innate immunity [20] . In addition , SKN-1 could act downstream of p38 MAPK signaling pathway in the regulation of innate immunity [15] . These results imply the existence of signaling cascade of PMK-1-SDZ-24-SKN-1 in the intestine in the regulation of innate immunity . Meanwhile , after P . aeruginosa PA14 infection , intestinal RNAi knockdown of skn-1 did not significantly affect the sdz-24 expression ( S9 Fig ) . In C . elegans , sdz-24 encodes an ortholog of human RPA1 , and may have nucleic acid binding activity based on the protein domain information [21 , 34] . However , it is still unclear whether sdz-24 can potentially have the nucleic acid binding activity with its targeted genes for the reason of lack of direct evidence . In this study , we further found that intestinal SDZ-24 could act downstream of DAF-2 in the insulin signaling pathway to regulate the innate immunity , since RNAi knockdown the sdz-24 gene reduced the survival , increased the CFU of P . aeruginosa PA14 , and suppressed the expression of antimicrobial genes ( lys-1 , dod-22 , K08D8 . 5 , and F55G11 . 7 ) in daf-2 ( e1370 ) mutant nematodes ( Fig 10 ) . In C . elegans , both the daf-2 gene and the sdz-24 gene can act in the intestine to regulate the innate immunity ( Fig 5 ) [8] . Therefore , in the intestine , the DAF-2-SDZ-24-SKN-1 signaling cascade may be also formed to be involved in the control of innate immune response to P . aeruginosa PA14 infection . In conclusion , in this study , we examined the tissue-specific activity of let-7 in the regulation of innate immune response to P . aeruginosa PA14 infection . We found that let-7 could function in both the intestine and the neurons to regulate the innate immunity . During the control of innate immune response to P . aeruginosa PA14 infection , we identified the SDZ-24 protein as a direct target for let-7 in the intestine . let-7 could regulate the innate immune response to P . aeruginosa PA14 infection by suppressing both the expression and the function of intestinal SDZ-24 . Moreover , we raised the SDZ-24-mediated signaling cascades potentially formed in the intestine for nematodes against the P . aeruginosa PA14 infection . Our results provide an important molecular basis for the intestinal let-7 in the regulation of innate immunity . Our results will further highlight the crucial role of intestinal miRNAs for animals against the pathogen infection .
Nematodes strains used in the present study were wild-type N2 , mutants of let-7 ( mg279 ) , eat-2 ( ad465 ) , sdz-24 ( yd101 ) , sid-1 ( qt9 ) , tag-38 ( tm470 ) , nhr-43 ( tm1381 ) , mtl-1 ( tm1770 ) , nex-4 ( gk102 ) , daf-2 ( e1370 ) , Y95B8A . 6 ( tm5312 ) , nhr-43 ( tm1381 ) ;let-7 ( mg279 ) , sdz-24 ( dy101 ) ;let-7 ( mg279 ) , and sdz-24 ( dy101 ) ;daf-2 ( e1370 ) , and transgenic strains of zaEx5[let-7::GFP] , let-7 ( mg279 ) Ex ( Pges-1-let-7 ) , let-7 ( mg279 ) Ex ( Punc-14-let-7 ) , let-7 ( mg279 ) Ex ( Pmyo-2-let-7 ) , let-7 ( mg279 ) Ex ( Pmyo-3-let-7 ) , let-7 ( mg279 ) Ex ( Pdpy-7-let-7 ) , Is ( Pges-1-let-7 ) , Is ( Punc-14-let-7 ) , Ex ( Psdz-24-sdz-24::GFP ) , Ex ( Psdz-24-sdz-24::GFP ) ;let-7 ( mg279 ) , VP303/kbIs7[nhx-2p::rde-1] , Ex ( Pges-1-sdz-24 ) , Ex ( Pges-1-sdz-24 ) ;Is ( Pges-1-let-7 ) , Ex ( Pges-1-sdz-24-3’-UTR ) , Ex ( Pges-1-sdz-24-3’-UTR ) ;Is ( Pges-1-let-7 ) , daf-16 ( RNAi ) ;Ex ( Pges-1-sdz-24 ) , Ex ( Pges-1-sdz-24 ) ;dbl-1 ( RNAi ) , Ex ( Pges-1-sdz-24 ) ;pmk-1 ( RNAi ) , Ex ( Pges-1-sdz-24 ) ;sek-1 ( RNAi ) , Ex ( Pges-1-sdz-24 ) nsy-1 ( RNAi ) , and Ex ( Pges-1sdz-24 ) ;skn-1 ( RNAi ) . Nematodes were basically maintained on nematode growth medium ( NGM ) plates seeded with E . coli OP50 at 20°C as described [35] . Age synchronous populations of L4-larvae were prepared as described [36] . The L4-larvae were infected with P . aeruginosa PA14 as described [37] . P . aeruginosa PA14 was cultured in Luria broth , and seeded on killing plates containing a modified NGM ( 0 . 35% instead of 0 . 25% peptone ) . P . aeruginosa PA14 was incubated first at 37°C for 24-h and then at 25°C for 24-h . The P . aeruginosa PA14 infection was started by adding 60 animals to each plate at 25°C . Full-lawn PA14 killing plates were prepared for P . aeruginosa PA14 infection . Survival assay was performed basically as described [38] . During the P . aeruginosa PA14 infection , nematodes were scored for dead or live every 12-h . Nematodes would be scored as dead if no response was detected after prodding with a platinum wire . The hermaphrodite nematodes were transferred daily at 25°C for the first 5 days of adulthood . For survival assay , three replicates were performed . For the survival assay , the killing plates contained the modified NGM , as well as of fluoro-29-deoxyuridine ( FUdR , 75 μ g/mL ) to prevent the growth of progeny . Survival curve data were statistically analyzed using the log-rank test . The survival curves were considered to be significantly different from the control when the p-values were less than 0 . 01 . CFU of P . aeruginosa PA14 was analyzed as described [39] . After P . aeruginosa PA14 infection for 24-h , nematodes were transferred into 25 mM levamisole to paralyze nematodes and to stop the pharyngeal pumping . Nematodes were then placed on a NGM plate containing ampicillin ( 1 mg/mL ) and gentamicin ( 1 mg/mL ) for 15-min in order to eliminate the P . aeruginosa PA14 stuck to the body of animals . Nematodes were transferred onto a new NGM plate containing ampicillin ( 1 mg/mL ) and gentamicin ( 1 mg/mL ) for 30-min to further eliminate the external P . aeruginosa PA14 . After that , the nematodes were lysed with a motorized pestle . The lysates were serially diluted with M9 buffer , and plated on Luria-Bertani plates containing rifampicin ( 100 μg/mL ) for the selection of P . aeruginosa PA14 . After incubation at 37°C overnight , colonies of P . aeruginosa PA14 were counted to determine the CFU per nematode . Ten nematodes were examined per treatment , and six replicates were performed . The nematodes were infected with P . aeruginosa PA14 for 24-h . Total nematode RNA ( ~ 1 μg ) was extracted using RNeasy Mini kit ( Qiagen ) , and reverse-transcribed using a cDNA synthesis kit ( Bio-Rad Laboratories ) . qRT-PCR was performed at the optimized annealing temperature of 58°C . The examined targeted genes were lys-1 , lys-8 , clec-85 , dod-22 , K08D8 . 5 , F55G11 . 7 , F55G11 . 4 , T24B8 . 5 , F08G5 . 6 , and F35E12 . 5 . Relative quantification of targeted genes in comparison to reference tba-1 gene , encoding a tubulin , was determined . The final results were expressed as the relative expression ratio between the targeted gene and the reference tba-1 gene . The designed primers for targeted genes and reference tba-1 gene were shown in S1 Table . Six replicates were performed . The brood size was analyzed as described [40] . To assay brood size , the number of offspring at all stages beyond the egg was counted . Ten nematodes were used for each reproduction assay , and three replicates were performed . The corresponding targeted genes for let-7 were firstly predicted using the TargetScan software by searching for the presence of conserved sites that match the seed region of let-7 ( version 6 . 2 , http://www . targetscan . org/worm_52/ ) . This prediction results were confirmed by mirBase , PicTar , and miRanda assays . After that , we further screened whether the predicted targeted genes of let-7 could also be dysregulated by P . aeruginosa PA14 infection [25] . To generate entry vector carrying promoter sequence , ges-1 promoter used for the intestine-specific expression , unc-14 promoter used for the neuron-specific expression , myo-2 promoter used for the pharynx-specific expression , myo-3 promoter used for the muscle-specific expression , or dpy-7 promoter used for the hypodermis-specific expression was amplified by PCR from C . elegans genomic DNA . The ges-1 , unc-14 , myo-2 , myo-3 , or dpy-7 promoter was inserted into pPD95_77 vector in the sense orientation . The let-7 , or sdz-24/K07E8 . 3 . 1 cDNA lacking 3’-UTR or containing 3’-UTR was amplified by PCR , verified by sequencing , and inserted into the corresponding entry vector behind the ges-1 , unc-14 , myo-2 , myo-3 , or dpy-7 promoter . Transgenic nematodes were generated as described by coinjecting the testing DNA at a concentration of 10–40 μg/mL and a marker DNA ( Pdop-1::rfp ) at a concentration of 60 μg/mL into the gonad of nematodes [41] . The designed primers for DNA construct generation were shown in S2 Table . The method was performed as described [26] . To generate deletion allele of sdz-24 , we targeted the fragment corresponding to the first exon . To clone the sequence with the target sites into the sgRNA expression vector , we designed the primers of Guide F ( 5’-CCCTATTCCATCACTCACTC-3’ ) , and Guide R ( 5’-TCCTGCTCACCGACTCGTT-3’ ) . To target Cas9 to the genomic sequence of sdz-24 , we inserted the desired targeting sequence into the Cas9-sgRNA construct ( pDD162 ) using the Q5 site-directed mutagenesis kit ( New England BioLabs ) using the forward primer ( 5′-N19GTTTTAGAGCTAGAAATAGCAAGT-3′ ) and the reverse primer ( 5′-CAAGACATCTCGCAATAGG-3′ ) . For this deletion construction , we injected 20 animals with a mixture containing 5 ng/ml Pmyo-3::mCherry ( pCFJ104 ) , and 50 ng/ml of each of the two sgRNAs using standard microinjection procedures as described [41] . From transgenic F1 animals expressing mCherry , a region surrounding the target site of sdz-24 was PCR amplified using primers of sdz-24 PL ( 5’- CACTTTCACAAATGCTCCGCCTA-3’ ) , and sdz-24 PR ( 5’- GGCCTGTGCGATTTGGATATCTT-3’ ) in order to confirm the knockout animals . The homozygous mutant lines were established by isolating single F2 nematodes and determining their genotype by PCR and sequence analysis . The 3’-UTR ( wt ) of sdz-24 was PCR amplified from genomic DNA . A sdz-24 3’-UTR ( mut ) reporter was constructed by replacing putative let-7 binding site with an oligonucleotide containing the exact complementary sequence of let-7 . The synthesized sdz-24 3’-UTR ( mut ) sequence is ATAGAATTCTTTGCCGTGTGTAACCGAATGGCTCAATAAAGAGGGAAAGTGTCCAACATGCCGCAAGTTGCTTCTCAACCCCGGGATGATGGAGTATTGAATTAAATTTATAATATTTTTAGTGTCTCAAGTTTGTATTTTGAATGTATGAGAATATTTCGAAAAAATTATATCATGAATTACTTTCTTCTTATAACCTGGAACACAACAAAGGGGCCCTAT . The 3’ UTR reporter construct ( Pges-1::GFP-3’ UTR ( sdz-24 wt ) or Pges-1::GFP-3’ UTR ( sdz-24 mut ) ) and mCherry internal control ( Pges-1::mCherry-3’ UTR ( tag-192 ) ) plasmid were coinjected into the gonad of nematodes as described [41] . The expression of GFP and mCherry was observed and analyzed under a fluorescence microscope ( Olympus BX41 , Olympus Corporation , Japan ) . The designed primers for related DNA construct generation were shown in S2 Table . RNAi was performed by feeding nematodes with E . coli strain HT115 ( DE3 ) expressing double-stranded RNA that is homologous to a target of sdz-24 , K01A2 . 10 , daf-16 , dbl-1 , pmk-1 , sek-1 , nsy-1 , or skn-1 gene as described [42] . E . coli HT115 ( DE3 ) was first grown in LB broth containing ampicillin ( 100 μg/mL ) at 37°C overnight , and then plated onto NGM plants containing ampicillin ( 100 μg/mL ) and isopropyl 1-thio-β-D-galactopyranoside ( IPTG , 5 mM ) . L2 larvae of certain strain were transferred onto RNAi plates for 2 days at 20°C until they developed into the gravid . The gravid adults were further transferred onto the fresh RNAi-expressing bacterial lawns to let them lay eggs for 2 h to obtain the second generation of RNAi population . The eggs were allowed to develop at 20°C to young adults for the subsequent survival , CFU of P . aeruginosa PA14 , and gene expression pattern assays . The method was performed as described previously [43] . Nematode protein was extracted and electrophoresed on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( SDS-PAGE ) gel . The gel was transferred to a nitrocellulose membrane using a Bio-Rad semi-dry transfer apparatus . After pre-incubation with 5% nonfat milk in TBST buffer ( 10 mM Tris , pH 8 . 0 , 150 mM NaCl and 0 . 5% Tween 20 ) for 30 min , the membrane was incubated with a primary antibody in TBST buffer with 5% nonfat milk for 12 h at 4°C . After washing with the TBST buffer , the membrane was further incubated with a horseradish peroxidase ( HRP ) -conjugated secondary antibody for 1 . 5 h . The membrane was then developed with ECL system ( Thermo Scientific ) . Anti-phospho-p38 MAPK monoclonal antibody 28B10 ( 1:500 ) was from Cell Signaling , and anti-Actin monoclonal antibody MAB1501 ( 1:5000 ) was from EMD Millipore . The goat anti-mouse IgG antibody ( H&L ) [HRP] ( 1:10 000 ) was from GenScript . Three replicates were performed . All data in this article were expressed as means ± standard deviation ( SD ) . Graphs were generated using Microsoft Excel ( Microsoft Corp . , Redmond , WA ) . Statistical analysis was performed using SPSS 12 . 0 ( SPSS Inc . , Chicago , USA ) . Differences between groups were determined using analysis of variance ( ANOVA ) . Probability levels of 0 . 05 and 0 . 01 were considered statistically significant . | Some microRNAs ( miRNAs ) have been identified recently to play important roles in the regulation of innate immunity in Caenorhabditis elegans . let-7 is one of important miRNAs identified to be involved in the control of innate immune response . However , the underlying molecular mechanism for let-7 in the regulation of innate immune response is still largely unknown . In C . elegans , let-7 could function in both the intestine and the neurons to regulate the innate immunity . We here focused on the examination of molecular basis for the intestinal let-7 in the regulation of innate immune response to Pseudomonas aeruginosa PA14 infection . We identified SDZ-24 , an ortholog of human replication protein A1 , as a direct target for intestinal let-7 in the regulation of innate immune response . For the molecular mechanisms of intestinal let-7 in the regulation of innate immunity , let-7 might negatively regulate the function of SKN-1 by suppressing the expression and function of its target of SDZ-24 . Our results imply the important function of intestinal miRNAs , such as let-7 , in the regulation of innate immune response to pathogenic infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"pathogens",
"immunology",
"microbiology",
"parasitic",
"diseases",
"pseudomonas",
"aeruginosa",
"nematode",
"infections",
"epigenetics",
"bacteria",
"bacterial",
"pathogens",
"digestive",
"system",
"pseudomonas",
"genetic",
"interference",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"immune",
"response",
"immune",
"system",
"gastrointestinal",
"tract",
"biochemistry",
"rna",
"anatomy",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms"
]
| 2017 | Molecular Control of Innate Immune Response to Pseudomonas aeruginosa Infection by Intestinal let-7 in Caenorhabditis elegans |
Little is known of how gene expression and its plasticity evolves as populations adapt to different environmental regimes . Expression is expected to evolve adaptively in all populations but only those populations experiencing environmental heterogeneity are expected to show adaptive evolution of plasticity . We measured the transcriptome in a cadmium-enriched diet and a salt-enriched diet for experimental populations of Drosophila melanogaster that evolved for ~130 generations in one of four selective regimes: two constant regimes maintained in either cadmium or salt diets and two heterogeneous regimes that varied either temporally or spatially between the two diets . For populations evolving in constant regimes , we find a strong signature of counter-gradient evolution; the evolved expression differences between populations adapted to alternative diets is opposite to the plastic response of the ancestral population that is naïve to both diets . Based on expression patterns in the ancestral populations , we identify a set of genes for which we predict selection in heterogeneous regimes to result in increases in plasticity and we find the expected pattern . In contrast , a set of genes where we predicted reduced plasticity did not follow expectation . Nonetheless , both gene sets showed a pattern consistent with adaptive expression evolution in heterogeneous regimes , highlighting the difference between observing “optimal” plasticity and improvements in environment-specific expression . Looking across all genes , there is evidence in all regimes of differences in biased allele expression across environments ( “allelic plasticity” ) and this is more common among genes with plasticity in total expression .
Phenotypic plasticity is the phenomenon of one genotype producing different phenotypes ( e . g . , physical or behavioral ) when exposed to different environments . To produce different phenotypes from a single genotype , one of the essential intermediate steps is the induction of gene expression changes by the environment [1 , 2] . Studying expression plasticity can provide insights on how different phenotypes are generated by the interactions between genotype and the environment [3] . Moreover , transcriptomics allow us to examine plastic responses for a large set of traits ( expression levels of many genes ) that are relatively unbiased compared to traditional phenotypic traits with respect to preconceived notions of their importance or ease of measurement , though the link between expression traits and ecological importance is typically more tenuous than for traditional phenotypes . Plasticity may be beneficial , allowing organisms developing in different environments to produce phenotypes better suited to those environments . On the other hand , a plastic response can be deleterious if it shifts the phenotype away from the optimum for the organism , perhaps reflecting the inability of the organism to buffer against the imposed environmental perturbation [4 , 5] . Finally , a plastic phenotypic change can be neutral ( or nearly so ) , e . g . , a by-product of physiological response to the environment that has little effect on fitness [6] . A variety of transcriptional changes may occur as a result of exposure to a novel environment . Multigenerational selection in the new environment may engender genetic responses that reinforce initially beneficial plastic changes ( e . g . , genetic assimilation [5] ) . Alternatively , long-term selection may result in genetic responses that oppose deleterious plastic responses that were deleterious , resulting in a pattern of “counter-gradient” variation [7] . Plasticity can only be shaped adaptively in populations that evolve in a variable environment . A naïve population first exposed to a variable environment may initially exhibit beneficial plasticity with respect to some expression traits but deleterious plasticity for others . Subsequent evolution in a variable environment is expected to reshape plastic responses . For most traits , including expression traits , selection does not act directly on plasticity itself but rather plasticity evolves as a by-product of adaptation of trait means to each encountered environment [6] . How plasticity evolves depends on how the phenotypes initially produced in a novel environment differ from the optimal phenotype in each environment . Both decreases and increases in phenotypic plasticity could contribute to adaptation to variable environments [8] . Further , alternative forms of heterogeneity ( e . g . , temporal vs . spatial ) may select on plasticity differently [9 , 10] . Although expression plasticity could be beneficial [11 , 12] or deleterious [13] and genetic variation for plasticity has been found in different organisms [14–17] , we still have little understanding of how expression plasticity evolves , in terms of the rate and the directions . Can plasticity evolve adaptively on short time scales ? Yampolsky et al . [18] used microarrays to survey the transcriptome of Drosophila populations maintained in homogeneous environments ( regular or ethanol medium ) or spatially heterogeneous environments ( mixed of two types of mediums ) for more than 300 generations but found that the selective regime had limited effect on expression plasticity for the two mediums . They suggested that evolution of expression plasticity might require a longer timescale . Here we examine expression plasticity in experimental Drosophila melanogaster populations that have evolved under constant conditions or with either spatial or temporal heterogeneity in larval diets . We have previously used these populations to examine how environmental heterogeneity affects inbreeding depression [19] , genome-wide molecular diversity [20] , quantitative genetic variation [21] and adaptive potential [22] . Here we use these populations , after ~130 generations of evolution , to study expression plasticity in larvae . We address three types of questions:
We first examine expression differences between populations that evolved under alternative constant conditions by contrasting the samples from the cadmium-selected ( Cad ) and salt-selected ( Salt ) populations . 546 genes show a “selection history” effect between Cad and Salt populations ( FDR ( q ) < 0 . 1; we use a liberal q-value because we are interested in the properties of the genes in the list—tested in downstream analyses—rather than the genes themselves ) . A significant “selective history” effect can be loosely interpreted as an evolved difference in the level of expression averaged across diets . Previously , we examined allele frequency divergence between the Cad and Salt populations [20] . Combining those results with expression divergence , we tested whether the set of genes with significantly divergent SNP frequencies between the cadmium-selected and the salt-selected populations are enriched for genes showing divergent expression between Cad and Salt . We separated the genes based on whether the SNPs are located in exon , intron or intergenic regions . There is significant enrichment for differential expression for genes with differentiated SNPs located within 1kb outside the genic regions ( χ2 = 8 . 0 , df = 1 , p = 0 . 0047 for genes with SNPs 1kb around the gene ) , but not for genes with SNPs located in exons ( χ2 = 2 . 3 , df = 1 , p = 0 . 12 ) or introns ( χ2 = 1 . 7 , df = 1 , p = 0 . 19 ) . These results suggest that cis-acting factors contribute to the evolved divergence in expression ( but these results do not exclude an important role for trans-acting factors as well ) . This is consistent with previous studies showing cis-regulatory variants are important for expression divergence between populations or species in Drosophila [23–25] . Next , we examine whether the direction of expression changes in the ancestral population when exposed to cadmium- versus salt-enriched food match evolved expression differences between cadmium- and salt-selected populations ( i . e . , do genes that are up-regulated by cadmium exposure relative to salt exposure in the ancestor evolve higher or lower expression in cadmium-selected populations than salt-selected populations ? ) . Using the Grand Ancestor ( GA ) population , which is naïve to both diets , we identify 905 genes that have log2 fold change ( log2FC ) in expression between cadmium and salt diets greater than 0 . 4 . ( This plasticity reflects a change in expression then the ancestor is reared on cadmium versus salt; it is not a contrast with expression on benign cornmeal . ) Using samples from the five Cad and the five Salt populations , there are 546 genes showing a “selection history” effect with q < 0 . 1 . ( The “selection history” history reflects differentiation between Cad and Salt regimes from one another , not necessarily differentiation from the Grand Ancestor . ) A hundred and eight genes overlap between the two sets of genes ( gene names are listed in S7 Table ) . The functional categories “Membrane” and “Transmembrane” are significantly enriched among these 108 genes ( ~40 of the 108 genes are related to membranes ) . For 91% of these genes , the evolved divergence is in the opposite direction to the naïve plastic response ( 98/108; significantly different than 50%: χ2 = 71 . 7 , df = 1 , p < 2 . 2e-16; Fig 2A ) , meaning , for example , that a gene which is up-regulated when the ancestor is reared in cadmium ( relative to salt ) has evolved lower average expression in the Cad than in the Salt populations . Using Cad/Salt population pairs ( rather than genes ) as the unit of replication confirms this result where evolved responses oppose rather than reinforce plastic ones ( Fig 2B ) . This type of “counter-gradient” response [7] is emerging as a major theme in evolutionary expression studies [26–29] . There are at least two reasons that a countergradient pattern could occur . If selection favours the same expression level ( or phenotype ) across environments but one environment induces a change in the phenotype , then opposing genetic changes are expected to evolve . Heuristically , genetic and environmental effects of opposite sign combine additively to yield little net change in phenotype/expression when genotypes are assayed in their adaptive environment . This type of explanation has been used for the countergradient pattern observed for the rate of tadpole development along an altitudinal gradient [30 , 31] . A different reason for the appearance of a countergradient pattern is related to the stress experienced by a population exposed to a novel environment . Abnormal expression responses occur because of the direct perturbation by the novel environment or reflect the stress response ( or its cascading effects ) employed to cope with environmental perturbation . Adaptation may render the environment benign ( for example , via the evolution of efficient cadmium detoxifaction ) such that there is no longer a direct perturbation or a stress response . In this case , genetic and environmental effects are not additive as each population is affected differently by the environment . A closer inspection of our data indicates that the observed counter-gradient response does not appear to involve a genetic effect acting in opposition to but additively with the diet effect; rather Salt and Cad regimes are affected differently by diet ( S1 Fig ) . In this gene set , a large proportion of genes is up-regulated in the naïve ancestor ( GA ) when it is reared in cadmium compared to when it is reared in salt; a similar pattern of plasticity is observed in Salt populations . In contrast , this plasticity has largely been evolutionarily lost in Cad populations such that there is little up-regulation of these genes in cadmium . A likely scenario is that , after Cad populations adapt reasonably well to cadmium , they no longer show the perturbed response of a population naïve to cadmium . In addition to the two homogeneous selective regimes ( Cad and Salt ) , this experiment involved two regimes with heterogeneous selection ( Temp and Spatial ) . To visualize the expression divergence among experimental samples from all regimes , we plot the first five principle components of expression ( Fig 3 ) . The most striking pattern occurs with respect to PC2 . Salt populations reared in cadmium have a very different value for PC2 than all other treatment/diet combinations , reinforcing the idea that cadmium perturbs the expression profile of populations without an evolutionary history of cadmium exposure . From further inspection of Fig 3 , various effects of diet , regime , and their interaction are apparent; there are dimensions in which each regime differs from others either with respect to average expression across diets or plasticity between diets ( S1 Table ) . A functional category enrichment analysis for the genes that strongly load on to each PC dimension is shown in S2 Table . A summary of gene-level pairwise comparisons among all four selective treatments is given in the Supporting Information and S3 Table . A transcriptome-wide view of plasticity is given by the last column of Fig 3 but this perspective does not attempt to distinguish between genes showing plastic responses that are beneficial , deleterious , or selectively neutral . Though plasticity is expected to be more adaptive in the heterogeneous regimes than homogeneous regimes , we cannot test this prediction by looking at the transcriptome as a single unit . Rather , we need to first identify genes on which we expect selection to result in ( i ) increased or ( ii ) decreased plasticity . To do so , we leverage the history of our experimental treatments . The focal populations in all four selective regimes were originally created by crossing two diet-adapted populations , Ancestral Cadmium ( AC ) and Ancestral Salt ( AS ) , followed by ~130 generations of selection within each of the four treatments . Using the expression patterns of the two diet-adapted ancestral populations ( AC and AS ) , we screen for genes where we predict plasticity to evolve under heterogeneous environments and then examine levels of plasticity in these genes in our four experimental treatments . If optimal expression differs across environments , then , in the absence of constraints , we would expect populations evolving in heterogeneous environments to evolve adaptive plasticity as a by-product of selection to produce different expression patterns in each environment [6] . Though we cannot know the “optimal” expression level , we can use expression of each diet-adapted ancestor assayed in its respective diet as a first approximation of the optimum . To identify potential targets for the evolution of increased plasticity in heterogeneous regimes , we used the data from the two diet-adapted ancestors ( AC and AS ) . We screened for genes that met the two criteria ( see Methods for further details ) . First , we required a reasonably large difference ( |log2FC| > 0 . 4 ) in the “optimal” expression level for each diet ( given by AC in cadmium and AS in salt ) . Second , to exclude genes that are initially highly plastic , we required relatively low levels of plasticity within each ancestor ( the |log2FC| between diets for each of AC and AS is less than half the |log2FC| between AC and AS each in its own diet ) . This screen , based on the ancestral populations , identified 109 genes ( S7 Table ) ; no functional categories are significantly enriched for this set of genes . We now consider their plasticity in the four experimental treatments derived from these ancestors . For each gene in each population , we calculated the log2FC change across diets such that positive values indicate plasticity in the adaptive direction . For each population we averaged across the 109 genes to obtain a single measure of adaptive plasticity per population ( i . e . , “population” , not “gene” , is used as the unit of replication ) . The mean score for average adaptive plasticity is significantly greater than zero for both heterogeneous regimes ( t = 3 . 6 , p = 0 . 02 for Temp , t = 4 . 7 , p = 0 . 009 for Spatial; Fig 4A ) . In contrast , the score is close to zero in both homogeneous treatments ( p > 0 . 2 for each ) . A direct contrast of the heterogeneous versus the homogeneous treatments confirms the prediction that adaptive plasticity evolves to a greater extent in populations subject to variable environments ( χ2 = 14 . 3 , df = 3 , p = 0 . 0025 ) . The two alternative forms of heterogeneity ( Temp and Spatial ) appear to have very similar levels of adaptive plasticity . If optimal expression is similar in the two diets , then ideally there would be little or no plasticity . To identify potential targets for reduced plasticity in heterogeneous regimes , we again used the data from the diet-specific ancestors . We screened for genes meeting the following criteria ( see Methods for details ) . First , we required that optimal expression was similar in the two diets . Second , to exclude genes that begin with little plasticity , we required that both ancestors ( AC and AS ) had a plastic response to the other diet that was large relative to the difference between the two optima . ( Further , we required that both ancestors showed the same direction of plasticity between their adapted diet and their non-adapted diet; this requirement simplifies the interpretation of changes in plasticity . ) This screen could include genes where selection always favours the same level of expression but that are misregulated under the stress of a novel environment . Alternatively , or in addition , this screen could include genes that are part of a stress response that is activated upon exposure to a novel diet but not when consuming a diet to which a population is adapted . Using this ancestor-based screen , we obtained a set of 121 genes ( S7 Table ) ; no functional categories are significantly enriched for this set of genes . We now consider their plasticity in the four regimes . Because this gene set should ideally have little or no plasticity , we calculated the absolute value of expression change |log2FC| for each gene in each population , then averaged across the 121 genes to obtain a single value for each population . Though we expected to find lower values of plasticity in heterogeneous than homogeneous treatments , there was no evidence of this; Spatial had the lowest average plasticity and Temp had the highest average plasticity but there was no significant variation among treatments ( Fig 4B ) . Based on these results it seems that expression has not evolved as expected in the heterogeneous regimes , especially not in the Temp treatment ( but see below ) . Plasticity measures the change in expression across diets but does not measure how adaptive expression is in either environment . To do the latter , we created a metric Φ to represent the relative distance to the optimum for expression in diet d of gene i of population j: Φd , i , j=Ed , i , j−Od , iNd , i−Od , i where Od , i is the expression for the sample representing the “Optimal” state for diet d ( i . e . , AC in cadmium diet or AS in salt diet ) and Nd , i is the expression for the sample representing the “Non-adapted” state for diet d ( i . e . , AS in cadmium diet or AC in salt diet ) . When expression of a focal population is intermediate between values of the “Optimal” and “Non-adapted” states , the scaled distance to adaptive expression value is 0 ≤ Φd , i , j ≤ 1 , with 0 meaning the expression in the focal population is at the “optimal” expression and 1 meaning the focal population is as poor as the non-adapted ancestor . We first consider the set of 109 genes that we identified as potential targets to evolve increased ( adaptive ) plasticity . For each population , we calculated the average Φd over all the genes of interest for each diet separately . As expected for the constant regimes ( Cad and Salt ) , Φ¯ values are close to 0 ( optimal expression ) in their respective adapted diets but the Φ¯ are far above 0 in the alternative diets ( Fig 4C ) . Both Temp and Spatial regimes have Φ¯ values almost as low as the constant regime in its adapted diet , particularly in cadmium , providing evidence that heterogeneous populations are also relatively adapted to both diets for this set of genes . We next consider the set of 121 genes that we identified as potential targets to evolve reduced plasticity . As expected for the constant regimes ( Cad and Salt ) , Φ¯ is close to 0 ( optimal expression ) in their “native” diet but Φ¯ is far above 0 in the alternative diet , especially for the Salt regime in cadmium ( Fig 4D ) . The expression for Temp and Spatial regimes are close to optimal expression ( Φ¯ is close to 0 ) in both diets and significantly lower than the constant regime that is not adapted to that diet , suggesting that populations in heterogeneous populations are relatively well adapted to both diets with respect to expression for this set of genes . So far , we have focused on “abundance plasticity” , the difference in the total expression of a gene ( summing across alleles ) between diets . RNAseq provides us an opportunity to study another type of plasticity , “allelic plasticity” , which is the relative expression difference of two alleles for a polymorphic gene between diets ( i . e . , plasticity in allelic expression bias ) . This type of plasticity reflects variation in cis-regulatory elements whose effects are environmentally-dependent . Because each population is assayed in both environments , significant differences in SNP frequencies between environments in the RNAseq data reflects plasticity in allelic expression bias . In the context of pooled-seq data , allelic plasticity can be due to between-diet differences in expression between alternative homozygotes or because of between-diet differences in allele expression within heterozygotes . Alternatively , apparent allelic plasticity could be due to selection but there is little opportunity for this ( see Methods ) . For each regime , we screened for polymorphic sites and then selected the most informative site within each gene . We detect evidence of allelic plasticity at numerous genes ( ~7% of genes tested ) ; the average number of genes across the four regimes with significant ( at p < 0 . 01 ) allelic plasticity is 516 whereas the average number of genes expected by chance based on a permutation analysis is 49 . Moreover , allelic plasticity is approximately twice as common among those genes with significant abundance plasticity than those without ( Fig 5 ) . This pattern of enrichment would not be expected if a strong from of compensatory expression in which increased expression of one allele is balanced by reduced expression of the alternative allele in order to keep total expression reasonably constant . Rather , this enrichment likely exists because expression of one allele is substantially more sensitive to the environment but than the other , resulting in plasticity in total expression ( abundance plasticity ) as well as plasticity in the relative expression of the two alleles ( allelic plasticity ) . These patterns are likely driven by variation in cis-regulatory elements . There is no indication that this enrichment varies among treatments . We attempted to assess whether biased expression between diets is adaptive by examining the difference SNP frequencies in the RNAseq data between salt and cadmium assays for a given regime matched the direction of the difference in SNP frequencies from genomic data for salt- and cadmium-selected populations [20] . For the genes we identified as showing allelic plasticity ( p < 0 . 01 ) , the numbers showing significant allele frequency differentiation ( q < 0 . 001 ) are: Cad , 204; Salt , 167; Temp , 196; Spatial , 177 . However , for these genes we find no evidence , in any regime , that the allele favoured in a given environment is the more strongly expressed one . However , this method for detecting adaptive allelic expression bias is crude because it assumes that the favoured allele should be expressed more but in some cases reduced expression will be favoured .
This study demonstrates that alternative selective histories cause extensive divergence in gene expression and expression plasticity within 130 generations . We observe a striking pattern of countergradient variation with respect to expression divergence in a manner suggesting genetic responses evolve to restore an optimum perturbed by environmental effects . Though “countergradient” responses have been reported in several recent expression studies [26–29] , not all studies use the term in exactly the same manner . Classic usage of the term implicitly assumes plasticity is constant ( at least in sign ) across populations [7] but in a recent study [28] , and in our own , patterns of plasticity differ among populations . As suggested above , adaptation may often involve physiologically “managing” an environmental stimulus ( e . g . , cadmium detoxification ) so that it no longer creates a cascade of expression perturbations . If so , we would not expect that the genes showing countergradient responses to be the direct targets of adaptive evolution . Ghalambor et al . [28] argued that genes that have their expression maladaptively perturbed by a novel environment would be under strong selection to evolve genetic changes in expression . From this perspective , one might expect many of the genes with countergradient responses to be the direct targets of adaptive evolution . Future studies attempting to resolve these issues should strive to understand mechanistically ( i ) why plasticity occurs , ( ii ) what genetic changes underlie adaptation , and ( iii ) how these are responsible for changes in transcription . The expression perturbations experienced by a naïve population relative to an adapted one are not necessarily bad , and some may be beneficial in the short-term . Naïve populations may respond to novel environmental stimuli by altering gene expression in helpful but non-ideal ways that mitigate harmful environmental effects , e . g . , beneficial stress responses that are no longer needed ( and possibly harmful ) after better mechanisms of coping with the environment have evolved . Such genes could contribute to the gene set used in our countergradient analysis as well as the gene set predicted to evolve reduced plasticity , though there is no obvious indication of this from GO analysis . Regardless , these expression changes reflect the lack of a good solution to the environment even if some expression changes are beneficial initially . Thus , inclusion of such genes does not interfere with the objective of either analysis , though this possibility should be considered in interpreting the results . Specifically , we cannot assume that plastic responses that occur a naïve population but not in an adapted one are all necessarily deleterious in the naïve population . Expression studies like our own cannot identify which specific plastic changes are beneficial or deleterious in populations with different genetic backgrounds . In principle , selection on gene expression within a population could be studied using the same framework used to study selection on traditional phenotypes [32] though in practice this would be very difficult given the high dimensionality . Changes in transcriptome-wide expression plasticity can be difficult to interpret because selection that results in adaptive reaction norms will not occur in populations living in homogeneous environments . For populations experiencing heterogeneity , selection within each environment may ultimately result in increases or decreases in plasticity for different genes . Our approach has been to use the diet-adapted ancestors as a guide to optimal expression within each environment . Using this approach , we inferred the ideal level of plasticity in the absence of constraints and identified gene sets that we expected to evolve increased or reduced levels of plasticity . Focusing on these gene sets , we found that expression in heterogeneous regimes was more adaptive in each environment than that of the non-adapted homogeneous regime . For genes predicted to increase plasticity , we found , as expected , higher levels of adaptive plasticity in heterogeneous regimes than homogeneous regimes . However , we did not find a reduction in plasticity in heterogeneous regimes for genes predicted to reduce plasticity despite the evidence of adaptive levels of expression . Two reasons may contribute to the seeming discrepancy with respect to this latter gene set for which we see evidence of adaptive expression yet not the expected reduction in plasticity in heterogeneous regimes . First , any measurement error in expression ( in addition to true plasticity ) will contribute to our estimates of |log2FC| because we are using the absolute value of the difference in observed expression values between diets . Though such error should not artificially create differences in |log2FC| among regimes , it may reduce our power to detect true variation among regimes . Second , the observed results could arise simply by evolution toward optimal expression proceeding faster in one environment than the other within heterogeneous regimes ( S2 Fig ) . Consider a case where expression in both environments is initially higher than the optimum , which is the same in both environments . If expression levels evolve down toward the optimum in both environments but adaptation proceeds faster in one environment than the other , then this would result in increased plasticity in expression across environments despite improvement in both . The observation of adaptive expression without the expected reduction in plasticity serves as a reminder that selection does not truly act directly on plasticity for most types of traits [6] . If improvement is possible in only one environment , then this may result in increased plasticity ( at least transiently ) even if optimal expression is the same in both environments . Because the averaging across environments works differently with temporal and spatial heterogeneity [33] , plasticity may evolve differently with these alternatives forms of heterogeneity , though existing models make different predictions about the nature of this difference [9 , 10] . With respect to expression in the gene sets predicted to evolve increased and decreased plasticity , we see some hints of differentiation between the Spatial and Temp treatments ( Fig 4 ) but these are not statistically significant . Given that there are SNP frequency differences between the Spatial and Temp treatments [20] , a remaining challenge is to understand mechanistically how and evolutionarily why adaptation occurs differently with alternative forms of heterogeneity . When other studies investigate differences in plasticity in populations with spatial versus temporal heterogeneity , it will be possible to ask if there are general patterns in how plasticity evolves with these two common forms of heterogeneity . A fundamental question of expression evolution is the relative importance of cis and trans effects . Previous work has established several lines of evidence that cis effects are of considerable importance [23–25] . Two of our results add to this . First , we observed that genes differentially expressed between Cad and Salt regimes where enriched having significantly differentiated SNP frequencies located in nearby intergenic regions . Second , we found evidence of extensive diet-dependent differences in allelic bias and that this allelic plasticity is associated with abundance plasticity . The simplest explanation for these observations is environmentally-sensitive cis-acting factors . This observation , along with other recent studies [34 , 35] , raises the possibility that a substantial fraction of the genetic variation for expression may be manifest under particular environments ( i . e . , a large G×E component ) . Our view of expression plasticity in this study is limited in several respects . We have measured expression at only a single developmental stage ( very young larvae ) and patterns may differ at other stages . For example , we see evidence that expression of Salt larvae is strongly perturbed in cadmium but comparatively little expression perturbation of Cad larvae in salt . Because we know that egg to adult survivorship is low for both situations [21 , 22] , more severe expression perturbations are expected for Cad larvae in salt but these may not become apparent until later in development . Second , our analysis , like most expression studies , is biased towards detecting expression differences for genes that are reasonably highly expressed . Third , the pattern of plasticity evolves over time and may not yet have reached its equilibrium so measuring plasticity at multiple time points for evolving populations would be informative; this study represents only a single snapshot of expression evolution . Despite these limitations , various patterns are apparent and we have no a priori reason to believe these are misrepresentative , though other patterns might emerge with other types of expression data . While we have studied expression and its plasticity in well-controlled experimental populations , comparing expression plasticity for populations living in different natural habitats should generate insights into how plasticity facilitates adaptation on long timescales [26 , 27] . What aspects of the regulatory networks mediating plasticity evolve differently in short versus long evolutionary timescales ? Combining different approaches from laboratory experiments to field studies should yield a more comprehensive understanding of the selective forces and constraints on the evolution of plasticity .
A full description of the selective history of these populations can be found in Huang et al . 2014 [20] . Briefly a population collected from the wild was maintained in standard benign conditions ( approximately 2000–4000 adults ) , referred to here as the “Grand Ancestor” ( GA ) . Two subsets of flies from the GA population were used to initiate a population maintained in a cadmium-enriched medium and a population maintained in a salt-enriched medium , each with population size ~1000 , referred to as the “Ancestral Cadmium” ( AC ) population and the “Ancestral Salt” ( AS ) population , respectively . 448 males and 448 virgin females were collected from each the AC and AS populations and crossed with flies from the other population . The offspring from the next generation were used to found 20 populations ( each with 448 adults ) that were distributed evenly among four regimes: ( i ) constant cadmium-enriched ( CdCl2 ) medium ( “Cad” ) , ( ii ) constant salt-enriched ( NaCl ) medium ( “Salt” ) , ( iii ) alternating each generation between salt- and cadmium-enriched media ( “Temp” ) , and ( iv ) half the rearing vials containing cadmium-enriched medium and the other half containing salt-enriched medium ( “Spatial” ) . For the Spatial regime , an equal number of adult flies produced from each type of medium were mixed to produce offspring for the next generation ( i . e . , a “soft” selection regime sensu [36] ) . The Cad and Salt regimes show the expected patterns of local adaptation based on the different fitness assays measured at ~50 and ~130 generations , i . e . , populations from each regime perform better in their own diet than in the alternative [21 , 22] . The Temp and Spatial regimes have intermediate fitness between the two constant regimes in both diets . This suggests that the heterogeneously selected populations are relatively well adapted to both diets , though less adapted than the homogeneously selected populations in their own selective diets . Starting at ~127 generations of experimental evolution , samples were prepared in five blocks . Each block used one replicate population from each of the four regimes and assayed in both diets , resulting in eight samples ( perfectly balanced ) per block . Block 1 was performed at generation 127; Blocks 2 and 3 were performed at generation 128; Blocks 4 and 5 were performed at generation 129 . Samples from all three ancestral populations were prepared at generation 129 , regarded as Block 6 . One generation before the collection , each replicate experimental population was reared in regular cornmeal diet to control for the maternal environment . In next generation , the emerged adult flies ( Day 12 ) of each population mated for 1 . 5 days and then laid eggs on either salt or cadmium treatment diet . Before the actual egg-laying , pre-lay plates were supplied for two hours to provide ample opportunity for females to dump any stored eggs and ensure synchronization of the developmental stage of subsequently laid eggs . Laying plates were then supplied for two hours before the adults were removed . After ~ 20 hours ( at 25°C ) , any hatched larvae were removed and discarded . Newly hatched larvae were transferred to the same selective diet within a one-hour window . After 12 hours , 80 larvae were collected per sample in 1 . 5 ml tubes with PBS solution ( i . e . , all larvae were 12–13 hours old ) . The PBS solution was removed after centrifuging at 9500 rmp for 1 min . The samples were immediately frozen in dry ice and store in -80°C prior to RNA extraction . Total RNA was extracted using the NucleoSpin RNA Kit ( MACHEREY-NAGEL ) . Strand-specific single-end libraries were prepared and sequenced in six lanes of HiSeq2000 ( The McGill University and Génome Québec Innovation Centre ) . All eight samples from the same block were sequenced in the same lane except the sample of replicate population 2 of Temp in cadmium diet treatment because of an error at the sequencing centre . The six samples from the ancestral populations were sequenced in one lane . The single-end reads were mapped to the D . melanogaster transcriptome and genome ( FlyBase release version 5 . 41 ) using Tophat2 with library-type as fr-firststrand [37] . Only the alignment with the highest alignment score was used . If multiple alignments with the same score existed , one alignment of them was randomly retained . The aligned reads were sorted and viewed using samtools v . 0 . 1 . 16 [38] and then assigned to features of the transcriptome using HTSeq with default settings [39] . Because differences in coverage among treatments can result in different statistical power , we performed down-sampling of the mapped reads to obtain equivalent level of coverage across diet treatments and selective regimes . We first ranked the eight samples per experimental replicate block by read number . For each block i ( i ϵ {1–5} ) , we found the minimum coverage across the eight samples , ni . We sampled without replacement ni reads from the ith block for each sample . As a result , the numbers of mapped reads remaining per sample within each block were as follows: block 1 , 21707402; block 2 , 23714035; block 3 , 22732039; block 4 , 23378572 and block 5 , 20909145 . The numbers of useful reads for the six samples from the ancestral populations ( Grand Ancestor ( GA ) , Ancestral Cad ( AC ) and Ancestral Salt ( AS ) in both diets ) were in the range of 30 . 7 to 33 . 5 million . The gene expression counts were analyzed by the DESeq2 package [40] of the BioConductor suite [41] with empirical Bayes estimation . The expression counts for each gene were normalized as a quantity proportional to the concentration of cDNA from the gene in each sample and transformed to log2 scale [40] . To examine divergence between each pair of selective regimes , the transformed expression value for each gene of the 20 samples ( two regimes each with five replicate populations in two diets ) was analyzed by a generalized linear model with a logarithmic link: Expression ~ diet + selective history + diet × selective history + block . The selective history effect represents the difference between regimes in expression averaged across the diets . To highlight whether expression differs between regimes in one diet more than the other , we also examined expression separately for each diet: Expression ~ selective history + block . To examine levels of parallel plasticity among five replicates within each regime , we examined each regime alone ( 10 samples ) : Expression ~ diet + block . The Benjamini-Hochberg procedure was used to control the false discovery rate ( FDR . i . e . , q-value; [42] ) in R ( version 3 . 2 . 0 , R-Development-Core-Team 2015 ) . Gene expression changes between diets or regimes were calculated as log2 fold changes ( log2FC ) between two tested groups . To examine whether the initial plasticity in the GA population tends to be reinforced or opposed during adaptive differentiation , we screened for genes that have log2FC between the two diets greater than 0 . 4 in GA and show a strong “selective history” effect ( q < 0 . 1 ) in linear model of expression comparing Cad and Salt populations . To compare plasticity of the GA to evolved differences between Cad and Salt , we used the log2FC between diets for the GA and then calculated the log2 fold change between the replicate Cad population and the replicate Salt population for gene i in block j as log2FCi , j= ½Z ( log2 ( Ecadmium , i , Cad_j ) + log2 ( Esalt , i , Cad_j ) −log2 ( Ecadmium , i , Salt_j ) −log2 ( Esalt , i , Salt_j ) ) where Ed , i , j is the normalized expression in diet d for gene i in population j ( number of expression counts divided by the total number of counts of the sample ) . Z serves as an indicator of whether plasticity in GA is in the same or opposite direction of adaptive divergence between Cad and Salt: Z = 1 if expression was up-regulated in cadmium for GA and Z = -1 if expression was down-regulated in cadmium for GA . We averaged the expression changes across the screened genes for each replicate pair ( i . e . , block ) . To visually assess the overall patterns of variation in the transcriptome among samples , we first performed principle component analysis for all samples , including the ancestral populations , using DESeq2 . The DESeq dataset object was constructed from the matrix of the count data and the sample information table , with design format as ~ regime + diet . After regularized-logarithm transformation ( rlog ) , the top 1000 genes with highest variance across samples at the transformed scale were used for principle component analysis ( PCA ) . The principal component value for each sample was obtained by the function plotPCA . The values for all samples with respect to the first and second principal components are plotted in S3 Fig . The samples from ancestral populations are somewhat distinct from the experimental population samples along the PC1 axis . The separation between samples from ancestors and experimental populations may be due to subtle life history differences because the ancestral populations are maintained slightly differently ( in terms of density and other maintenance procedures ) or because the ancestral populations were collected for RNAseq in a different week ( i . e . , block effect ) . To qualitatively assess whether block effects tend to be large , we repeated the same PC analysis without the ancestors , with the design format changed to ~ regime + block . From visual inspection , there is no indication of strong block effects among the experimental populations , either in the PCA above or in a PCA based on only the experimental populations , i . e . , excluding the ancestors ( S4 Fig ) . This PCA ( without the ancestors ) is the one represented in Fig 3 . To further explore the functionality of different PC axes , we extracted the loading value for each of the 1000 genes on different PC axes using prcomp function . Using the R package “gage” , we tested , for each PC , whether different GO Ontology categories were significantly associated with either positive or negative loadings on that PC . To identify genes expected to evolve increased plasticity in heterogeneous regimes , we used a screen based on the Ancestral Cadmium ( AC ) and Ancestral Salt ( AS ) populations . We treated the samples of Ancestral Cadmium ( AC ) in the cadmium diet and Ancestral Salt ( AS ) in the salt diet as “Optimal” and samples of Ancestral Cadmium ( AC ) in the salt diet and Ancestral Salt ( AS ) in the cadmium diet as “Non-adapted” . We identified candidate genes that should be selected for increased plasticity in heterogeneous regimes by finding genes that meet the following criteria: ( i ) large expression differences between the two “Optimal” states ( |log2FC| > 0 . 4 ) ; and ( ii ) low levels of plasticity relative to regime effects ( the |log2FC| between diets for both AC and AS populations is less than half the |log2FC| between AC and AS within each diet ) . The cut-off values used represent a compromise between high stringency to obtain a set of genes with the desired properties and ensuring a reasonable number of genes ( ~100 ) pass the screen to allow for meaningful downstream analysis . For this set of genes , we calculate the relative adaptive plasticity ( polarized plasticity ) as: log2FCi , j , p=Y* log2FCi , j where Y = {-1 , 1} is an indicator of whether the direction of plasticity matches the direction of difference between the two “Optimal” ancestral states . log2FCi , j is the log2 fold change for gene i in population j , calculated by DESeq2 . We calculated the average log2FCi , j , p across the genes for each population . The paired comparisons of the average log2FCi , j , p among selective regimes were based on ANOVA Tukey HSD tests . To test whether the heterogeneous populations differ from homogeneous populations in relative adaptive plasticity , we analyzed the average polarized plasticity among the gene set log2FCi , j , p¯ using the lmer function in the lme4 package in R: log2FCi , j , p¯ = regime + selective history + block where regime is homogeneous or heterogeneous; selective history ( Cad or Salt , Temp or Spatial ) is nested within the homogeneous ( Cad and Salt ) or heterogeneous regime ( Temp and Spatial ) ; block was treated as random effect . The regime effect was tested by comparing the full model with a model without the regime effect . To identify genes expected to evolve reduced plasticity in heterogeneous regimes , we again treated the samples of AC in the cadmium diet and AS in the salt diet as “Optimal” and AC in the salt diet and AS in the cadmium diet as “Non-adapted” . We used two liberal criteria to screen for genes for hypothesis testing: ( i ) genes must be differentially expressed between the “Optimal” and “Non-adapted” states: |log2FC| > 0 . 3 and the difference between “Optimal” and “Non-adapted” must be in the same direction for both AC and AS ( i . e . , “Optimal” states both have higher or both have lower expression than the “Non-adapted” states ) ; and ( ii ) differences in the adaptive state of the ancestors must be relatively low: the |log2FC| between “Optimal” states for AC vs . AS must be less than half as large as the |log2FC| for “Optimal” vs . “Non-adapted” for both AC and AS . For each gene passing the two criteria , we calculated the scaled absolute plasticity ( |log2FC| ) based on DESeq2 and took the average across genes for each population . The comparisons among selective regimes were based on ANOVA Tukey HSD tests . To further examine how selective history alters expression on genes of interest , we calculated the scaled distance to the “adaptive” optimum for expression in diet d of gene i of population j ( Φd , i , j ) : Φd , i , j=Ed , i , j−Od , iNd , i−Od , i where Od , i is the expression for the sample representing the “Optimal” state for diet d ( AC in cadmium diet or AS in salt diet ) and Nd , i is the expression for the sample representing the “Non-adapted” state for diet d . For each population , we calculated the average Φd , i , j over all the genes of interest for each diet separately . The average value of Φd , i , j across all genes in each gene set was calculated for each population . These average values were used in comparisons among regimes . Gene Ontology enrichment test was performed with the R package “gage” [43] with ranked based two-sample t-test . Different sets of genes were tested for functional enrichment: genes for principle component analysis , genes for differential expression analysis between diets/regimes . Selection of overrepresented GO terms among all the tested GO ( only considering the terms that do not associate with the child terms ) was based on FDR ( q ) < 0 . 05 and were reported on different directions separately ( i . e . , positive or negative loading values on each PC axis; up-regulated in one or the other diet/regime ) . For the genes involved in comparing initial plasticity in the GA and evolved divergence , we obtained the functionality information and tested for overrepresented GO terms using the Database for Annotation , Visualization and Integrated Discovery ( DAVID ) [44 , 45] . We performed the same GO analysis for the gene sets predicted to evolve increased or decreased plasticity . To identify genes with different levels of allele expression bias in different diets ( “allelic plasticity” ) for each regime , we assumed the allele frequencies are the same for the two samples in different diets from a replicate population . Therefore , the difference in the ratio of mRNA levels of two alleles between diets is due to expression changes but not DNA . There will be some difference due to sampling of larvae ( “genetic drift” ) but this should be minimal because 80 larvae where taken for each sample . More importantly , the direction of drift should not be the same across replicates ( reducing the statistical power to detect allelic plasticity rather than creating false positives ) . It is possible that during the 12 hours of diet feeding , selection changed the allele frequency . However , the opportunity for selection seems very small as dead larvae were rarely observed and almost all the larvae on the medium were collected for RNA extraction . However , if different genotypes grow at different rates at different diets , the changes in the allelic expression ratio will be due to the changes of relative contribution of mRNA from different genotypes . Bearing this caveat in mind , we examined between-diet differences in allelic expression within each regime . ( An additional analysis described below found no evidence of “allelic plasticity” being in the direction of the allele favoured in a given diet , providing further evidence that selection is not responsible for observed instances of allelic plasticity . ) To test for allelic plasticity , we first used Popoolation2 [46] to obtain the counts for different nucleotides in each position of the genes . To control the statistical power of identifying allelic expression among diets , we down-sampled the nucleotide counts for each site to the minimum coverage of the sites in each block ( eight samples ) . For all the samples from regime i in both diets , we screened for sites that ( i ) have average diversity 2piqi > 20% ( where pi is the average mRNA nucleotide count frequency across the 10 samples for regime i , qi = 1 –pi ) , ( ii ) the total count for the site passes ni > 100 ( where ni is the total count for that site among the 10 samples for regime i ) , and ( iii ) ni is at least half of the total read count for that gene . We chose the most informative site within each gene for each regime ( highest nipiqi ) , to calculate its allelic expression . Significant allelic plasticity was identified by Cochran-Mantel-Haenszel ( CMH ) test in R ( p < 0 . 01 ) . Two samples in different diets from the same replicate population were paired in the CMH test . To test whether allelic plasticity is overrepresented among genes showing abundance plasticity , we only considered genes that are included in both the allelic plasticity analysis and the abundance plasticity analysis . Significant abundance plasticity ( i . e . , differential expression between diets ) was identified by DESeq2 ( p < 0 . 05 for “diet” effect on expression ) . We calculated the fraction of genes showing allelic plasticity amongst genes with significant abundance plasticity ( f1 ) and the fraction of genes showing allelic plasticity amongst genes without significant abundance plasticity ( f2 ) . The difference between the two fractions is diff_f = f1 –f2 . To evaluate whether the observed diff_f is significantly different than expected by chance , we used permutations to produce an empirical null distribution of diff_f for each regime . For each site within each population , we permutated the two alleles across diets , keeping total read count within each diet unchanged . Each permuted site was tested by CMH test for allelic bias expression . “Significant” genes in this permutation test result from chance associations ( i . e . , classic false positives ) but may be more likely to occur for some genes than others ( i . e . , genes with low coverage in one or both diets ) . The number of pseudo-significant genes was much lower than the observed number of sites with significant allelic plasticity . To complete the enrichment test , we designated “significant allelic bias” to randomly chosen sites until the total number of genes with “significant allelic bias” was the same as the actual number . The genes with significant abundance expression plasticity were based on the actual data ( i . e . , this feature of the data was not permuted ) . For each permuted data set , we calculate the difference in fraction of allelic expression bias genes between genes with or without significant abundance expression plasticity ( diff_f_permuted ) . We performed 5000 permutations to generate the distribution of diff_f_permuted for each regime . The p-value was computed as the twice of the proportion of the permuted statistics that were equal or more extreme than the actual diff_f . We obtained sites that are significantly differentiated with respect to genotypic SNP frequencies between cadmium- and salt-selected populations from previously reported genotypic sequencing data at generation 42 [20] . To determine whether the genes with ecologically differentiated SNPs are enriched for genes showing differential expression across regimes or diets , we only analyzed genes that not only have polymorphic SNPs for the allele frequency differentiation test [20] but are also involved in differential expression analysis . Further , we divided the genes into different categories based on whether the polymorphic SNPs are located in exons , introns and 1kb upstream and downstream of the genic region for enrichment test; genes with multiple SNP-types ( e . g . , SNPs in both exons and introns ) are used in each relevant test . χ2 tests were used to test for significant enrichment . In addition , we searched for evidence for adaptive differential expression of alternative alleles between diets . To obtain high statistical power , we identified SNPs that show significant genomic allele frequency differentiation between the six cadmium- and six salt-selected populations ( by including AC and AS ) , only using the sites identified as having alternative alleles that are differentially expressed between diets , i . e . , allelic plasticity ( p < 0 . 01 ) . For the set of genes that show both allelic plasticity and significant genomic allele frequency differentiation , we examined whether the direction of allelic plasticity between cadmium and salt diets is aligned with the direction of allele frequency change between cadmium- and salt-selected populations . For each of these genes , the allelic bias expression was assigned as positive if the direction of change between environments is the same , otherwise it was negative . The average allelic bias expression in the direction of allele frequency change was tested based on whether the bootstrapped distribution overlaps with 0 . | Different developmental environments change how genes are expressed and what phenotypes are produced . Here we examine how the responsiveness of gene expression to different environments ( “expression plasticity” ) evolves in populations adapted to constant environments or heterogeneous ones ( temporal or spatial heterogeneity ) using experimental populations of D . melanogaster . We find the plastic response of the ancestral population that is naïve to both environments is generally opposed by the evolved differences between populations adapted to alternative environments . Populations that live in heterogeneous environments show evidence of adaptive expression evolution in genes predicted to evolve changes in plasticity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"sodium",
"chloride",
"salts",
"diet",
"developmental",
"biology",
"nutrition",
"evolutionary",
"adaptation",
"cadmium",
"gene",
"expression",
"life",
"cycles",
"chemistry",
"evolutionary",
"genetics",
"genetic",
"screens",
"evolutionary",
"processes",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"evolutionary",
"biology",
"larvae",
"chemical",
"elements"
]
| 2016 | Experimental Evolution of Gene Expression and Plasticity in Alternative Selective Regimes |
Dengue virus ( DENV ) causes the most prevalent arthropod-borne viral disease in humans . Although Aedes mosquitoes transmit DENV when probing for blood in the skin , no information exists on DENV infection and immune response in the dermis , where the blood vessels are found . DENV suppresses the interferon response , replicates , and causes disease in humans but not wild-type mice . Here , we used mice lacking the interferon-α/β receptor ( Ifnar–/– ) , which had normal cell populations in the skin and were susceptible to intradermal DENV infection , to investigate the dynamics of early DENV infection of immune cells in the skin . CD103+ classical dendritic cells ( cDCs ) , Ly6C– CD11b+ cDCs , and macrophages in the steady-state dermis were initial targets of DENV infection 12-24 hours post-inoculation but then decreased in frequency . We demonstrated recruitment of adoptively-transferred Ly6Chigh monocytes from wild-type and Ifnar–/– origin to the DENV-infected dermis and differentiation to Ly6C+ CD11b+ monocyte-derived DCs ( moDCs ) , which became DENV-infected after 48 hours , and were then the major targets for virus replication . Ly6Chigh monocytes that entered the DENV-infected dermis expressed chemokine receptor CCR2 , likely mediating recruitment . Further , we show that ∼100-fold more hematopoietic cells in the dermis were DENV-infected compared to Langerhans cells in the epidermis . Overall , these results identify the dermis as the main site of early DENV replication and show that DENV infection in the skin occurs in two waves: initial infection of resident cDCs and macrophages , followed by infection of monocytes and moDCs that are recruited to the dermis . Our study reveals a novel viral strategy of exploiting monocyte recruitment to increase the number of targets for infection at the site of invasion in the skin and highlights the skin as a potential site for therapeutic action or intradermal vaccination .
The skin is the barrier to the environment and provides a first line of defense against invasion of microbial pathogens . Dendritic cells ( DCs ) and macrophages ( MΦs ) serve as immune sentinels in the skin [1] . DCs take up antigen , sense the presence of invading pathogens , and migrate to draining lymph nodes ( LNs ) , where they prime naïve T cells [2] . MΦs are tissue-resident cells that are specialized in phagocytosis and local antigen presentation to effector and memory T cells [3] . Several subsets of DCs have been identified in the steady-state skin . The epidermis contains Langerhans cells ( LCs ) that self-renew [4] . The dermis of mice contains CD103+ classical DCs ( cDCs ) and CD11b+ DCs [5] , [6] that are replenished by blood-derived precursors . In other non-lymphoid tissues , CD103+ cDCs are derived from pre-cDCs – precursors down-stream of common DC progenitors [7]–[10] . CD11b+ DCs are derived from pre-cDCs as well as from monocytes [11] , suggesting that CD11b+ DCs are heterogeneous and need to be further resolved . Additionally , the entry of pre-cDCs into the steady-state dermis and replenishment of dermal DCs has not been demonstrated . Inflammation drastically changes the network of immune cells in the skin . Ultraviolet light , chemicals , or herpes simplex virus-1 infection induce the migration of epidermal LCs [4] and dermal DCs [12] , [13] to LNs , where they prime CD4+ and CD8+ T cell responses . Ly6Chigh monocytes enter the inflamed epidermis to replenish LCs [14] and are recruited to other inflamed tissues , where they differentiate to monocyte-derived DCs ( moDCs ) [15] . Two studies showed monocyte recruitment and differentiation to moDCs in the inflamed dermis during Leishmania major infection [16] and contact hypersensitivity reaction [17] . Yet , many questions remain as to how DCs are replenished in the inflamed dermis and how pathogens overcome the immune response in the skin to establish infection . The four dengue virus serotypes ( DENV1–4 ) cause the most common arthropod-borne viral disease of humans , with 390 million infections and up to 96 million cases of dengue per year [18] . No specific vaccine or therapeutic exists against dengue . DENV is a Flavivirus that contains a positive-strand RNA genome encoding 3 structural ( C , prM/M , E ) and 7 non-structural proteins [19] . Aedes aegypti and Ae . albopictus mosquitoes transmit DENV when probing for blood vessels in the dermis [20] . After systemic spread , monocytes , DCs , and MΦs are the main targets for DENV replication [21]–[23] . The few studies that have examined the skin found DENV infection in epidermal LCs [24]–[26]; however , no information exists about DENV infection and the immune response in the dermis , where DENV is most likely transmitted . Memory responses raised during a DENV infection modulate disease severity during a subsequent DENV challenge . Most primary ( 1° ) DENV infections are subclinical or manifest as dengue fever and induce protective immunity against the same DENV serotype . In contrast , subsequent infection with a different DENV serotype may lead to potentially fatal dengue hemorrhagic fever/dengue shock syndrome , due to antibody-dependent enhancement ( ADE ) [27] and/or serotype cross-reactive T cells [28] . During ADE , antibodies from a previous DENV infection bind , but do not neutralize , the secondary DENV serotype , facilitate DENV infection of Fcγ-receptor expressing cells , and may thus increase disease severity [27] , [29] , [30] . By the time symptoms of dengue develop 4–8 days after the bite of a DENV-infected mosquito , the site of DENV transmission is no longer apparent . Therefore , biopsies of naturally DENV-infected human skin are not available , and animal models must serve to study dynamics of the immune response in the skin . DENV suppresses the interferon ( IFN ) response , replicates , and causes disease in humans but not wild-type ( WT ) mice [31]–[33] . Most DENV infection models use mice deficient in the IFN pathway , such as IFN-α/β and -γ receptor-deficient ( AG129 ) mice that display virus tropism similar to humans and a vascular leak syndrome with key features of severe dengue disease [34] , [35] . We recently improved this model by using the virulent DENV2 strain D220 in the less immunodeficient Ifnar–/– mice in the C57BL/6 background , which lack the IFN- receptor but express functional IFN-γ receptor [36] . Here , we establish an intradermal ( i . d . ) DENV infection model in Ifnar–/– mice to study the early immune response during DENV infection of the dermis . Comparing cell populations in the steady-state dermis and performing adoptive transfers of WT or Ifnar–/– monocytes , we confirm the normal phenotype , frequency , and response of Ifnar–/– monocytes and DCs in the skin . Using Ly6C expression , we resolved the heterogeneity of dermal DCs . We find that dermal Ly6C– CD11b+ cDCs and MΦs are the initial targets for DENV infection , but Ly6Chigh monocytes are recruited to the dermis , differentiate to Ly6C+ CD11b+ moDCs , and become the major targets for DENV infection after 48 h . Our study unveils a novel viral strategy of exploiting monocyte recruitment to increase the targets for virus replication at the site of transmission in the skin .
To establish an intradermal DENV infection model , we evaluated DENV infection of the skin of WT and Ifnar–/– mice . We infected WT and Ifnar–/– mice i . d . with DENV2 under 1° or ADE conditions . As expected , DENV2 did not cause disease in WT mice after i . d . inoculation under either condition ( Figure S1 ) . In contrast , Ifnar–/– mice developed mild disease after i . d . inoculation with 106 plaque-forming units ( PFU ) of DENV2 under 1° conditions and showed significantly more severe and prolonged disease under ADE conditions ( Fig . 1 , A and B ) . Whereas 1° infections were sublethal , 60% of Ifnar–/– mice succumbed to ADE infection 5-7 days post-i . d . inoculation ( Fig . 1C ) via a vascular leak-like syndrome , which displays features similar to severe dengue disease in humans [29] . Since little information exists on early DENV infection in the skin before onset of illness , we focused on the first 3 days of DENV infection ( Fig . 1 , A and B ) . No DENV E protein was detectable in CD45+ hematopoietic cells in the epidermis or dermis of WT mice ( Fig . 1 , D and E ) 48 h post-i . d . inoculation ( hpi ) with 106 PFU DENV2 under 1° or ADE infection conditions . In contrast , CD45+ cells in the epidermis and dermis of Ifnar–/– mice displayed DENV E staining ( Fig . 1 , D and E ) . We further stained skin samples from Ifnar–/– mice for intracellular DENV non-structural protein NS3 , indicative of active viral replication [22] . No significant difference in DENV infection of total CD45+ cells in the skin of Ifnar–/– mice existed between 1° and ADE conditions ( Fig . 1 , F and G ) , implying that the differences in pathogenesis occurred later , after systemic viral spread . While the epidermis of Ifnar–/– mice contained <0 . 54% DENV-infected CD45+ cells ( Fig . 1F ) , the dermis contained up to 17% DENV-infected CD45+ cells ( Fig . 1G ) and 3 . 3-fold more total CD45+ cells than the same area of epidermis . Hence , the dermis of Ifnar–/– mice contained ∼100-fold more DENV-infected CD45+ cells than the epidermis at 72 hpi . Next , we set out to identify the cell types that DENV infects in the skin and first examined the epidermis more closely . We evaluated the frequency of LCs in the epidermis of WT and Ifnar–/– mice in steady state , gating LCs as CD45+ MHC II+ Langerin+ ( Fig . 2A ) . The phenotype and frequency of LCs were similar in WT and Ifnar–/– mice , with LCs comprising ∼46% of CD45+ cells in both mouse strains ( Fig . 2B ) . Twelve and 24 hpi with DENV , LC frequencies significantly decreased in the epidermis of Ifnar–/– mice , with a decline of 42% at 24 h ( Fig . 2 , C and D ) compared to steady-state untouched ears . Primary and ADE infection conditions as well as inoculation with PBS reduced LCs frequencies similarly . This was likely caused by exit from the epidermis and migration to LNs , as DENV inoculation did not increase LC death in the epidermis . By 48 hpi , LC frequency recovered to steady-state levels ( Fig . 2D ) . Staining for DENV proteins demonstrated infection of epidermal LCs ( Figure 2E ) , which peaked 48–72 hpi and was similar under 1° and ADE conditions ( Fig . 2F ) . LCs were the main DENV-infected hematopoietic cells in the epidermis , as over 90% of NS3+ E+ cells were LCs ( Fig . 2G and Figure S2 ) , and the remaining epidermal CD45+ cells that were mostly γδT cells were not DENV-infected . Our results in the Ifnar–/– model are in line with DENV2 inoculation of human skin explants that have found DENV infection of epidermal LCs [24] , [25] . We next focused on the dermis , evaluating DC , monocyte , and MΦ frequencies in the dermis of WT and Ifnar–/– mice in steady state . The dermis contained similar cell populations in steady-state WT ( Fig . 3A ) and Ifnar–/– mice ( Fig . 3B ) . We separated dermal CD45+ cells by MHC II expression into MHC IIhigh DCs and MHC IIlow/– non-DCs . The main dermal population was CD11b+ DCs ( on average ∼14% of all CD45+ cells , Fig . 3C ) , gated as MHC IIhigh CD103– CD11b+ Langerin– . Additionally , the dermis contained MHC II+ CD103– CD11b+ Langerin+ cells , which were likely dermal LCs migrating from the epidermis through the dermis to LNs [37] . Both WT and Ifnar–/– mice contained dermal CD103+ cDCs ( ∼2 . 5% of CD45+ cells; Fig . 3D ) , gated as MHC IIhigh CD103+ Langerin+ . The steady-state dermis contained only 0 . 87% Ly6Chigh monocytes of all CD45+ cells ( Fig . 3E ) , gated as MHC II– CD11b+ Ly6G– Ly6Chigh , and did not contain MHC II– CD11b+ Ly6G+ granulocytes . Dermal MΦs were MHC IIlow/– CD11b+ Ly6G– Ly6Clow/– , expressed F4/80 and a high FSC/SSC profile ( Figure S3 ) , and constituted ∼4 . 6% of CD45+ cells ( Fig . 3F ) . The hematopoietic cells in the skin that did not stain positive for the indicated markers mostly consisted of T cells expressing the γδT-cell receptor , other T cells , or mast cells . No statistically significant difference existed between WT and Ifnar–/– mice in any cell population examined in the steady-state skin ( Fig . 3C–F ) . These results indicate that IFNAR signaling is not required to maintain hematopoietic cells in the steady-state skin and substantiate Ifnar–/– mice as a model to study immune cells in the skin . We continued to examine DENV infection in the dermis of Ifnar–/– mice because WT mice did not develop disease or support DENV2 replication , but Ifnar–/– mice carried equivalent immune cell populations in steady-state skin , supported DENV infection , and developed key features of human dengue disease . No information existed on the cells that first encounter DENV in the dermis and the dynamics of the DENV-induced immune response . We monitored dermal cell populations after i . d . inoculation with DENV2 or PBS in Ifnar–/– mice ( Fig . 4A and Figure S4 ) . Notably , dermal Ly6Chigh monocytes increased significantly during DENV infection compared to steady-state untouched ears and peaked after only 12 h ( Fig . 4B ) . The increase of Ly6Chigh monocytes was significantly higher during ADE compared to 1° infection conditions ( p<0 . 05 at 12 hpi ) with a 39- and 30-fold peak increase , respectively . PBS-injected controls showed some increase of Ly6Chigh monocytes compared to steady state , but significantly less than during DENV infection ( 24 h , p<0 . 01; 48 h , p<0 . 0001; and 72 h , p<0 . 01 ) . Furthermore , CD11b+ DCs increased significantly , ∼2 . 5-fold in the dermis 48 hpi with DENV ( Fig . 4C ) . As Ly6Chigh monocytes enter inflamed tissues and give rise to moDCs [15] , [38] , [39] , we hypothesized that moDCs contributed to the increase in CD11b+ DCs in the DENV-infected dermis . Dermal CD11b+ DCs were heterogeneous and , in line with recent studies [16] , [17] , could be separated based on Ly6C expression ( Fig . 4D and Table 1 ) . The steady-state dermis contained mostly Ly6C– CD11b+ cDCs , which decreased 24 hpi with DENV2 ( Fig . 4E ) . In contrast , Ly6C+ CD11b+ moDCs significantly increased and peaked at 48 h ( Fig . 4F ) . The increase of Ly6C+ CD11b+ moDCs was significantly higher during ADE compared to 1° infection conditions ( p<0 . 05 at 48 hpi ) with a 21- and 16-fold peak increase , respectively . PBS-injected controls showed some increase of Ly6C+ CD11b+ moDCs ( Fig . 4F ) , but significantly less than during DENV infection ( 24 h , p<0 . 01; 48 h , p<0 . 0001; and 72 h , p<0 . 05 ) . Ly6C expression thus resolved the heterogeneity of CD11b+ DCs into Ly6C– CD11b+ cDCs resident in steady-state dermis and Ly6C+ CD11b+ moDCs that increased during DENV infection . In addition , CD103+ cDCs ( Fig . 4G ) and MΦs ( Fig . 4 H ) significantly decreased in the dermis by 12 and 48 hpi with DENV2 , respectively , and remained low through 72 h . PBS-injected controls also showed reduction of CD103+ cDCs and MΦs , but less pronounced than after inoculation with DENV2 ( at 24 hpi , p<0 . 05 for CD103+ cDCs and p<0 . 01 for MΦs ) . In contrast , granulocytes significantly increased 16-fold after i . d . inoculation with DENV2 , showing a temporary peak after 12–24 h , but then decreased ( Fig . 4I ) . No increased cell death was observed in the dermis after inoculation with DENV , suggesting that the decreased number of cDCs was due to exit of cDCs from the dermis and migration to LNs . Overall , the immune response in the dermis was slightly more pronounced during ADE compared to 1° DENV infection . The increase in dermal Ly6Chigh monocytes and moDCs , as well as decrease in cDCs and MΦs , likely impacted the number of targets for DENV replication . Previous studies examined DENV infection only in the epidermis . Here , we analyzed DENV infection via intracellular staining of DENV proteins E and NS3 in the dermis ( Fig . 5A–F ) . Dermal Ly6Chigh monocytes ( Fig . 5G ) and Ly6C+ CD11b+ moDCs ( Fig . 5H ) showed no or low levels of DENV infection 12–24 hpi , but became DENV-infected by 48 hpi and then steeply increased DENV infection , up to 37% for moDCs at 72 h . In contrast , more than 20% of Ly6C– CD11b+ cDCs , CD103+ cDCs , and MΦs became DENV-infected within the first 12–24 h ( Fig . 5I-K ) . While DENV infection of Ly6C– CD11b+ cDCs and CD103+ cDCs remained high throughout ( 20–34% ) , DENV infection of MΦs declined after 24 h ( Fig . 5K ) . Less than 10% of granulocytes showed DENV infection at all times ( Fig . 5L ) . A trend of higher DENV infection during ADE compared to 1° conditions appeared at some time-points and populations , but no significant difference existed ( Fig . 5G–L ) . For the first time , we show that Ly6C– CD11b+ cDCs , CD103+ cDCs , and MΦs , which are present in the steady-state dermis , are the initial targets for DENV replication . During a second wave of infection , dermal Ly6Chigh monocytes and Ly6C+ CD11b+ moDCs become highly DENV-infected . The abundance of a certain cell type together with its susceptibility to virus infection determines its contribution to virus replication . We multiplied the frequency of a dermal cell type ( Fig . 4 ) by its percent DENV infection ( Fig . 5 ) to determine its contribution to overall infection ( Fig . 6 ) . DENV-infected Ly6C– CD11b+ cDCs contributed to virus replication at all times , representing ∼1 . 5–2 . 5% of total CD45+ cells 12–72 hpi ( Fig . 6 ) . In contrast , DENV-infected MΦs contributed 0 . 8% of total CD45+ cells during the first 24 h , but then waned . Although infected , CD103+ cDCs did not contribute considerably to DENV replication at any time-point examined ( ≤0 . 1% ) due to their low frequency after inoculation with DENV . Classical DCs and MΦs that reside in the steady-state dermis thus were the main DENV-infected cells in the initial phase , up to 24 hpi . In a second phase , 48–72 hpi , monocytes and moDCs became the main targets of DENV replication in the dermis ( Fig . 6 ) . While Ly6Chigh monocytes were present in high numbers as early as 12 hpi ( Fig . 4B ) , they became DENV-infected only after 48 h and then contributed ∼2 . 3% of CD45+ cells ( Fig . 6 ) . Ly6C+ CD11b+ moDCs showed substantial DENV infection starting at 48 hpi , when they peaked in frequency ( Fig . 4F ) , and DENV-infected moDCs constituted 7 . 6% of total CD45+ cells at 72 hpi ( Fig . 6 ) . The mean % CD45+ cells consisting of DENV-infected Ly6C+ CD11b+ moDCs was significantly larger during ADE ( mean 6 . 4%±0 . 75 SEM ) compared to 1° infection conditions ( 3 . 8%±0 . 61 ) at 48 hpi ( p<0 . 05 , unpaired t-test ) . Together , DENV-infected monocytes and moDCs made up 79% of all infected cells by 72 hpi and were thus the major targets for virus replication in a second wave of infection . Next , we aimed to determine the mechanism of how monocytes and moDCs increase in the DENV-infected dermis . To test the hypothesis that Ly6Chigh monocytes enter the DENV-infected dermis , differentiate to Ly6C+ CD11b+ moDCs , and sustain DENV infection , we determined CCR2 expression , which has been shown to be essential for repopulation of dermal DCs in the inflamed skin [40] , and performed adoptive transfer experiments . Ly6Chigh monocytes and Ly6C+ CD11b+ moDCs in the DENV-infected dermis expressed chemokine receptor CCR2 on their surface 24 hpi with DENV under 1° or ADE conditions ( Fig . 7 ) , similar to Ly6C– CD11b+ cDCs , which is in line with previous studies [17] . In contrast , granulocytes that were also recruited to the dermis did not express CCR2 . These results suggest that Ly6Chigh monocytes may be recruited to the DENV-infected dermis via signaling through CCR2 , while other migratory markers mediate the recruitment of granulocytes . Next , we isolated monocytes from the bone marrow of steady-state WT or Ifnar–/– mice via negative magnetic-bead selection , yielding purities of 91–95% and a Ly6Chigh CCR2high phenotype ( Figure S5 ) . Monocytes were labeled with CFSE , transferred intravenously into Ifnar–/– hosts , and allowed to home for 24 h . Then , recipients were inoculated i . d . with DENV2 into one ear under 1° infection conditions ( Fig . 8A ) . Before extracting skin samples for analysis , euthanized recipients were perfused with PBS to exclude cells circulating in the blood from the samples . Mice that were infected with DENV2 but had not received monocyte transfers served as gating controls for CFSE+ graft-derived cells . Adoptively transferred CFSE+ Ly6Chigh monocytes from WT ( Fig . 8B ) or Ifnar–/– ( Fig . 8C ) origin engrafted robustly into the dermis of the DENV-infected ear of Ifnar–/– hosts 48 hpi . In contrast , the dermis of control steady-state recipients ( Figure S5 ) or the non-infected side of DENV-infected recipients ( Fig . 8 , B and C ) showed minimal engraftment . No monocyte engraftment was detected in the epidermis . These results provide direct evidence that monocytes were specifically recruited and entered the DENV-infected dermis . Further , 34% of engrafted monocyte-derived WT cells ( Fig . 8B ) and 39% of Ifnar–/– cells ( Fig . 8C ) in the DENV-infected dermis expressed MHC II and Ly6C and had thus differentiated to Ly6C+ CD11b+ moDCs . The remaining graft was still Ly6Chigh monocytes . Therefore , the recruitment and differentiation of WT and Ifnar–/– monocytes during early DENV infection was equivalent and thus independent of IFNAR signaling . Although , as expected , WT cells were not infected ( Fig . 8D ) , approximately 22% of de novo-recruited monocytes and moDCs that derived from the Ifnar–/– graft were DENV-infected 48 hpi ( Fig . 8E ) . We show for the first time that the increase of Ly6Chigh monocytes in the DENV-infected dermis is due to specific recruitment from the blood and entry into the DENV-infected dermis , which was likely mediated by CCR2 . We further demonstrate the differentiation of Ly6Chigh monocytes to Ly6C+ CD11b+ moDCs and DENV infection of de novo-recruited cells , leading to the increase in targets for DENV replication in the dermis . Lastly , we examined the activation state of immune cells in the DENV-infected dermis , comparing 1° versus antibody-enhanced DENV infection to steady-state control mice . Relative to isotype-matched control stains , Ly6C+ CD11b+ moDCs and Ly6C– CD11b+ cDCs expressed substantial amounts of the activation marker CD80 ( B7-1 ) on the surface already in steady state , whereas Ly6Chigh monocytes and MΦs expressed lower levels of CD80 ( Fig . 9A ) . Forty-eight hours after DENV infection , non-infected bystander monocytes , Ly6C+ CD11b+ moDCs , and Ly6C– CD11b+ cDCs upregulated CD80 in the dermis , whereas upregulation of CD80 was diminished in DENV-infected cells ( Fig . 9A ) . The upregulation of CD80 in the dermis was more pronounced during ADE than during primary DENV infection . Steady-state and non-infected bystander MΦs expressed relatively low levels of CD80 , and DENV-infected MΦs significantly down-regulated CD80 expression compared to steady-state or non-infected bystander cells . We also measured CD86 ( B7-2 ) surface expression and observed similar trends as for CD80 expression on Ly6C– CD11b+ cDCs and MΦs , i . e . , that DENV-infected cells expressed significantly lower levels of CD86 than non-infected bystander cells ( Fig . 9B ) . Our results show that DENV-infected monocytes , DCs , and MΦs express lower levels of activation markers than non-infected bystander DCs in the dermis . This confirms earlier studies that showed that DENV blocks the activation of DENV-infected human moDCs compared to non-infected bystander cells in vitro [41] , [42] .
While several subsets of DCs in the steady-state skin have been identified , many questions about pathogen invasion via the skin remain unanswered . Here we studied the immune response during DENV infection in the dermis . We identify resident cDCs and MΦs as initial targets for DENV replication and show that the recruitment of monocytes to the dermis and differentiation to moDCs greatly increased the number of targets for DENV replication in a second wave , as summarized in Fig . 10 . Infected Aedes mosquitoes transmit DENV when probing for blood vessels in the skin , yet few studies have examined DENV infection in the skin . DENV was shown to infect DCs in human skin explants , but epidermal LCs were not distinguished from dermal DCs [24] . Inoculating the surface of human skin explants with DENV later established infection of epidermal LCs [25] , which was confirmed in a mouse model [26] . No information , however , existed about DENV infection in the dermis . Mosquitoes likely deposit DENV into the dermis when probing for blood , as the dermis , but not the epidermis , contains vasculature . We performed i . d . inoculation of DENV that was produced in C6/36 mosquito cells into the ears of mice to mimic the natural route of transmission . While confirming low levels of DENV infection of epidermal LCs , we establish that DENV infects ∼100-fold more CD45+ cells in the dermis than the epidermis , stimulating future studies to focus on the dermis as a major site for initial DENV replication . In addition , factors in the saliva of the mosquito vector may impact DENV infection and the host response in the skin . For example , saliva from Ae . aegypti mosquitoes decreased DENV infection of moDCs in vitro [43] . In contrast , two mouse models showed increased or prolonged DENV serum viremia when DENV was inoculated in the presence of mosquito saliva [44] , [45] . Clearly , further studies are required to determine the role of mosquito-derived factors in early DENV infection in the skin . Investigating the immune cell network in the dermis is important to identify the cells that first encounter DENV . We extended recent advances in surface markers that characterize immune cell subsets in the steady-state dermis [11] and further dissect the DC network in the dermis to identify the cells that first encounter DENV . Combining MHC II and CD11b staining with Ly6G and Ly6C was important for dissecting Ly6G+ Ly6C+ granulocytes from Ly6G– Ly6Chigh monocytes and DC populations––and staining for Gr-1 , which detects both Ly6G and Ly6C , would not have been sufficient . Similarly , Langerin staining was necessary to separate dermal Langerin+ CD103+ cDCs from Langerin+ CD11b+ LCs and Langerin– CD11b+ DCs . We further dissected dermal CD11b+ DCs according to Ly6C expression into Ly6C– CD11b+ cDCs and Ly6C+ CD11b+ moDCs during DENV infection , which is in line with a recent study showing that CD64– Ly6C– CD11b+ cDCs are different from CD64+ Ly6C+ CD11b+ moDCs in steady state , the latter of which clustered with Ly6Chigh monocytes in gene expression analysis [17] . The combination of 7 cellular markers and 2 DENV intracellular proteins established Ly6C as valuable marker to dissect dermal CD11b+ DCs and allowed us to identify Ly6C– CD11b+ cDCs , CD103+ cDCs , and MΦs that reside in the steady-state dermis as the initial targets for DENV replication in the skin . Knowing the initial targets of DENV infection may help develop therapeutic strategies to block DENV from establishing infection . In steady state , Ly6Chigh monocytes give rise to some CD11b+ DCs in non-lymphoid tissues [11] , [46] , but not in lymphoid tissues [47] . After depletion of CD11b-expressing cells , first Ly6Chigh monocytes reappeared in the dermis , then after 7 days came Ly6Chigh CD11b+ moDCs , and after 20 days Ly6Clow CD11b+ moDCs [17] . This suggested that few monocytes continuously enter the steady-state dermis and differentiate to moDCs . During DENV infection , we found here a large increase in dermal Ly6Chigh monocytes already at 12 hpi . Adoptive transfers provided direct evidence that Ly6Chigh monocytes are recruited to the DENV-infected dermis and rapidly differentiate to Ly6C+ CD11b+ moDCs . More than 90% of engrafted monocytes and moDCs in the dermis retained Ly6C expression 72 h after adoptive transfer of Ly6Chigh monocytes . While activation may lead to the upregulation of Ly6C in some skin-resident cells , the dynamics of monocyte recruitment and differentiation to Ly6C+ CD11b+ moDCs suggest that most Ly6C-expressing cells in the DENV-infected dermis were monocyte-derived . These results are consistent with a parasite infection where Ly6Chigh monocytes were recruited to L . major-infected dermis , differentiated to Ly6C+ CD11b+ moDCs , and became targets for infection [16] . Further , monocytes were found to replenish skin DCs during herpes simplex virus-1 infection [48] , but monocytes and DCs are not the major targets for HSV infection . During human immunodeficiency virus-1 ( HIV ) infection , DCs produce chemokines that attract CD4+ T cells , which then serve as targets for HIV replication [49] , [50] . Although DCs bind and shuttle HIV to CD4+ T cells for infection [51] , HIV does not efficiently replicate in monocytes and DCs [49] , [50] . In our study , de novo-recruited monocytes and moDCs became DENV-infected . To the best of our knowledge , our report is the first demonstration of a viral infection that uses the dermis as a primary port of entry and targets de novo-recruited monocytes and moDCs for local replication . CCR2 and its ligand CCL2 are required for the mobilization of monocytes from the bone marrow [52] and recruitment to the skin during murine cytomegalovirus infection [53] and after exposure to ultraviolet light [40] or chemical irritants [17] . In dengue patients , CCL2 was increased in the plasma and positively correlated with disease severity [54] . Here , CCR2 was present on the surface of Ly6Chigh monocytes and moDCs that were recruited to the DENV-infected dermis and likely mediated the recruitment . Nevertheless , further studies need to determine whether chemokine ligands for CCR2 are upregulated in the DENV-infected dermis , and whether CCR2 is required for the recruitment of Ly6Chigh monocytes . Alternatively , CCR6 may play a role , although this is controversial , as CCR6 was found to be important for the recruitment of monocytes into the dermis after application of chemical adjuvants [55] , but not for the repopulation of dermal DCs in the inflamed skin after exposure to ultraviolet light [40] . Exit from the skin , recruitment of precursors , and cell death determine the abundance of cell types that can serve as targets for virus replication . We did not observe increased cell death during DENV infection , and thus the decrease in dermal cDCs was likely due to exit from the dermis and migration to skin-draining LNs . DC migration from the skin to LNs is well established after exposure to ultraviolet light [40] , chemical irritants [5] , and during infection with herpes simples virus-1 [48] , [56] . Pre-cDCs replenish cDCs in other non-lymphoid tissues [11] , [46] , [57] , but direct evidence for cDCs entering the dermis is still missing . Also , the migratory ability of moDCs to LNs is under debate . De novo-generated moDCs migrated from the inflamed dermis to LNs during L . major infection [16] but had limited migratory ability during contact hypersensitivity reactions [17] . We found that moDCs accumulated in the DENV-infected dermis and served as targets for virus replication , but migration of moDCs to draining LNs remains to be determined . In contrast , migration of LCs from the inflamed epidermis to LNs is well established [4] . Most LCs that exited the DENV-infected epidermis of AG129 mice [26] and human skin explants [25] were not infected . We found that LCs decreased in the epidermis ( 24 hpi ) before they became DENV-infected by 48 hpi , and inoculation with PBS induced a similar decrease in LCs . Thus , inflammation rather than infection caused the decline in epidermal LCs . While LCs self-renew in the steady-state epidermis [4] , recruitment of Ly6Chigh monocytes replenishes LCs 4-7 days after irradiation with ultraviolet light [14] . We did not observe entry of Ly6Chigh monocytes into the DENV-infected dermis up to 72 hpi , thus proliferation of resident LCs likely replenished LCs when frequencies recovered by 48 h . Our findings in mice are consistent with DENV infection of primary human cells . In vitro-generated human moDCs were more susceptible to DENV infection than monocytes , cDCs , or macrophages under 1° infection conditions [23] , [24] , [58] . Here , we also find the highest DENV infection in moDCs , followed by cDCs , and less infection of monocytes and MΦs in the dermis . DC-SIGN ( CD209 ) is an attachment factor for DENV , positively correlates with DENV infection under 1° conditions in vitro , and is highly expressed by human moDCs [58] , [59] . In vivo , moDCs that accumulated in the LNs of mice during bacterial infection also expressed high levels of DC-SIGN [60] , which likely explains the high susceptibility of moDCs to DENV infection that we find in the dermis . In contrast , human monocytes [22] , [58] and MΦs [61] show some DENV infection under 1° conditions , but efficient DENV infection requires Fcγ receptor-mediated uptake of DENV-antibody complexes during ADE . Remarkably , Ly6Chigh monocytes were recruited to the dermis by 12 hpi but became DENV-infected only 48-72 hpi under 1° or ADE conditions . Thus , local DENV titers , cell activation , expression of virus attachment factors , and/or the microenvironment likely influence DENV infection . Indeed , activation via the cytokines GM-CSF and IL-4 was necessary for DC-SIGN expression and DENV infection of human cDCs freshly isolated from the blood [23] . In our hands , ADE resulted in only a minor increase in DENV-infected moDCs in the skin , despite inducing severe disease and mortality in Ifnar–/– mice later on . This suggests that ADE acts mostly after systemic virus spread to increase infection and pathogenesis . We further find that DENV infection blocked the expression of activation markers CD80 and CD86 on monocytes , DCs and MΦs in the dermis and that this effect was more pronounced during ADE compared to 1° infection . However , activation markers were upregulated in non-infected bystander cells within the DENV-infected dermis . These findings are in line with previous studies showing that non-infected bystander moDCs upregulate MHC I and II , as well as CD80 , CD83 , and CD86 , in DENV-infected cultures [41] , [42] . However , this activation was inhibited in DENV-infected moDCs within the same cultures , as revealed by intracellular staining for DENV proteins [41] , [42] . Further studies are needed to determine the impact of impaired activation of DENV-infected DCs on the priming of DENV-specific naïve T cells and on maintaining effector T cell responses in the periphery . Human cells infected with DENV become deficient in IFN- receptor signaling and production of IFN- because DENV proteins NS5 and NS2B/3 degrade human STAT-2 [62] and STING [33] , respectively . However , DENV proteins cannot bind and degrade the mouse homologs for STAT-2 [32] and STING [33] . We found here that i . d . -inoculated DENV did not replicate in the skin or cause disease in WT mice , similar to studies showing that intravenously infected WT mice did not show systemic DENV replication [31] . Previous models of DENV infection have used AG129 mice deficient in the IFN- as well as IFN-γ receptor [29] , [34] , [63]-[66] . This model was improved recently by generating DENV strains , such as D220 used here , that efficiently replicate in less immunocompromised Ifnar–/– mice and cause disease with key features of dengue in humans [36] , [67] . We found equal cell populations in the steady-state skin of WT and Ifnar–/– mice , and adoptive transfers of monocytes from WT and Ifnar–/– origin showed equivalent recruitment and differentiation during DENV infection . These results support that the current Ifnar–/– model is suitable to study early DENV infection and the recruitment of immune cells in the skin . Nevertheless , although suitable to study early DENV replication and recruitment of monocytes to the skin , the absence of IFNAR signaling may have effects on the subsequent priming of adaptive immune responses . In summary , we demonstrate that dermal cDCs and MΦs are the initial targets for DENV infection at the site of transmission in the skin . Further , we reveal a new viral strategy exploiting monocyte recruitment and differentiation to moDCs to increase the number of targets for DENV replication in the dermis . These results should stimulate future studies on the role of dermal DC subsets in dengue pathogenesis and in priming protective immunity during vaccination or natural infection . Thus , these findings open possibilities for early DENV control , as the skin may be a site for therapeutic action or intradermal vaccination .
Mice were bred and experiments were performed at the University of California Berkeley Animal Facilities strictly following guidelines of the American Veterinary Medical Association and the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The Animal Care and Use Committee of the University of California Berkeley has approved all experiments ( protocol R252-1012B ) . Trained laboratory personnel performed anesthesia of mice via isoflurane inhalation and euthanasia of mice using exposure to isoflurane followed by cervical dislocation . C57BL/6 wild-type ( WT ) mice were obtained from Jackson Laboratory , and C57BL/6 mice deficient in the IFN- receptor-1 ( Ifnar1tm1Agt , here called Ifnar–/– ) [68] were obtained from Dr . Daniel Portnoy ( University of California Berkeley , Berkeley , USA ) . DENV2-infected animals were monitored using a morbidity scale as follows: 1 , healthy; 2 , mild signs of lethargy; 3 , mild signs of lethargy , fur ruffling , hunched posture; 4 , increased lethargy , limited mobility , ruffled fur , hunched posture; 5 , moribund with minimal mobility and inability to reach food or water [36] . Moribund mice were euthanized immediately , scored as 5 and omitted from the mean morbidity on later days . The study performed here examines the initial infection events in the skin; therefore , we used mosquito-derived virus , produced in Aedes albopictus C6/36 cells , that best mimics the natural cycle of DENV transmission from an infected mosquito to the mammalian host . Ten passages of the clinical DENV2 isolate PL046 between C6/36 cells and serum of 129/Sv mice deficient in IFN-α/β and -γ receptors generated the strain D2S10 [34] . Ten further passages of D2S10 by the same scheme resulted in the strain D220 [36] used throughout this study . Defined mutations that resulted from the passaging procedure have been identified [34] , [36] . To grow DENV2 stocks , C6/36 cells ( obtained from Paul R . Young , University of Queensland , Brisbane , Australia ) were maintained at 28°C in M199 medium supplemented with 10% fetal bovine serum ( FBS ) , 100 U/ml penicillin-streptomycin , 10 mM HEPES , and GlutaMAX ( all obtained from LifeTech ) , rinsed with serum-free medium , and inoculated with DENV2 in RPMI 1640 medium containing 2% FBS . DENV2 was harvested on days 5 through 8 post-infection and concentrated with Amicon Ultra-15 Centrifugal Filter Units with 100 kDa molecular weight cut-off ( Millipore ) . Virus titers were determined by plaque assay using BHK-21 clone 15 ( BHK ) cells maintained at 37°C and 5% CO2 in α-MEM medium supplemented with 5% FBS , 100 U/ml penicillin-streptomycin , and 10 mM HEPES . BHK cells were seeded in 12-well plates ( Becton Dickinson ) , and at 60% confluence were inoculated with 150 µl of 10-fold serial dilutions of DENV2 stocks and incubated for 2 h before overlaying with 10% low-melting Seaplaque Agarose ( Cambrex ) in MEM medium . Plaques were counted 7 days later after fixation in 10% buffered formalin phosphate ( Fisher Scientific ) , and titers were expressed as PFU/ml . WT and Ifnar–/– mice were inoculated i . d . with 106 PFU DENV2 in 20 µl PBS using 30-gauge , 25-mm long , 10°–12° beveled removable needles and 25- µl glass syringes ( Hamilton ) . For i . d . inoculations , ears of anesthetized mice were immobilized with cover slip forceps , and the needle was inserted parallel to the skin’s surface . DENV2-injected ears were compared to PBS-injected and steady state , untouched ears . Separate sets of needles and syringes were reserved for DENV and PBS injections and were cleaned by flushing with Lysol , sterile PBS , and 70% ethanol . For 1° infection conditions , naïve mice were used . For ADE infection conditions , 5 µg anti-DENV E monoclonal antibody 4G2 were injected in 200 µl PBS into the peritoneum 24 h prior to i . d . DENV infection . Mice were euthanized 12 , 24 , 48 , and 72 hpi , and organs were harvested . Ears were removed at the base , incubated for 5 min at room temperature with hair remover lotion ( Nair ) , washed in D-PBS , and split into dorsal and ventral halves using fine-tip tweezers ( TDI ) . Ear halves were digested with 2 U/ml Dispase II in HBSS with no Ca++/Mg++ ( LifeTech ) in 5% CO2 for 90 min at 37°C [69] , while floating with the epidermal side up to achieve digestion through the dermal side [70] . The epidermis was removed as a sheet from the dermis , and both layers were cut into small pieces in separate 1 . 5 ml tubes using scissors . Epidermal and dermal layers were digested for 45 and 75 min , respectively , in 1 . 6 mg/ml collagenase type 1 ( LifeTech ) and 10 U/ml DNase 1 ( Roche ) in RPMI 1640 medium supplemented with 10% FBS at 37°C while shaking at 220 rotations per min . Homogeneous cell suspensions were generated via pipetting , and samples were filtered through 100 µm nylon meshes . Staining with Zombie Aqua ( BioLegend ) or 7-AAD ( eBiosciences ) viability dyes excluded dead cells from general analysis . Specific staining of surface markers distinguished cell types using monoclonal antibodies from BioLegend , if not stated otherwise: CCR2 ( clone 475301 , R&D ) , CD11b ( M1/70 ) , CD11c ( N418 ) , CD45 ( 30-F11 ) , CD80 ( 16-10A1 ) , CD86 ( PO3 ) , CD103 ( 2E7 ) , Ly6C ( AL21 ) , Ly6G ( 1A8 ) , MHC II ( I-A/I-E , M5/114 . 15 . 2 ) , and Armenian hamster IgG ( HTK888 ) or rat IgG2b isotype controls ( RTK4530 , BioLegend; or 141945 , R&D ) , which were conjugated to PacificBlue , Brilliant Violet 605 , PE , PE-CF594 , PE-Cy7 , Alexa Fluor 700 , APC-Cy7 , or biotin . Antibody stains were performed in D-PBS with no Ca++/Mg++ containing 2% FBS and 2 mM EDTA ( LifeTech ) . Biotinylated antibodies were visualized using streptavidin conjugated to Brilliant Violet 605 or PE-Cy7 ( BioLegend ) . After fixation with 2% formaldehyde ( Ted Pella ) , cells were permeabilized with 1 mg/ml saponin solution ( Sigma ) containing 2% FBS and 1% normal mouse serum obtained from steady-state mice . Intracellular staining for DENV proteins E ( 4G2 , ATCC ) and NS3 ( E1D8 , [35] ) , which were conjugated to Alexa Fluor 488 or Alexa Fluor 647 , respectively , using protein-labeling kits ( LifeTech ) , identified DENV2-infected cells . Intracellular staining for Langerin ( 4C7 ) was used to further dissect cell populations . Flow cytometry data were recorded with an LSR Fortessa cell analyzer ( BD Biosciences ) with 405 , 488 , 561 , and 632 nm laser excitation lines and were analyzed using FlowJo 8 . 8 . 7 software ( TreeStar ) . Gating FSC-A/SSC-A , SSC-H/SSC-W , FSC-H/FSC-W , and negative for Zombie Aqua or 7-AAD defined single , live-cell populations . Total bone marrow cells were isolated from steady-state Ifnar–/– or WT mice by crushing femur , tibia and vertebral column with a mortar and pestle and filtering through nylon meshes . Centrifugation on Ficoll-Paque Plus ( GE Healthcare ) isolated bone marrow mononuclear cells , and incubation with 5% normal rat serum blocked unspecific binding to Fc receptors . The MACS Isolation Kit ( Miltenyi Biotec ) was then used to isolate monocytes via depletion of other cell types , without using the Fc receptor block provided . After incubation with the monocyte biotin-antibody cocktail and anti-biotin MicroBeads , monocytes were collected as run-through from LS columns in a magnetic field . Monocytes were labeled for 8 min at 37°C with 2 µM CFSE ( Invitrogen ) in PBS with 1% FBS and washed twice in ice-cold PBS with 10% FBS before adoptive transfer . Kaplan-Meier curves display survival data , and statistical significance between experimental groups was determined using the Log-rank ( Mantel-Cox ) test . The Mann-Whitney unpaired , non-parametric test was used to determine significant differences between experimental groups of data pooled from independently repeated experiments , depicted in Figures as mean ±SEM , if not stated differently . P-values were considered significant at values ≤0 . 05; and p-values summarized on graphs are shown as non-significant ( NS ) for p>0 . 05; * for p≤0 . 05; ** , p≤0 . 01; *** , p≤0 . 001; and **** , p≤0 . 0001 . Data were plotted and statistically analyzed using Prism 6 . 0 software ( GraphPad ) . | The skin and its immune cells are an important barrier against invading pathogens . Dengue is a major public health problem worldwide , with no specific therapeutic or vaccine available . Aedes mosquitoes transmit dengue virus ( DENV ) to humans via the skin when taking a blood meal . Previous studies have examined DENV infection only in the epidermis , the uppermost layer of the skin , but no information existed about DENV infection in the dermis , the layer below that contains blood vessels . We established a model of DENV infection in the skin of mouse ears , as biopsies from naturally-infected human skin are unavailable . The normal dermis contains classical dendritic cells ( DCs ) and macrophages , which we found to be the initial targets of DENV infection . Monocytes that circulate in the blood were then recruited to the dermis and differentiated to monocyte-derived DCs , an inflammatory DC subset . These newly-recruited monocytes and monocyte-derived DCs became DENV-infected in a second wave and then were the major targets for DENV replication . Our study identifies how DENV exploits the immune response by infecting cells that are recruited to the skin as part of antiviral defense . These results should help future research to develop new strategies for vaccination and therapeutics against dengue . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"skin",
"blood",
"cells",
"langerhans",
"cells",
"dermatology",
"flow",
"cytometry",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"integumentary",
"system",
"antigen-presenting",
"cells",
"immunology",
"animal",
"models",
"dendritic",
"cells",
"infectious",
"disease",
"immunology",
"model",
"organisms",
"skin",
"infections",
"epidermis",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"spectrum",
"analysis",
"techniques",
"white",
"blood",
"cells",
"inflammation",
"animal",
"cells",
"dengue",
"fever",
"arboviral",
"infections",
"mouse",
"models",
"immune",
"response",
"spectrophotometry",
"cytophotometry",
"anatomy",
"cell",
"biology",
"monocytes",
"clinical",
"immunology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"viral",
"diseases",
"macrophages",
"dermis"
]
| 2014 | Monocyte Recruitment to the Dermis and Differentiation to Dendritic Cells Increases the Targets for Dengue Virus Replication |
Cell division and development are regulated by networks of kinases and phosphatases . In early Drosophila embryogenesis , 13 rapid nuclear divisions take place in a syncytium , requiring fine coordination between cell cycle regulators . The Polo kinase is a conserved , crucial regulator of M-phase . We have recently reported an antagonism between Polo and Greatwall ( Gwl ) , another mitotic kinase , in Drosophila embryos . However , the nature of the pathways linking them remained elusive . We have conducted a comprehensive screen for additional genes functioning with polo and gwl . We uncovered a strong interdependence between Polo and Protein Phosphatase 2A ( PP2A ) with its B-type subunit Twins ( Tws ) . Reducing the maternal contribution of Polo and PP2A-Tws together is embryonic lethal . We found that Polo and PP2A-Tws collaborate to ensure centrosome attachment to nuclei . While a reduction in Polo activity leads to centrosome detachments observable mostly around prophase , a reduction in PP2A-Tws activity leads to centrosome detachments at mitotic exit , and a reduction in both Polo and PP2A-Tws enhances the frequency of detachments at all stages . Moreover , we show that Gwl antagonizes PP2A-Tws function in both meiosis and mitosis . Our study highlights how proper coordination of mitotic entry and exit is required during embryonic cell cycles and defines important roles for Polo and the Gwl-PP2A-Tws pathway in this process .
The cell cycle is largely driven by networks of kinases and phosphatases that coordinate the sequential events of cell division in addition to regulating each other [1] . Kinases of the Polo , Aurora and cyclin-dependent families play particularly important roles in this process [2]–[4] . Phosphatases compete with kinases for the same substrates , and the balance between their activities is subjected to a fine regulation through the cell cycle [5] , [6] . While in budding yeast the Cdc14 phosphatase plays a crucial role in promoting mitotic exit by dephosphorylating Cdk1 substrates and promoting its inactivation [7] , it is becoming increasingly clear that a form of Protein Phosphatase 2A ( PP2A ) bound to a B-subtype adaptor subunit fulfills this function in vertebrates [8] , [9] . Yet , a clear picture of the dynamic interplay between specific kinases and phosphatases during the cell cycle is still missing . The developmental program of a complex organism requires that the cell cycle machinery adapt to the situation and contribute to the integration of cell divisions in various tissue contexts and cell types . Early embryogenesis typically involves rapid cell cycles where S-phases and M-phases alternate rapidly with little or no gap phases , growth or transcription . In Drosophila , the first 13 mitotic cycles occur in a syncytium at around 10–15 min intervals with virtually no zygotic transcription , and are driven by maternally contributed proteins and mRNAs [10] . At that stage , nuclei migrate in the syncytium , first in an axial fashion , and then towards the cortex to form the blastoderm [11] . In addition to organizing mitotic spindles , centrosomes are tethered to nuclei in the syncytium and constitute anchors for the nuclei to a network of anti-parallel astral microtubules ( MTs ) that push nuclei away from each other and towards the cortex [11] . Because of the absence of G1 at that stage , DNA replication and centrosome duplication occur shortly after mitotic exit and as early as telophase [10] . The nuclear envelope does not completely break down during syncytial mitoses , but becomes fenestrated to allow MTs to penetrate nuclei [12]–[14] . While centrosomes are dispensable for cell division in many cell types , they are absolutely essential for early embryogenesis in Drosophila [15] , [16] . The Polo kinase is a conserved , central regulator of M-phase [2] , [17] , [18] . Polo promotes mitotic entry by activating Cdc25 phosphatases that activate Cdk1/Cdc2 [19] , [20] . Cyclin B-Cdk1 triggers nuclear envelope breakdown and chromosome condensation [21] , [22] . Polo also plays important roles in centrosome maturation , chromosome attachment to MTs , bipolar spindle assembly and cytokinesis [2] , [18] . However , how Polo functions contribute to various developmental contexts has been little explored . We have recently identified the Greatwall kinase genetically in an antagonistic functional relationship with Polo in the Drosophila syncytial embryo [23] . Decreasing Polo activity and increasing Gwl activity together lead to a failure in early embryogenesis characterized by centrosome detachments from nuclei . Discovered in Drosophila as an important mitotic kinase [23] , [24] , Greatwall has rapidly emerged as a crucial regulator of M phase in Xenopus extracts [25] and in human cells [26] , [27] . The precise nature of the functional relationship between Polo and Greatwall that we uncovered genetically remained elusive [23] . Free centrosomes can occur in several distinct ways in syncytial embryos , detaching from mitotic spindles or interphase nuclei , or duplicating independently from nuclei [28] . In response to DNA damage , centrosomes can be inactivated , leading to the loss of the damaged nucleus which sinks into the yolk [29] . This response depends on the Chk2 kinase [30] . Mutations inactivating chk2 did not prevent the centrosome detachments observed in embryos where Polo function was decreased and Gwl function was increased [23] , suggesting that those events are not due to an activation of the DNA damage checkpoint , but reflect problems in coordinating the early mitotic divisions at another level . We have conducted a genetic screen to identify additional genes functioning with polo and gwl . We uncovered two genes encoding subunits of PP2A as the strongest hits in this screen: the catalytic subunit gene ( microtubule star/mts ) and the B-type adaptor subunit gene ( twins/tws ) . Phenotypic examination of the nuclear divisions in single mutants shows that Polo promotes the cohesion between centrosomes and nuclei around prophase while PP2A-Tws promotes centrosome attachment to nuclei during mitotic exit . Compromising Polo and PP2A-Tws functions simultaneously strongly enhances centrosome detachments and leads to failures in syncytial embryonic development . Moreover , we show that the Gwl kinase functions to antagonize PP2A-Tws in both meiosis and mitosis . Our results indicate that precise coordination of mitotic entry and exit is crucial during embryonic cell cycles , and implicate Polo , Gwl and PP2A-Tws as key regulators . We discuss our findings in the context of recent studies from the Xenopus extract and Drosophila systems , which also implicate a pathway linking Gwl and PP2A in regulating M phase .
Halving the levels of Polo kinase in the fly does not cause any obvious problems of development or fertility . However , embryos laid by polo heterozygous females are made inviable by a gain of Greatwall ( Gwl ) function or by overexpression of Map205 , a strong physical interactor of Polo [23] , [31] . In both cases , defective embryos are characterized by a strikingly penetrant phenotype of centrosome detachments from nuclei . These observations prompted us to examine more closely the effects of compromising Polo function in the syncytial embryo . Immunofluorescence in embryos from mothers heterozygous for the null polo11 allele [23] shows that several nuclei in prophase have one dislocated centrosome ( Figure 1A , 1B left , arrowheads , quantified in Figure 5B ) , consistent with previous results [23] . This phenotype is also observed in embryos from mothers heterozygous for polo9 , a strong hypomorph ( data not shown ) . Centrosome detachment can also be observed in prometaphase/metaphase ( Figure 1B right ) . Thus , the cohesion between centrosomes and nuclei in early mitosis is sensitive to Polo kinase levels . The Polo kinase is essential for a wide variety of functions during cell division , including centrosome maturation , spindle assembly and cytokinesis . Yet , these processes do not appear compromised in embryos receiving half their normal amount of Polo , while centrosome attachment to nuclei is partially defective . This suggested that centrosome-nuclei cohesion in the embryo is particularly sensitive to a decrease in Polo activity , while the other functions of Polo are satisfied with a lower Polo level . We sought to examine the effects of a more severe reduction in Polo function during syncytial embryogenesis . Because Polo is contributed to the embryo maternally , and because it is essential for viability and for meiosis , we could not examine the effects of a complete genetic inactivation of Polo on embryogenesis . To circumvent these limitations , we used chemical inhibition . We found that the Polo-like kinase 1 inhibitor BI2536 [32] is a potent inhibitor of Drosophila Polo . Treatment of D-Mel or S2 cells in culture with BI2536 phenocopied RNAi depletion of Polo , leading to a higher mitotic index and an accumulation of prometaphase cells that often displayed monopolar spindles ( data not shown ) . As predicted , treatment of embryos with BI2536 led to a high frequency of centrosome detachment , along with other defects typical of Polo inhibition in cultured cells , including incomplete or absent spindles and misaligned chromosomes ( Figure 1D , 1E ) . Not surprisingly , defective , neighbouring spindles were often fused in the syncytium ( Figure 1Db , 1E ) . However , spindle defects were less penetrant than centrosome detachments ( Figure 1E ) . All defects were almost never observed ( <1% ) in control embryos treated with DMSO alone . Altogether , our results suggest that a lower level of Polo activity is sufficient for spindle assembly and function , while proper cohesion between centrosomes and nuclei requires a higher level of Polo activity . Centrosome-nuclei cohesion is crucial to embryonic development , since it provides the link between nuclei and a skeleton of anti-parallel astral microtubules ( MTs ) that pushes the nuclei apart and towards the cortex ( Figure 2A ) , although overlapping MTs are difficult to observe [11] . Moreover , centrosomes are essential to the assembly of bipolar mitotic spindles in syncytial embryos [15] . Because the examined polo-compromised embryos are able to complete development , the dislocation between centrosomes and nuclei observed either leads to problems that occur at a tolerated rate , or is transient and does not usually lead to mitotic or nuclear migration defects . To examine these possibilities , we used time-lapse microscopy , following GFP-D-TACC as a centrosomal and spindle marker [33] , and histone H2A-RFP as a nuclear marker . In embryos from polo11/+ mothers , we could readily find several nuclei where one centrosome was dislocated from its nucleus of origin ( Figure 2B , arrowheads ) . Yet , dislocated centrosomes were recaptured during spindle assembly ( 420 s , arrows ) , and nuclear divisions could then be completed ( see also Video S1 ) . This detachment was not observed in embryos from WT mothers ( Figure 2C and Video S2 ) . These observations explain how the centrosome dislocations seen in polo-compromised embryos are transient and have no lethal consequences on their own ( Figure 2D ) . This centrosome recapture has also been observed in embryos with reduced Polo function and elevated Gwl function , although those embryos were much sicker and inviable , and many centrosomes were not recaptured [23] . That Polo is required to keep the cohesion between centrosomes and nuclei is consistent with its known functions at mitotic entry , in promoting centrosome maturation and Cdk1 activation [2] , [20] . Polo also assists Cyclin B-Cdk1 in promoting nuclear envelope breakdown [34] , [35] . In Drosophila syncytial embryos , the nuclear envelope does not completely break down in mitosis , but becomes fenestrated near centrosomes , allowing MTs to penetrate nuclei and reach chromosomes [12]–[14] . At that stage , centrosome attachment to the nuclear envelope is weakened and is replaced by MT attachments to chromosomes . Thus , the detached centrosomes observed in prophase when Polo activity is decreased could be explained by a failure to coordinate centrosome function with nuclear envelope fenestration at mitotic entry . Consistent with this idea , in polo-compromised embryos , MTs emanating from the detached centrosome can be seen as if pressing on the nuclear envelope , which caves in deeply but does not allow MT penetration at the prophase/prometaphase transition , while MTs from the attached centrosomes have already invaded the nucleus ( Figure 1C ) . The observed centrosome-nucleus cohesion defects in embryos from polo-heterozygous females suggested that it could provide a good sensitized background to screen for genes that function with polo in the embryo . We conducted such a screen using the DrosDel deficiency core kit , consisting of a sub-collection of large genomic deletions which altogether uncover approximately 60% of the fly genome [36] . Females heterozygous for the polo11 null mutation were systematically crossed to males heterozygous for a single deficiency ( Figure 3A ) . In F1 progeny , females heterozygous for both polo11 and the deficiency were tested for their ability to produce viable embryos , hatching into larvae . The same scheme was applied to test each deficiency in combination with one copy of gwlScant , a gain-of-function allele of gwl identified previously as a dominant synthetic lethal enhancer of polo hypomorphic embryos [23] , [37] . The large majority of deficiencies allowed full fertility when combined with polo11 or gwlScant . Only 6 deficiencies resulted in less than 50% embryo hatching when combined with either polo11 or gwlScant ( Figure 3B ) . Interestingly , the deficiencies that interacted with polo11 tended to also interact with gwlScant ( and vice-versa ) , further suggesting a very close functional link between Polo and Gwl . We reasoned that the deletions identified were likely to uncover genes that function with polo and gwl . Among the deficiencies identified in the screen , the strongest genetic interactor with polo and gwl was Df ( 3R ) ED5474 , which resulted in complete failure of the embryos to hatch ( Figure 3B ) . Using overlapping deficiencies , we were able to limit the interval of interest to a region containing only 15 genes on chromosome 3R ( Figure 3C ) . One of those genes was twins ( tws ) , which encodes a B-type adaptor subunit of PP2A previously implicated in cell cycle regulation [38]-[40] and is the sole ortholog of the human B55 group of adaptor subunits [41] . Combining one copy of twsP ( a strong hypomorphic allele due to a P-element insertion ) with one copy of polo11 in the maternal genotype resulted in a complete failure of embryos to hatch ( Figure 3D ) . Similar results were obtained with another allele , twsaar-1 . We did not test if mutations in the other 14 genes uncovered in the interval of interest genetically interact with polo . In any case , the genetic interaction between polo and tws pointed at a functional interdependence between Polo and PP2A in the embryo . We then tested systematically all PP2A subunit genes in Drosophila for which we could obtain mutants , for potential genetic interactions with polo in the same assay . Of the three PP2A adaptor subunit genes tested ( tws , widerborst , PP2A-B' ) only tws showed a genetic interaction with polo11 ( CG4733 ( PP2A-B'' ) was not tested ) . Females heterozygous for both polo11 and a mutant allele of microtubule star ( mts ) , the catalytic subunit gene [42] , produced semi-viable embryos ( Figure 3D ) . Similar genetic interactions were observed between polo9 , a strong hypomorph , and tws or mts ( data not shown ) . Interestingly , mts is uncovered by Df ( 2L ) ED12527 , identified as another strong hit in our screen ( Figure 3B ) . One copy of a mutant allele of Pp2A-29B , which encodes the structural subunit of PP2A had no effect in combination with polo11 . This could be explained if this subunit were to be present in excess relative to the other subunits of the holoenzyme . No genetic interactions were detected with deficiencies uncovering widerborst , PP2A-B' or Pp2A-29B ( data not shown ) . Strong genetic interactions were also observed between gwlScant and tws or mts in the same assay . The percentage of embryos hatching from mts/+; gwlScant /+ and gwlScant +/+ twsP was zero in both cases ( 0; N = 4 ) . These results point at PP2A-Tws as a critical functional interactor of Polo and Gwl during syncytial embryo development . The fact that the two strongest hits out of the 60% of the genome that was screened are subunits of PP2A-Tws highlights the specificity of the pathway uncovered here . We note that a strong genetic interaction was also detected between Df ( 2L ) ED1315 and polo11 or gwlScant; and therefore this deletion could uncover another important gene functioning with Polo , Gwl and PP2A . However , Df ( 2L ) ED1315 disrupts 94 genes , and the deficiency mapping of the gene of interest is proving challenging . The synthetic maternal-effect embryonic lethality between gwlScant and tws or mts heterozygous mutations suggested an antagonistic relationship between Gwl and PP2A-Tws . The gwlScant mutant allele of gwl leads to a K97M substitution that makes the kinase hyperactive in vitro [23] . To test directly if embryos with reduced PP2A-Tws function are sensitive to a gain in Gwl kinase activity , we overexpressed Gwl in the egg and early embryos . This was achieved by the late female germline expression of Gal4 under control of the maternal α-tubulin promoter , leading to the Gal4-driven expression of UASp-GWL in that tissue . Overexpression of Gwl in eggs/embryos from twsP heterozygous mothers was almost completely lethal , while overexpression of a kinase-dead form of Gwl had no effect on embryonic viability ( Figure 4A ) . Therefore , excessive Gwl kinase activity relative to PP2A-Tws activity disrupts either oogenesis and/or embryonic development . Fertilised eggs or embryos overexpressing Gwl and with a reduced Tws dosage were examined by immunofluorescence . Strikingly , most eggs/embryos were blocked in metaphase of meiosis I , even after 4 to 6 hours post-laying ( Figure 4Ba , 4C ) . Embryos overexpressing Gwl alone or with a reduced dose of Tws alone were almost all cellularized by that time , like WT embryos ( Figure 4B right , 4C ) . These results suggest a role for Gwl in promoting the meiotic arrest by antagonizing PP2A-Tws . Conversely , we previously showed that a loss of Gwl function in oocytes leads to a failure to arrest in meiosis , associated with unstable sister chromatid cohesion [23] . Our results are consistent with recent biochemical studies showing that Gwl promotes M-phase in Xenopus egg extracts by antagonizing PP2A-B55δ ( ortholog of Twins ) , which has itself been shown to promote M-phase exit by dephosphorylating Cdk1 substrates ( Figure 4D , discussed below ) [9] , [43] , [44] . Although most eggs overexpressing Gwl and with a reduced amount of Tws arrested in meiosis , a few eggs managed to complete meiosis and attempted to initiate embryonic development . However , theses embryos usually aborted in the first mitoses , displaying several small aberrant structures with condensed chromatin at the center of small spindles ( Figure 4Bb , 4C ) . The few mitotic nuclei observed showed a very high incidence of detached centrosomes ( Figure 4Bb , c , arrowheads ) . These observations suggest that the Gwl-PP2A-Tws pathway can also regulate mitotic divisions . We reasoned that if normal levels of Gwl regulate mitosis by antagonizing PP2A-Tws , then decreasing the activity of Gwl could rescue the viability of embryos from mts XE-2258/+; polo11/+ mothers , which are semi-viable and show severe mitotic defects ( Figure 5A , 5B ) . Indeed , those embryos were partially rescued by the introduction of one copy of the null allele gwl6a , or one copy of gwlSr18 , which abolishes the only maternally contributed splice variant of gwl ( Figure 4E ) [23] . This result strongly suggests that Gwl normally negatively regulates PP2A-Tws during the mitotic divisions in addition to meiosis in vivo . Phenotypic examination of embryos from polo-heterozygous mothers revealed an elevated frequency of transient centrosome dislocations from nuclei in prophase ( Figure 1 , Figure 5B ) . Because of the strong genetic interaction between polo and tws or mts , we examined embryos from mothers heterozygous for tws and mts mutations to reveal any potential defects . Interestingly , we detected a significant incidence of centrosome detachments in both cases . However , unlike those observed in polo-compromised embryos which occurred mostly in prophase , centrosome dislocation tended to occur in late M-phase ( between anaphase and karyokinesis ) in embryos with reduced PP2A-Tws ( Figure 5A , 5B ) . As expected , embryos combining reductions in Polo and PP2A-Tws levels displayed a strongly elevated incidence of centrosome detachments . Embryos from mothers heterozygous for polo11 and mtsXE-2258 , of which a minority were able to hatch into larvae ( Figure 3D ) , showed a high frequency of detached centrosomes at any stage of the mitotic cycles ( Figure 5A , 5B ) . Embryos from mothers heterozygous for polo11 and twsP , which all failed to hatch ( Figure 3D ) , aborted very early during syncytial divisions , with most or all centrosomes detached already in the first few cycles ( Figure 5A ) . Therefore , Polo and PP2A-Tws collaborate to ensure proper centrosome cohesion to nuclei and cell cycle progression during early embryogenesis . Since the centrosome detachments that occur upon reduction in Polo function are seen mostly in prophase while those that occur when PP2A-Tws is reduced occur mostly in late M-phase , many centrosomes in double mutants may never be able to recover their attachment and drift away , leading to a failure in nuclear division ( Figure 5C ) , as seen in embryos from polo11-Scant females [23] .
Our results shed new light on cell cycle regulation and syncytial embryogenesis . We clearly show that high Polo activity is needed to promote the normal cohesion between centrosomes and nuclei , and this is mostly observable around the time of mitotic entry . Interestingly , transiently detached centrosomes can be recaptured by the assembling spindle and nuclear division can then be completed . This centrosome recapture is probably essential for successful development of the syncytial embryo . Our systematic genetic screen unveiled a very strong and specific functional link between Polo and a specific form of PP2A associated with its B-type subunit Tws . We also show that PP2A-Tws activity is required for centrosome cohesion with nuclei , although in late M-phase , around the time of mitotic exit . This is consistent with a recent study where centrosome defects were observed in late M-phase when the small T antigen of SV40 , which binds PP2A , was expressed in Drosophila embryos [45] . PP2A-B55δ has been recently implicated in promoting mitotic exit in vertebrates , by inactivating Cdc25C and by directly dephosphorylating Cdk1 mitotic substrates [43] , [46] . The closely related isoform PP2A-B55α has been shown to promote the timely reassembly of the nuclear envelope at mitotic exit [8] . Thus , the failure to reattach centrosomes to nuclei during mitotic exit in PP2A-Tws compromised embryos could be due to problems or a delay in nuclear envelope resealing . Our results indicate that the proper regulation of the events of mitotic entry and exit by Polo and PP2A-Tws is crucial . This may be particularly true in the syncytial embryo due to the rapidity of the cycles , where one mitosis is almost immediately followed by another , and because of the obligatory cohesion between centrosomes and nuclei for their migration towards the cortex of the syncytium . Combining partial decreases in the activities of Polo and Tws strongly enhances the frequency of centrosome detachments observed ( Figure 5 ) . This suggests that when centrosomes fail to attach properly for too long between mitotic exit and the next mitotic entry , they become permanently detached from nuclei , leading to failures in mitotic divisions ( Figure 5C ) . The differences in timing between the detachments observed in polo and tws hypomorphic situations lead us to propose that the two enzymes act in parallel pathways , of which the disruption can lead to a failure in centrosome-nucleus cohesion . This is also supported by the prominent roles of Polo in regulating centrosome maturation and mitotic entry [2] , and the specific requirements of PP2A-Tws/B55 at mitotic exit . However , we cannot exclude that Polo , Gwl and PP2A-Tws could function on a common substrate , or even in the same linear pathway , where the different players of the pathway could become more or less influential at different times of the cell cycle . In has been proposed that PP2A promotes full expression of Polo in larval neuroblasts and in S2 cells [47] . It has also been shown that depletion of Tws by RNAi leads to centrosome maturation defects in S2 cells [48] , which could be explained by a reduction in Polo levels . However , we have repeatedly failed to detect a significant difference in Polo levels in embryos from gwlScant/+ or tws/+ females , compared to wild-type controls by Western blotting ( data not shown ) . Deeper genetic and molecular dissection of those pathways should lead to a clearer understanding of the regulation of centrosome and nuclear dynamics during mitotic entry and exit . Our results add strong support to an emerging model for a pathway that controls entry into and exit from mitosis and meiosis in animal cells . It is increasingly clear that a form of PP2A associated with a B-type regulatory subunit plays a crucial and conserved role in competing with Cdk1 . In Xenopus egg extract , PP2A-B55δ activity is high in interphase and low in M phase [9] . PP2A-B55δ must be down-regulated to allow mitotic entry , and conversely , it appears to promote mitotic exit both by inactivating Cdc25C and by dephosphorylating Cdk1 substrates [9] , [46] . In human cells , depletion in B55α delays the events of mitotic exit , including nuclear envelope reassembly [8] . Already some years ago , mutations in Drosophila tws were found to lead to a mitotic arrest in larval neuroblasts [38] , and extracts from tws mutants were shown to have a reduced ability to dephosphorylate Cdk substrates [49] . Mutations in mts resulted in an accumulation of nuclei in mitosis in the embryo [42] . The budding yeast now appears to be a particular case , as its strong reliance on the Cdc14 phosphatase to antagonize Cdk1 may reflect the need for insertion of the anaphase spindle through the bud neck prior to mitotic exit [50] , a constraint that does not exist in animal cells . Nevertheless , additional phosphatases to PP2A , including PP1 are likely to play conserved roles in promoting mitotic and meiotic exit , and this remains to be dissected . Our identification of PP2A genes as functional interactors of polo and gwl is the result of an unbiased genetic screen . We found that an elevation in Gwl function combined with a reduction in PP2A-Tws activity leads to a block in M phase , either in metaphase of meiosis I or in the early mitotic cycles . However , our positioning of Gwl as an antagonist of PP2A-Tws was facilitated by reports that appeared subsequent to our screen , proposing that the main role of Gwl in promoting M-phase was to lead to the inactivation of PP2A-B55δ in Xenopus egg extracts [43] , [44] . Results consistent with this idea were also obtained in mammalian cells [26] . More recently , two seminal biochemical studies using Xenopus egg extracts showed that the antagonism of PP2A-B55δ by Gwl is mediated by α-endosulfine/Ensa and Arpp19 , two small , related proteins which , when phosphorylated by Gwl at a conserved serine residue , become able to bind and inhibit PP2A-B55δ [51] , [52] . By this mechanism , Gwl activation at mitotic entry leads to the inhibition of PP2A-B55δ , which results in an accumulation of the phosphorylated forms of Cdk1 substrates . Depletion of human Arpp19 also perturbs mitotic progression in Hela cells [51] , suggesting a conserved role among vertebrates . In an independent study , the group of David Glover has recently identified mutations in Drosophila endosulfine ( endos ) as potent suppressors of the embryonic lethality that occurs when gwlScant ( the gain-of-function allele ) is combined with a reduction in polo function , in a maternal effect ( see accompanying paper by Rangone et al [53] ) . endos is the single fly ortholog of Xenopus α-endosulfine and Arpp19 . That the identification of endos by Rangone et al came from another unbiased genetic screen testifies of the specificity and conservation of the Gwl-Endos-PP2A pathway in animal cells . The authors went as far as showing that the critical phosphorylation site of Gwl in Endos is conserved between frogs and flies , and is critical for the function of Endos in antagonizing PP2A-Tws in cultured cells . These findings are consistent with a previous report showing that mutations in endos lead to a failure of oocytes to progress into meiosis until metaphase I [54] . Moreover , loss of Gwl specifically in the female germline also leads to meiotic failure , although in that case oocytes do reach metaphase I but exit the arrest aberrantly [23] . Although the meaning of those phenotypic differences is not yet understood , Gwl and Endos are both required for meiotic progression in Drosophila . Conversely , we show here that excessive Gwl activity relative to PP2A-Tws prevents exit from the metaphase I arrest , suggesting that the inhibition of PP2A-Tws by Gwl and Endos must be relieved to allow completion of meiosis . Moreover , Rangone et al show that the Endos pathway also regulates the mitotic cell cycle in the early embryo , in larval neuroblasts and in cultured cells [53] . Together , the systematic and unbiased identifications of mutations in PP2A-Tws subunit genes as enhancers ( this paper ) , and of mutations in endos as suppressors [53] of gwlScant provide strong evidence for a pathway connecting these genes to control M phase in flies . Our studies provide a convincing genetic and functional validation of the recent biochemical results from Xenopus extracts , and show that the Gwl-Endos-PP2A-Tws/B55 pathway is conserved and plays a key role in regulating both meiosis and mitosis in a living animal .
Flies were kept at 25°C on standard food . The wild-type strain used was Oregon R . For the genetic screen , the DrosDel Core deletion kit [36] ( obtained from John Roote , Cambridge , UK ) was used and uncovered approximately 60% of the genome with 200 deficiencies . For fertility tests , 3 to 5 well-fed , 1 to 4 days-old virgin females were given 3 to 5 Oregon R males per tube and allowed to mate for one day . Flies were then transferred on grape juice-containing agar and yeast . After one day , flies were removed ( usually transferred to a new tube ) . Between 24 and 30 hours later , the percentage of hatched embryos was counted . At least 100 embryos were counted . polo11 , gwlScant and gwlSr18 alleles were previously published [23] , [37] . twsP ( twsJ11C8 ) and twsaar-1 were from David Glover . mtsXE-2258 , Pp2A-29BEP2332 , PP2A-B'A131 , and wdb07 were from Bloomington stock center . GFP-D-TACC and H2A-RFP stocks were kindly provided by Pat O'Farrell and Jordan Raff . UASp-GWL-MYC and UASp-GWL-KD-MYC ( K87R ) expressed the long splice variant of Gwl , and were made in pPWM ( Drosophila Genomics Resource Center ) . Transgenic flies were generated using the P-element-based method by BestGene Inc ( Chino Hills , CA , USA ) . UASp-GWL flies were reported previously [23] . Overexpression in the early embryo was driven by maternal α-Tubulin Gal4 , which was obtained from Adelaide Carpenter . For immunofluorescence , embryos were collected on grape juice agar , and dechorionated and fixed as described [23] . Antibodies used for stainings are: α-Tubulin: YL1/2 ( 1:50; Serotec ) , γ-Tubulin: GTU-88 ( 1∶50; Sigma ) and lamin B/Lamin Dm0 ( 1∶100; Developmental Studies Hybridoma Bank ) . Secondary antibodies were coupled to Alexa-488 ( 1∶200; Invitrogen ) or Texas red ( 1∶200; Invitrogen ) . DNA was marked with DAPI . Images were acquired on a Laser scanning confocal microscope LSM 510 Meta ( Zeiss ) , using a 100X oil objective . Embryos from WT or polo11/+ females and expressing GFP-D-TACC and H2A-RFP were dechorionated and imaged on a Swept-Field confocal microscope ( Nikon Eclipse Ti ) , using a 100X oil objective . Chemical treatments of embryos used a protocol modified from Sibon et al [29] . Embryos were dechorionated and incubated for 30 min in a 1∶1 mixture of Express Five Drosophila cell culture medium ( Invitrogen ) and heptane with 1 µM of BI2536 ( from a DMSO stock solution ) or DMSO alone . | The development and survival of all living organisms relies on the fine regulation of cell division at the molecular level . This coordination depends on kinases and phosphatases , enzymes that catalyze the addition and removal of phosphate groups on specific target proteins . The genes encoding these enzymes have been largely conserved between species during evolution . In a previous paper published in PLoS Genetics , we found an antagonism between the Polo and Greatwall mitotic kinases in the fruit fly model . In this study , we have used fly genetics to identify additional genes that function with polo and greatwall during early embryogenesis . We have found a specific form of the Protein Phosphatase 2A ( PP2A-Tws ) that collaborates with the Polo kinase at a stage when multiple nuclei rapidly divide in a large , single-cell early embryo . We found that Polo and PP2A-Tws are both required for the proper cohesion between nuclei and the centrosomes , which are essential structures for mitosis and embryonic development . We also found that the Greatwall kinase antagonizes the PP2A-Tws phosphatase to promote mitosis and meiosis . Our genetic study sheds new light on cell cycle regulation and is consistent with recent results from biochemical studies using frog cell extracts . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"meiosis",
"cellular",
"structures",
"gene",
"function",
"animal",
"models",
"mitosis",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"cell",
"division",
"molecular",
"development",
"molecular",
"genetics",
"cell",
"nucleus",
"cytoskeleton",
"biology",
"cell",
"biology",
"genetic",
"screens",
"genetics",
"molecular",
"cell",
"biology",
"gene",
"networks",
"genetics",
"and",
"genomics"
]
| 2011 | PP2A-Twins Is Antagonized by Greatwall and Collaborates with Polo for Cell Cycle Progression and Centrosome Attachment to Nuclei in Drosophila Embryos |
Next-generation sequencing of the exome and genome of prostate cancers has identified numerous genetic alternations . SPOP ( Speckle-type POZ Protein ) was one of the most frequently mutated genes in primary prostate cancer , suggesting SPOP is a potential driver of prostate cancer development and progression . However , how SPOP mutations contribute to prostate cancer pathogenesis remains poorly understood . SPOP acts as an adaptor protein of the CUL3-RBX1 E3 ubiquitin ligase complex that generally recruits substrates for ubiquitination and subsequent degradation . ER-localized isoform of the formin protein inverted formin 2 ( INF2 ) mediates actin polymerization at ER-mitochondria intersections and facilitates DRP1 recruitment to mitochondria , which is a critical step in mitochondrial fission . Here , we revealed that SPOP recognizes a Ser/Thr ( S/T ) -rich motif in the C-terminal region of INF2 and triggers atypical polyubiquitination of INF2 . These ubiquitination modifications do not lead to INF2 instability , but rather reduces INF2 localization in ER and mitochondrially associated DRP1 puncta formation , therefore abrogates its ability to facilitate mitochondrial fission . INF2 mutant escaping from SPOP-mediated ubiquitination is more potent in prompting mitochondrial fission . Moreover , prostate cancer-associated SPOP mutants increase INF2 localization in ER and promote mitochondrial fission , probably through a dominant-negative effect to inhibit endogenous SPOP . Moreover , INF2 is important for SPOP inactivation-induced prostate cancer cell migration and invasion . These findings reveal novel molecular events underlying the regulation of INF2 function and localization , and provided insights in understanding the relationship between SPOP mutations and dysregulation of mitochondrial dynamics in prostate cancer .
Large-scale exome/genome sequencing studies have recently revealed that recurrent mutations in the SPOP gene occur in up to 15% of prostate cancers [1–4] . Interestingly , the SPOP mutant subset of prostate cancers had some notable molecular features , including mutual exclusivity with ERG gene rearrangement , elevated levels of DNA methylation , homogeneous gene expression patterns , frequent deletion of CHD1 and overexpression of SPINK1 mRNA , supporting the concept that SPOP mutation tumors represent a distinct molecular subclass of prostate cancer [4] SPOP is one of the adaptor proteins of the CUL3-RBX1 E3 ubiquitin ligase complexes . It selectively recruits substrates via its N-terminal MATH domain , whereas its BTB and BACK domains mediate oligomerization and interaction with CUL3 [5] . SPOP has been linked to the ubiquitination and degradation of several substrates , including the steroid receptor coactivator 3 ( SRC-3 ) , androgen receptor ( AR ) , DEK , ERG , SENP7 and several others [6–11] . All prostate cancer-associated SPOP mutations identified so far affect evolutionarily conserved residues in the MATH domain , suggesting that these mutations may alter the interaction of SPOP with its substrates [1–4] . Inactivation of SPOP by knockdown or overexpression of prostate cancer-associated SPOP mutants leads to increased prostate cancer cell proliferation , migration and invasion , implying SPOP is a tumor suppressor [2 , 8–10] . However , limited numbers of SPOP substrates have been identified and functionally explored . Mitochondria are highly motile organelles that undergo constant fission and fusion , and are actively transported to specific subcellular locations [12] . Unbalanced mitochondrial fission and fusion events are associated with mitochondrial dysfunction and frequently linked to the pathogenesis of many human diseases , including cancer [12 , 13] . The majority of studies that have explored mitochondrial morphology in tumor cells support a pro-tumorigenic role for mitochondrial fission and tumor suppressor role for mitochondrial fusion [14] . Mitochondrial fragmentation has been observed in various types of tumor cells [15–17] . Inhibition of mitochondrial fission decreases cell proliferation , migration and invasion in various cancer models including lung , colon , breast , thyroid cancer and glioblastoma[16–20] . While cancer is a disease characterized by multiple genetic aberrations , little is known about whether cancer-associated mutations can directly affect mitochondrial dynamics , and how this impacts upon tumor phenotypes . Inverted formin 2 ( INF2 ) is a unique vertebrate formin protein that accelerates both actin polymerization and depolymerization [21] . In mammalian cells , INF2 can be expressed as two C-terminal splice variants: the prenylated ( CAAX ) isoform , which is tightly bound to endoplasmic reticulum ( ER ) [22] , and the nonCAAX isoform , which is cytoplasmic [23] . Recent studies have persuasively showed in mammalian cells that actin polymerization mediated by ER-localized INF2 CAAX isoform is required for mitochondrial fission [24] . By contrast , the cellular function of the nonCAAX isoform of INF2 has been less characterized . Suppression of INF2-nonCAAX isoform in cells causes Golgi dispersal , suggesting INF2 might be involved in maintenance of Golgi architecture [23] . Mutations in INF2 are linked to two human genetic diseases: focal and segmental glomerulosclerosis ( FSGS ) , a degenerative kidney disease [25] , and Charcot-Marie-Tooth disease ( CMTD ) , a neurological disorder [26] . However , little is known about how INF2 protein is physiologically regulated . In this study , we demonstrate that SPOP suppresses mitochondrial fission by promoting atypical ubiquitination and relocalization of ER-localized INF2 . Moreover , this effect is abrogated by the prostate cancer-associated SPOP mutations . Thus , our results provide a functional link between SPOP mutations and dysregulation of mitochondrial dynamics in prostate cancer .
To identify molecular mediators of the tumor suppressive function of SPOP , we performed a yeast two-hybrid screen in a human fetal brain cDNA library using the full length SPOP as bait . Among the positive clones identified , 4 clones were INF2 fragments . Considering INF2 is an important regulator of actin polymerization and mitochondrial fission , we explored whether INF2 is an authentic SPOP substrate and its function is dysregulated in SPOP-mutated prostate cancer . We first examined whether SPOP interacts with INF2 in cells . To do this , FLAG-INF2 , and Myc-SPOP were co-expressed in 293T cells . Cell lysates were subsequently prepared for co-immunoprecipitation ( co-IP ) with anti-FLAG antibody . As shown in Fig 1A , Myc-SPOP was immunoprecipitated by FLAG-INF2 , suggesting an interaction between SPOP and INF2 proteins . Similar results were also obtained in the reciprocal co-IP experiment in which FLAG-SPOP was able to immunoprecipitate Myc-INF2 ( Fig 1B ) . FLAG-SPOP was able to immunoprecipitate endogenous INF2 , and two known SPOP substrates ( AR and DEK ) in LNCaP cells ( Fig 1C ) . Next , we decided to extend our analysis by investigating whether endogenous SPOP and INF2 can interact with each other in prostate cancer cells . Immunoprecipitation using anti-INF2 antibody was performed using cell lysates prepared from LNCaP cells . As shown in Fig 1D , INF2 was able to immunoprecipitate SPOP and a known interactor IQGAP1 , suggesting that SPOP can interact with INF2 protein at endogenous level . SPOP contains two structural domains: a substrate-binding MATH domain at the N-terminus and a CUL3-binding BTB domain at the C-terminus . To determine which domain may mediate its interaction with INF2 , we generated two deletion mutants of SPOP ( SPOP-ΔBTB and ΔMATH ) , corresponding to the deletion of these two domains respectively ( Fig 1E ) . Co-IP assay was performed to examine the binding of INF2 with the full length SPOP ( SPOP-WT ) and the two deletion mutants . As shown in Fig 1F , SPOP-WT and SPOP-ΔBTB , but not SPOP-ΔMATH interacted with INF2 . Therefore , our findings demonstrate that SPOP binds INF2 via the MATH domain . We then explored whether SPOP can promote the ubiquitination and degradation of INF2 . Unexpectedly , overexpression of wild-type SPOP or its mutants ( SPOP-ΔBTB , ΔMATH ) did not alter the protein level of ectopically co-expressed INF2 ( Fig 2A ) . Moreover , we found that ectopic expression of SPOP in LNCaP or DU145 prostate cancer cells did not alter the protein level of endogenous INF2 ( Fig 2B ) . In contrast , SPOP overexpression in LNCaP cells ( AR positive ) decreased the expression of endogenous AR , a known SPOP substrate ( Fig 2B ) [7 , 27] . Consistent with these findings , depletion of endogenous SPOP by two independent shRNAs did not alter INF2 protein level in both LNCaP and DU145 ( AR-negative ) cells , but elevated AR protein level in LNCaP cells ( Fig 2C ) . Thus , these results demonstrate that SPOP does not affect INF2 protein level . To determine whether SPOP regulates INF2 polyubiquitination , HA-Ub and FLAG-INF2 were co-expressed in 293T cells with increasing doses of SPOP-WT or its mutants ( SPOP-ΔBTB , ΔMATH ) . As shown in Fig 2D , INF2 protein was robustly polyubiquitinated by co-expression of SPOP-WT , but not SPOP-ΔBTB or ΔMATH , in a dose-dependent manner . Accordingly , depletion of SPOP in LNCaP cells decreased the ubiquitination of endogenous INF2 ( Fig 2E ) . Since the INF2 construct used in above analysis is the CAAX isoform , we examined whether SPOP can ubiquitinate the nonCAAX isoform . As shown Fig 2F , INF2 nonCAAX isoform was also robustly polyubiquitinated by SPOP . Taken together , our data suggest that SPOP can promote INF2 ubiquitination , but not degradation . We then examined the linkage specificity of SPOP-mediated INF2 ubiquitination . In vivo ubiquitination assay was performed using a panel of ubiquitin mutants containing a single K/R mutation on each of the seven lysines in the ubiquitin sequence , potentially involved in the formation of polyUb chains . We also included a lysine-free ubiquitin mutant ( K-ALL-R ) , in which all of the lysines were replaced with arginines . As shown in Fig 2G , expression of the K-ALL-R mutant abolished SPOP-mediated INF2 ubiquitination , excluding the possibility that SPOP promotes multiple mono-ubiquitination of INF2 . Expression of K6R or K11R mutant marginally altered the amount of ubiquitinated INF2 ( Fig 2G ) , suggesting that K6 and K11 are largely dispensable for SPOP-mediated INF2 ubiquitination . By contrast , a significant reduction of INF2 ubiquitination is instead observed when other ubiquitin mutants , including K27R , K29R , K33R , K48R and K63R , were used ( Fig 2G ) . We next used a reciprocal series of mutants , where all the seven lysines in ubiquitin were converted to arginine residues , except one ( one-K-Only mutants ) . As shown in Fig 2H , expression of K27O , K29O , K33O , K48O or K63O mutants completely abolished SPOP-mediated INF2 ubiquitination . Therefore , these data indicate that SPOP catalyzes synthesis of mixed-linkage polyUb chains on INF2 , and K27 , K29 , K33 , K48 and K63 residues in Ub are all essentially involved . Having established that SPOP promotes atypical ubiquitination of INF2 , we set out to identify the ubiquitin attachment sites on INF2 . We co-expressed the FLAG-INF2 , Myc-SPOP and HA-Ub constructs in 293T cells , and the immunoprecipitated ubiquitin-INF2 conjugates were analyzed by liquid chromatography tandem mass spectrometry ( LC-MS/MS ) . It revealed that INF2 was ubiquitinated at least at 7 lysine residues ( Fig 2I ) . Interestingly , 5 of 7 ubiquitin attachment sites are localized in a short stretch of sequence ( amino acids 612–682 ) within the FH2 domain of INF2 ( Fig 2J ) . To evaluate whether this region is important for INF2 ubiquitination , we constructed a series of INF2 deletion mutants and performed in vivo ubiquitination assay . While these deletion mutants were capable of binding to SPOP in a manner similar to the full length INF2 ( Fig 2K ) , the N2 and ΔInter mutants , which lack the 612–682 aa region , were much less ubiquitinated by SPOP ( Fig 2l ) . These data suggest that the lysine residues located in the 612–682 aa of INF2 serve as the predominant ubiquitin attachment sites . Previous studies reported that one or several SBC motifs ( Φ-π-S-S/T-S/T; Φ: nonpolar residues , π: polar residues ) are present in known SPOP substrates [6–11 , 28] . We examined the protein sequence of INF2 that is required for SPOP-binding . To this end , we first deduced the minimal interacting region from the four INF2 fragments obtained in yeast two-hybrid screen . We found INF2 ( 1024~1249 aa ) corresponds to the smallest region necessary for SPOP interaction ( Fig 3A ) . Next we performed a protein motif search in the C-terminal region of INF2 and discovered a perfectly matched SBC motif ( Fig 3A ) . Moreover , this motif is very similar to the SBC motifs present in MacroH2A , DAXX and DEK ( Fig 3B ) . To examine whether this potential motif is actually required for SPOP-INF2 interaction , we generated an INF2 mutant in which the motif sequence was deleted . 293T cells were co-transfected with SPOP and wild-type INF2 or ΔSBC mutant . Co-IP assay demonstrated that SPOP only bound to the wild-type INF2 , but not the ΔSBC mutant although they were expressed at comparable levels ( Fig 3C ) , suggesting that the SBC motif of INF2 was required for SPOP binding . In vivo ubiquitination assay demonstrated that deletion of the SBC motif totally abolished SPOP-mediated INF2 ubiquitination ( Fig 3D ) . Collectively , we have identified a conserved SBC motif present in INF2 that is indispensable for SPOP-INF2 interaction . All the SPOP mutations detected thus far in prostate cancers exclusively occur in the MATH domain , which is responsible for substrate binding ( Fig 4A ) . We postulated that prostate cancer-associated mutants of SPOP may be defective in mediating INF2 polyubiquitination . To test this , we generated a series of Myc-tagged prostate cancer-associated mutants of SPOP , including Y87C , Y87N , F102C , S119N , F125V , K129E , W131G , W131C , F133L , F133V and K134N , and examined their interactions with INF2 by co-IP assays . As shown in Fig 4B , mutations of the residues at the MATH domain abrogated the ability of SPOP to interact with INF2 . Moreover , in vivo ubiquitination assay indicated that prostate cancer-associated SPOP mutants largely lost the capacity to promote INF2 polyubiquitination ( Fig 4C ) . Previous study showed that only one copy of SPOP allele is mutated in prostate cancer and SPOP mutants exert their tumor-promoting function in a dominant-negative manner to inhibit the wild-type SPOP [2] . We hypothesized that prostate cancer-associated mutations of SPOP might disrupt the interaction between wild-type SPOP and INF2 . Indeed , we found that co-expression of SPOP mutants ( Y87N , F125V or F133L ) reduced the interaction between wild-type SPOP and INF2 ( Fig 4D ) . Moreover , co-expression of SPOP mutants suppressed wild-type SPOP-induced INF2 ubiquitination ( Fig 4E ) . Taken together , our findings suggest that INF2 ubiquitination may be dysregulated by oncogenic prostate cancer-associated SPOP mutants . INF2 ( CAAX isoform ) is ER-anchored and INF2-mediated actin assembly is specially triggered at ER-mitochondrial intersections to ensure mitochondrial division [24] . Previous study showed that INF2-CAAX isoform was ER membrane-bound , but a pools of INF2 was cytosolic . [22] . SPOP was originally named as speckle-type POZ protein since ectopically expressed SPOP in COS-7 cells primarily exhibited a discrete speckled pattern in the nucleus [29] . Through quantitative analysis , we found that SPOP was localized exclusively in the nucleus as speckles in approximately 70% cells , but in both the cytoplasm and nucleus in the rest 30% cells , indicating that SPOP shuttles between cytoplasm and nucleus in a proportion of cells ( S1 Fig ) . Thus , we hypothesized that SPOP-INF2 interaction occurs in the cytoplasm and SPOP-mediated atypical ubiquitination may regulate the subcellular localization of INF2 . To test this hypothesis , we co-expressed GFP-tagged INF2 ( CAAX isoform ) and mApple-tagged Sec61β ( an ER marker ) in cells . We found that these two proteins were perfectly co-localized ( Fig 5A ) , confirming that INF2 CAAX isoform is ER-localized . However , in approximately 30% cells that HA-SPOP was localized in both the cytoplasm and nucleus , INF2 was primarily present as speckles in the cytoplasm and co-localized with SPOP , but not Sec61β ( Fig 5A ) . In contrast , in the rest 70% cells that HA-SPOP was localized exclusively in the nucleus , INF2 was still co-localized Sec61β ( Fig 5A ) . These results suggest that SPOP can inhibit the ER localization of INF2 , but this activity strictly depends its cytoplasmic localization . Moreover , deletion of the SBC motif ( ΔSBC ) or the region containing main ubiquitination sites ( ΔInter ) in INF2 did not alter its localization in ER ( S2 Fig ) , but SPOP-induced speckle pattern of INF2 in the cytoplasm was not observed ( S2 Fig ) , suggesting that SPOP-INF2 interaction and SPOP-induced INF2 ubiquitination are both required for INF2 localization outside of ER . Next , we investigated the impact of prostate cancer-associated mutants of SPOP on INF2 localization . To this end , we focused on three hotspot mutations Y87N , F125V and F133L . Interestingly , these mutants were exclusively localized as nuclear speckles in nearly 100% cells ( S1 Fig ) , implying that cytoplasmic retention ability of SPOP may be impaired by prostate cancer-associated mutations . Accordingly , we found that ectopic expression of SPOP mutants had no obvious effect on the ER localization of INF2 by immunofluorescence analysis ( Fig 5A ) . We used ER fractionation methods as a second method to corroborate the immunofluorescence analysis . As shown in Fig 5B , overexpression of wild-type SPOP , but not the prostate cancer-associated mutants of SPOP , reduced the protein amounts of GFP-INF2 in ER fractions . Lastly , we investigated whether SPOP would affect the localization of endogenous INF2 . As shown in Fig 5C , in a proportion of SPOP-WT-transfected cells , endogenous INF2 was present in cytoplasmic speckles and co-localized with SPOP , but this effect was not observed in cells expressing SPOP mutants . ER fractionation experiments demonstrated that stably overexpression of wild-type SPOP reduced the protein amounts of endogenous INF2 in ER fractions ( Fig 5D ) . In contrast , overexpression of SPOP mutants moderately increased the protein amounts of endogenous INF2 in ER fractions ( Fig 5D ) , probably those acting through a dominant-negative effect to inhibit endogenous SPOP . Taken together , our data suggests that wild-type SPOP , but not prostate cancer-associated mutants , can promote INF2 disassociation from ER . Considering that actin polymerization between mitochondria and INF2-enriched ER membranes is a critical step in mitochondrial fission [24] , we reasoned that SPOP might suppress mitochondrial fission by inhibiting INF2 localization in ER . To test this , DU145 cells were infected with lentivirus expressing wild-type SPOP or prostate cancer-associated SPOP mutants . The mitochondrial morphology was monitored by Mitotracker Red dye . As shown in Fig 6A and 6B , stably overexpression of HA-SPOP resulted in significant increases in mitochondrial average length , accompanying with endogenous INF2 speckles in cytoplasm . However , this effect was only observed in approximately 30% cells that HA-SPOP was localized in both the cytoplasm and nucleus , but not in those cells that HA-SPOP was exclusively localized in nucleus ( Fig 6A and 6B ) . These data suggest that SPOP-mediated suppression of mitochondrial fission strictly depends on its cytoplasmic localization . In contrast , the prostate cancer-associated SPOP mutants ( SPOP-Y87N , F125V and F133L ) lost the capacity to suppress mitochondrial fission monitored by immunofluorescence ( Fig 6A ) . Statistical analysis showed stably overexpression SPOP mutants even resulted in moderate decreases in mitochondrial average length probably those acting through a dominant-negative effect to inhibit endogenous SPOP ( Fig 6B ) . Previous study reported that the constitutive active mutant INF2-A149D can decreased mitochondrial length [24] . We found that the INF2-A149D-ΔSBC mutant , which can escape from SPOP-mediated ubiquitination , is more potent in decreasing mitochondria average length than INF2-A149D ( Fig 6C ) . Consistent with these findings , depletion of SPOP in DU145 cells resulted in decreases in mitochondria average length ( Fig 6D ) . Moreover , Co-depletion of DRP1 and SPOP by shRNAs reversed the effect of SPOP single depletion on mitochondria size ( Fig 6D ) . Thus , SPOP inactivation-induced mitochondrial fission occurs upstream of DRP1 . Taken together , our data suggest that wild-type SPOP , but not prostate cancer-associated mutants , can suppress INF2-mediated mitochondrial fission . Our above data indicated that SPOP regulates INF2-mediated mitochondrial fission strictly depends on its cytoplasmic localization , but the nuclear-cytoplasmic shuttling mechanism of SPOP was still poorly understood . It is clear that import of large proteins is generally mediated by nuclear localization signals ( NLS ) , which contain basic amino acids [30] . SPOP contains an evolutionarily conserved NLS sequence at its C-terminus ( S3A Fig ) . We found that SPOP lacking the NLS sequence ( SPOP-ΔNLS ) accumulated exclusively in the cytoplasm as puncta pattern and perfectly co-localized with GFP-INF2 ( S3B Fig ) . In contrast , two prostate cancer-associated SPOP mutants lacking the NLS sequence ( SPOP-F125V-ΔNLS , SPOP-F133L-ΔNLS ) accumulated exclusively in the cytoplasm as puncta pattern similar as SPOP-ΔNLS , but these mutants did not co-localize with GFP-INF2 , possibly due to impaired interaction with INF2 ( S3B Fig ) . Moreover , SPOP-ΔNLS cannot alter the ER localization of INF2-ΔSBC and INF2-ΔInter mutants ( S3C Fig ) , suggesting that SPOP-INF2 interaction and SPOP-induced INF2 ubiquitination are required for INF2 localization outside of ER . Proteins containing classic NLS are known to be transported into the nucleus by forming complexes with shuttling carriers , such as Karyopherin-alpha and-beta ( KPNA and KPNB ) [30] . Our yeast two-hybrid screen identified several clones corresponding to KPNA5 ( importin subunit alpha-6 ) . Indeed , deletion of the NLS sequence totally abolished the interaction between SPOP ( WT , F125V , F133L ) and overexpressed or endogenous KPNA5 ( S3D Fig ) , suggesting that KPNA5 might participate in nuclear transport of wild-type and prostate cancer-associated SPOP mutants . Not surprisingly , we found that SPOP-ΔNLS was able to immunoprecipitate more endogenous INF2 than SPOP-WT ( S3E Fig ) , and SPOP-ΔNLS was more effective to promote INF2 ubiquitination than SPOP-WT ( S3F Fig ) . It has been reported that INF2 promotes mitochondrial fission controls mitochondrial assembly of DRP1 [24] . DRP1 localized to cytoplasm and to mitochondrially associated puncta in cells . Depletion of INF2 reduced mitochondrially associated puncta , in addition to causing mitochondrial elongation[24] . We observed that SPOP-ΔNLS overexpression reduced DRP1 puncta associated with mitochondria and increased mitochondria length more efficient than SPOP-WT ( Fig 7A , 7B and 7C ) . The levels of DRP1 in purified mitochondrial fractions from SPOP-ΔNLS overexpressing cells were also lower than those from SPOP-WT overexpressing cells ( S3G Fig ) . In contrast , overexpression of SPOP-ΔBTB or ΔMATH mutant had no impact on DRP1 puncta , and mitochondria length ( Fig 7A , 7B and 7C ) . It is not surprising since INF2 protein cannot be polyubiquitinated by SPOP-ΔBTB or ΔMATH . Taken together , our data confirmed that cytoplasmic retention of SPOP is required for its regulation of mitochondrial fission . To determine the biological importance of SPOP regulation of INF2-mediated mitochondrial fission , we first used two independent shRNAs ( #1 targets total INF2 , #2 targets INF2 CAAX isoform only ) to knock down INF2 expression . Consistent with previous studies , INF2 depletion in LNCaP or DU145 cells resulted in a significant increase in mitochondrial average length ( S4 Fig ) . However , this change was not associated with major change in mitochondrial function , as the basal mitochondrial reactive oxygen species ( ROS ) production ( Fig 8A ) , oxygen consumption rate ( OCR ) ( Fig 8B ) , and membrane potential ( Fig 8C ) were not significantly altered following INF2 depletion . Moreover , we found that INF2 depletion marginally affected the cell cycle progression ( Fig 8D ) or overall cell growth ( Fig 8E ) . These results led us to explore other cancer cell phenotypes affected by INF2 depletion . Recently , emerging evidence supports a role for mitochondrial dynamics in tumor cell migration and invasion in various cancer models [16–20] . Indeed , we found that depletion of INF2 in DU145 cells markedly decreased cell migration and invasion ( Fig 8F and 8G ) . In contrast , depletion of SPOP enhanced cell migration and invasion ( Fig 8F and 8G ) . More importantly , co-depletion of SPOP and INF2 reduced cell migration and invasion compared with depletion of SPOP only ( Fig 8F and 8G ) . Similar results were obtained when we used SPOP-F133L mutant overexpression to replace knockdown of SPOP by shRNA ( S5A and S5B Fig ) . Previous studies demonstrated that INF2 functions upstream of DRP1[24] . We found that treatment with DU145 cells with DRP1 selective inhibitor Mdivi-1 or knockdown of DRP1 significantly reduced SPOP depletion-enhanced cell migration and invasion ( Fig 8H and 8I; S5C and S5D Fig ) . Similar effects were observed in another prostate cancer cells LNCaP ( S6 Fig ) . Together , our data suggests that SPOP suppresses prostate cell migration and invasion , at least in part , by regulating INF2-mediated mitochondrial fission .
Although SPOP mutation is now recognized as a distinct molecular feature in a subtype of prostate cancer , the underlying mechanisms remain poorly understood [4] . Previous studies showed that SPOP inactivation increased cell proliferation primarily in AR-positive prostate cancer cells , but increased prostate cell migration and invasion in an AR-independent manner [2 , 9 , 10] . These effects were partly dependent on stabilization of SPOP substrates such as AR and ERG [9 , 10] . ERG up-regulation leads to transactivation of its target genes , including ADAMTS1 , CXCR4 , OPN and MMP9 , all of which play important roles in promoting cell migration and invasion [9 , 10] . In this study , we revealed that the ER-localized isoform of the INF2 is ubiquitinated and regulated by SPOP . SPOP inactivation-induced prostate cancer cell migration and invasion is partly mediated by INF2 and mitochondrial fission ( Fig 9 ) . In the past few years , there is accumulating evidence that mitochondrial fission and fusion play active roles in regulation of cell movement , migration and invasion [14 , 31] . For example , there are higher levels of DRP1 and less Mfn1 ( a GTPase for mitochondrial fusion ) in the metastatic breast cancer cells compared with non-metastatic breast cancer cells . 18 Silencing DRP1 or overexpression of Mfn1 results in mitochondrial fusion , and significantly suppresses migration and invasion abilities of breast cancer cells [18] . Similar effect has been detected in glioblastoma and lung or thyroid cancer[17 , 19 , 20] . The possible mechanism for mitochondrial fission-enhanced cell movement is that mitochondria are usually trafficked to sites of high-energy demand , and in migrating cells , mitochondria are more frequently located at their leading edge where demands high energy [14 , 32] . Our study , for the first time , links prostate cancer-associated SPOP mutations to mitochondrial dynamics-related cell migration and invasion . Interestingly , a recent study demonstrated that aberrant activation of MAPK signaling by K-Ras ( G12V ) mutation in pancreatic cancer activates DRP1 via ERK-mediated phosphorylation , and DRP1-meditated mitochondrial fission is crucial for Ras-driven transformation [33] . Similarly , BRAF ( V600E ) , the most common mutation in melanoma , correlates with DRP1 phosphorylation in melanoma tumor tissues , whereas MAPK inhibition reverses DRP1-mediated mitochondrial fission , and sensitizes cells to mitochondrial-targeting drugs [34] . Therefore , cancer-associated mutations may promote mitochondrial fission through multiple signaling pathways in different tumors . It also should be noted that SPOP can ubiquitinate the cytoplasmic INF2 non-CAAX isoform similar as the ER-localized INF2 CAAX isoform ( Fig 2F ) . A recent study revealed that a proportion of cytoplasmic INF2 was localized in focal adhesion ( FA ) and the protruding edge of migrating cells [35] . We cannot rule out the possibility that SPOP-mediated ubiquitination of INF2 non-CAAX isoform also affects cell migration and invasion . Ubiquitination has critical functions in nearly all aspects of biological processes . Although ubiquitination is traditionally thought to only target proteins for degradation , recent studies suggest additional roles of ubiquitination in nonproteolytic functions involved in protein function regulation [36] . It is well known that K48-linked polyUb ubiquitin chains are sufficient to target substrates to the 26S proteasome for degradation and that K63-linked polyUb ubiquitin chains have been demonstrated to regulate a variety of nonproteolytic cellular functions , though the roles of other atypical ubiquitin linkages through M1 , K6 , K11 , K27 , K29 or K33 or mixed linkages within the same chain remain poorly understood [36] . Previous studies demonstrated that the mitochondrial ubiquitin ligase MITOL regulates mitochondrial-ER membrane bridges through K63-linked ubiquitination of mitochondrial Mfn2 ( a GTPase for mitochondrial fission ) , suggesting that atypical ubiquitination plays roles in mitochondrial dynamics [37] . In this study , we demonstrated SPOP catalyzes synthesis of mixed-linkage polyUb chains ( K27 , K29 , K33 , K48 and K63 ) on INF2 , which does not trigger INF2 degradation . Instead , these forms of ubiquitination cause INF2 dissociation from ER and impair its ability to promote mitochondrial fission . Until now , the majorities of known SPOP substrates are ubiquitinated and degraded by SPOP . But a previous study showed that SPOP is able to ubiquitinate the PcG protein BMI1 and the histone variant MacroH2A . These ubiquitinations do not affect the overall stability of BMI1 or MacroH2A , but facilitates PcG-mediated transcriptional repression and deposition of MacroH2A during stable X chromosome inactivation process [38] . These data and others reinforce a notion that SPOP can promote both degradative or non-degradative ubiquitination towards different substrates . Moreover , it is also possible that unknown deubiquitinase ( s ) might exist to recycle INF2 from cytoplasmic speckles to ER . Another interesting aspect of our work that needs further investigation is the potential molecular mechanism that accounts for nuclear-cytoplasmic shuttling of SPOP . One study demonstrated that hypoxia condition promotes SPOP cytoplasmic accumulation in clear cell renal cell carcinoma ( ccRCC ) cells [39] . However , our preliminary results found that hypoxia treatment did not affect SPOP localization at least , in prostate cancer cells ( S7 Fig ) . Considering that SPOP-mediated suppression of mitochondrial fission is strictly dependent on its cytoplasmic localization , elucidation the molecular mechanisms of cytoplasmic accumulation of SPOP is an important direction to pursue in the future . Moreover , three prostate cancer-associated SPOP mutants ( Y87N , F125V and F133L ) nearly lost their cytoplasmic localization compared with wild-type SPOP . It is possible that these mutations impair the capacity of SPOP to interact with proteins which facilitate cytoplasmic retention of SPOP . Our results showed that deletion of the NLS sequence forced prostate cancer-associated SPOP mutants to localize in cytosol as puncta , but these mutants cannot alter the ER localization of INF2 like SPOP-WT ( Fig 7B ) . So it is possible that the direct interaction with some cytoplasmic substrates of SPOP , including but not limited to INF2 , may cause a pool of SPOP to accumulate in cytosol by blocking access to Importin proteins . Prostate cancer-associated SPOP mutants lost the capacity to interact with its cytoplasmic binding partner , and localized exclusively in the nucleus . Taken together , our data suggest SPOP might exert its tumor-suppressive roles both in nucleus and cytoplasm .
293T , HeLa cells and prostate cancer cell lines ( LNCaP , DU145 , PC-3 ) were obtained from the American Type Culture Collection ( ATCC ) . 293T and HeLa cells were maintained in DMEM with 10% ( v/v ) FBS . LNCaP and DU145 cells were maintained in DMEM with 10% ( v/v ) FBS . All cells were grown at 37°C with 5% CO2 . Expression vectors for SPOP-WT or mutants are described previously . FLAG-INF2-CAAX was obtained from Dr . Miguel Angel Alonso ( Universidad Autónoma de Madrid ) . INF2 mutants were generated by KOD-Plus-Mutagenesis Kit ( TOYOBO ) following the manufacturer’s instructions . For WB detection of ER-localized INF2 from HeLa cells , the microsomal fraction from approximately 5×106 HeLa cells was prepared using an ER extraction kit ( ER0100 , Sigma-Aldrich ) . The mitochondrial fraction was prepared using a Mitochondria Isolation Kit ( MitoISO1 , Sigma-Aldrich ) . The pLKO . 3G GFP-shRNA plasmids were purchased from Addgene . The shRNA sequence of sh-SPOP#1: 5’-GGAGAACGCUGCAGAAAUU-3’; sh-SPOP#2: 5’-ATAAGTCCAATAACGACAGGC-3’; shINF2-#1: 5’- CCCUCUGUGGUCAACUACU-3’; shINF2-#2 ( target to CAAX isoform only ) : 5’-ACAAAGAAACTGTGTGTGTGA-3’;23 shDRP1: 5’-GCCAGCUAGAUAUUAACAACAAGAA-3’ . shControl: 5’- ACAGACUUCGGAGUACCUG-3’ . Viruses were collected from the medium 48 hr after transfection . For knockdown experiments , cells were infected with the collected viruses over 48 hr in the presence of polybrene , followed by GFP sorting for 3–4 days . pTsin- lentivirus vectors were used for overexpression of HA ( FLAG ) -SPOP-WT or mutants . The following antibodies were used: SPOP ( ab137537; Abcam ) , SPOP ( 16750-1-AP; proteintech ) , INF2 ( 20466-1-AP; proteintech ) , AR ( SC-816; Santa Cruz ) , DEK ( 16448-1-AP , Proteintech ) , IQGAP1 ( ab133490; Abcam ) , DRP1 ( 8570S; CST ) , KPNA5 ( A7731; Abcam ) , COX4 ( Abcam; ab14744 ) , Ubiquitin ( 6652–1; epitomics ) , Myc ( 9E10; Sigma ) , FLAG ( M2; Sigma ) , HA ( MM5-101R; Convance ) , Actin ( AC-74; Sigma ) . Mdivi-1 was purchased from Selleckchem . MitoSOX Red dye was purchased from Invitrogen . Ubiquitinated INF2 was prepared by transfecting FALG-INF2 , HA-Ub and Myc-SPOP in 293T cells ( 5x100 mm dish ) . After 48 hr , the cells were lysed in RIPA buffer and the transfected INF2 was immunopurified from cell lysates with anti-Flag M2 agarose beads ( Sigma ) before being resolved by 7 . 5% SDS-PAGE . After Coomassie blue staining , the band corresponding to ubiquitinated INF2 was excised . The liquid chromatography tandem mass spectrometry analysis was carried out at the Proteomics Center of our institute . For cell cycle analysis , cells were washed 48 h post-treatment with PBS and fixed in 70% ethanol overnight . The cells were washed again with PBS , stained with propidium iodide and analyzed by flow cytometry . Cell proliferation rate was determined using Cell Counting Kit-8 ( CCK-8 ) according to the manufacturer’s protocol ( Dojindo Laboratories , Japan ) . Briefly , the cells were seeded onto 96-well plates at a density of 1 , 000 cells per well . During a 2 to 8-d culture periods , 10 μl of the CCK-8 solution was added to cell culture , and incubated for 2 hr . The resulting color was assayed at 450 nm using a microplate absorbance reader ( Bio-Rad ) . Each assay was carried out in triplicate . Cell migration and invasion were determined by Transwell ( Costar ) migration and invasion assays . LNCaP cells were precultured in serum-free medium for 48 hr . For migration assay , 3x104 cells were seeded in serum-free medium in the upper chamber , and the lower chamber was filled with RPMI1640 containing 5% FBS . After 48 h , the non-migrating cells on the upper chambers were carefully removed with a cotton swab , and migrated cells underside of the filter stained and counted in nine different fields . Matrigel invasion assays were performed using Transwell inserts ( Costar ) coated with Matrigel ( BD Biosciences ) /fibronectin ( ( BD Biosciences ) . OCR was measured using a Seahorse XF24 Extracellular Flux Analyzer with the XF Cell Mito Stress Test Kit . Cells were seeded at 8 x 104 cells per well in 100μl DMEM containing 10% FBS and allowed to attach for 2 hr . 150μl DMEM-10% FBS was added per well and cells incubated overnight in 5% CO2 humidified incubator . Prior to assay run , cells were changed into assay media , unbuffered DMEM pH 7 . 4 and subjected to sequential injections of Oligomycin ( 1 μM ) , FCCP ( 0 . 3 μM ) , rotenone ( 1 μM ) and antimycin A ( 0 . 75 μM ) . Spare respiratory capacity was calculated by dividing the OCR response to FCCP by the basal respiration , having subtracted the non-mitochondrial respiration previously . All values were normalized to cell number per wells setup in parallel . Cells were seeded for 24hr and treated as indicated . TMRE ( 50 nM ) or MitoSOX Red ( 5 μM ) was added to the media , and the plates were incubated at 37°C in the dark for 30 min . Then cells were trypsinized and analyzed by flow cytometry . For immunofluorescence , cells were plated on chamber slides , fixed with 4% paraformaldehyde at room temperature for 30 min . After washing with PBS , cells were permeabilized with 0 . 1% Triton X-100 in PBS for 15 min . Cells were then washed with PBS , blocked with 0 . 5% BSA in PBS for 1hr , and incubated with primary antibodies in PBS for at 4°C for overnight . After washing with PBS , fluorescence-labelled secondary antibodies were applied and DAPI was counterstained for 1hr at room temperature . Cells were visualized and imaged using a confocal microscope ( LSM710 , Zeiss ) . The analytic method of mitochondrial length was described previously . 23 | Prostate cancer is the leading cause of global cancer-related death . The development of improved diagnoses and novel therapies has been confounded by significant patient heterogeneity . During recent years , significant progress has been made in identifying the molecular alterations in prostate cancer using next-generation sequencing . SPOP gene was frequently altered by somatic point mutations in a distinct molecular subclass of prostate cancer , although the precise role that SPOP mutation plays in the development of prostate cancer is unclear . Mitochondria are highly motile organelles that undergo constant fission and fusion . Unbalanced mitochondrial fission and fusion events are associated with mitochondrial dysfunction and frequently linked to human cancer . Here , we are the first to report that SPOP mutations are associated with dysregulation of mitochondrial dynamics in prostate cancer and this finding may have potential clinical implications in prostate cancer treatment . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"urology",
"cell",
"motility",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"cancer",
"cell",
"migration",
"293t",
"cells",
"pathogens",
"biological",
"cultures",
"cancers",
"and",
"neoplasms",
"genitourinary",
"tract",
"tumors",
"microbiology",
"retroviruses",
"oncology",
"viruses",
"developmental",
"biology",
"prostate",
"cancer",
"rna",
"viruses",
"mitochondria",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"ubiquitination",
"exocrine",
"glands",
"prostate",
"diseases",
"cell",
"lines",
"biochemistry",
"cell",
"biology",
"post-translational",
"modification",
"anatomy",
"viral",
"pathogens",
"cell",
"migration",
"biology",
"and",
"life",
"sciences",
"du145",
"cells",
"energy-producing",
"organelles",
"lentivirus",
"prostate",
"gland",
"organisms"
]
| 2017 | Dysregulation of INF2-mediated mitochondrial fission in SPOP-mutated prostate cancer |
Chronic infections are an increasing problem due to the aging population and the increase in antibiotic resistant organisms . Therefore , understanding the host-pathogen interactions that result in chronic infection is of great importance . Here , we investigate the molecular basis of chronic bacterial cystitis . We establish that introduction of uropathogenic E . coli ( UPEC ) into the bladders of C3H mice results in two distinct disease outcomes: resolution of acute infection or development of chronic cystitis lasting months . The incidence of chronic cystitis is both host strain and infectious dose-dependent . Further , development of chronic cystitis is preceded by biomarkers of local and systemic acute inflammation at 24 hours post-infection , including severe pyuria and bladder inflammation with mucosal injury , and a distinct serum cytokine signature consisting of elevated IL-5 , IL-6 , G-CSF , and the IL-8 analog KC . Mice deficient in TLR4 signaling or lymphocytes lack these innate responses and are resistant , to varying degrees , to developing chronic cystitis . Treatment of C3H mice with the glucocorticoid anti-inflammatory drug dexamethasone prior to UPEC infection also suppresses the development of chronic cystitis . Finally , individuals with a history of chronic cystitis , lasting at least 14 days , are significantly more susceptible to redeveloping severe , chronic cystitis upon bacterial challenge . Thus , we have discovered that the development of chronic cystitis in C3H mice by UPEC is facilitated by severe acute inflammatory responses early in infection , which subsequently are predisposing to recurrent cystitis , an insidious problem in women . Overall , these results have significant implications for our understanding of how early host-pathogen interactions at the mucosal surface determines the fate of disease .
Persistent microbial infections are a rapidly expanding problem because of increased antimicrobial resistance [1] . This trend is particularly concerning because of the increasingly appreciated role that chronic infections may play in cancer and chronic inflammatory diseases [2] , [3] , [4] . One key example is that of urinary tract infections ( UTI ) , which are common , highly recurrent , and can become chronic [5] , [6] . Females are disproportionately afflicted by UTI: 50% of all women will have an episode at some point in their lifetime , and 20 to 30% will have a recurrence within 3 to 4 months of the acute infection [7] . The high incidence of recurrent UTI ( rUTI ) suggests that many individuals do not develop protective immunity to uropathogens , though the capacity to do so has been demonstrated in both murine and primate experimental model systems and a phase 2 clinical vaccine trial in women [8] , [9] , [10] . This failure of adaptive immunity may be partially explained by host genetic and environmental factors , such as a maternal history of UTI and childhood exposure to uropathogens , which appear to play significant roles in determining susceptibility to rUTI [11] , [12] . However , the molecular basis for rUTI and how this may relate to mechanisms of chronic infection , where adaptive immunity also falls short , are unknown . Furthermore , while antibiotic therapy has been resoundingly successful in treating acute UTI , recent increases in the prevalence of antibiotic-resistant uropathogenic strains in the community threaten to make chronic UTI common again [5] , [13] , [14] . Thus , understanding the host mechanisms contributing to chronic UTI , and other chronic bacterial infections of the mucosae , is of critical importance . Uropathogenic Escherichia coli ( UPEC ) are by far the most common cause of UTI , accounting for 80% of outpatient infections and 25% of nosocomial infections [15] . During an acute episode , UPEC adhere to and invade the superficial facet cells of the urinary mucosal epithelium ( urothelium ) in a type 1 pili-dependent manner [16] , [17] . UPEC invasion has been reported to involve several components of lipid rafts such as caveolin-1 , an integral membrane protein found in the inner leaflet of the lipid bilayer [18]; Rac1 , a member of the Rho family of GTPases [19] , [20]; and microtubules [21] . After invasion , urothelial cells can expel UPEC via a TLR4-dependent exocytic pathway [22] . Alternatively , if UPEC escape into the cytoplasm , they can rapidly replicate , and subsequently aggregate into intracellular bacterial communities ( IBC ) [23] . Aggregation of UPEC into the biofilm-like IBC depends upon type 1 pili expression , independent of urothelial invasion , and is part of a mechanism for bacteria to evade extracellular host defenses while rapidly expanding in numbers during acute infection [24] . IBCs are transient in nature . Upon IBC maturation during approximately the first 12 to 16 hours of infection , the bacteria detach from the biomass and flux back into the lumen , spreading to neighboring epithelial cells where they are capable of initiating another IBC [25] . IBC formation has only been observed in the early acute stages of infection [25] . However , the ability of UPEC to expand in numbers via IBC formation has been shown to be a prerequisite for persistence as mutants that are defective in IBC formation are highly attenuated and rapidly cleared from the urinary tract [26] . IBC formation has been observed in multiple murine backgrounds with numerous UPEC strains [27] . Evidence of IBC formation has also been found in the bladders of mice infected with Klebsiella pneumoniae and in urine sediments from women with acute cystitis by UPEC , indicating that this intracellular pathogenic cycle is not unique to UPEC infection of mice [28] , [29] . UPEC colonization and invasion of the urothelium triggers innate host responses , which are mediated in part by Toll-like receptor 4 ( TLR4 ) , a pattern recognition receptor that responds to certain pathogen-associated molecular patterns such as lipopolysaccharide [30] , [31] , [32] . These early innate responses include bacterial expulsion , urothelial exfoliation , and bladder inflammation that is characterized by the production of the pro-inflammatory cytokine interleukin 6 ( IL-6 ) , granulocyte chemotactic cytokines such as IL-8 , the hormone granulocyte colony stimulating factor ( G-CSF ) , and the T cell-associated , pro-inflammatory cytokine IL-17A [17] , [33] , [34] , [35] , [36] , [37] , [38] . In humans , the ultimate outcome of UPEC infection of the urinary tract ranges from acute , self-limiting infection to asymptomatic bacteriuria ( presence of bacteria in the urine ) , to recurrent or chronic UTI [7] . Diminished TLR4 receptor expression and CXCR1 ( IL-8 receptor ) signaling have been associated with increased susceptibility to asymptomatic bacteriuria and severe pyelonephritis , respectively , in children , highlighting the importance of the innate immune response in determining disease outcome [39] , [40] . However , the host mechanisms that contribute to acute , recurrent and chronic cystitis in adult women are poorly understood . While mast cells , γδ T cells , and neutrophils have each been implicated in facilitating clearance of acute UTI in C57BL/6J mice [34] , [41] , [42] , [43] , recent studies have suggested that the role of neutrophils may be more complex than previously appreciated , highlighting the need for further studies [33] , [44] . Adding to this complexity is the fact that the outcome of experimental UPEC infection differs substantially between inbred mouse strains [45] . In C57BL/6J mice , which typically resolve acute infection and bacteriuria within 7–10 days , a small intracellular population of bacteria resistant to antibiotic therapy persists latently for months in the bladder urothelium of these mice [46] , [47] , [48] , [49] . This quiescent intracellular reservoir ( QIR ) is distinct from the IBC as it is comprised of fewer than 15 bacteria persisting in a membrane bound dormant state [49] . QIRs are capable of re-emerging from dormancy to cause recurrences of infection and bacteriuria , and may represent one mechanism for rUTI in humans [49] . In contrast , some C3H background mouse strains have been reported to develop chronic UTI for up to 2 weeks post-infection [45] , [50] . These latter strains may reflect in part the natural course of UTI in women , as placebo-controlled studies have demonstrated that a majority of women remain bacteriuric weeks after an acute UTI if not treated with antibiotics , despite overall improvement of symptoms [51] , [52] . Thus , these murine models of chronic UTI , which are not well understood , merit further study as they could reveal critical host determinants of UTI pathogenesis . In this study , we developed an experimental chronic infection model to study aspects of the host response that determine disease outcome . Using a C3H/HeN murine model of cystitis , we demonstrated that while acute , self-limiting infection occurs in a subset of mice , another potential outcome of UPEC infection is the establishment of chronic cystitis lasting months . We discovered acute biomarkers that are predictive for the development of chronic cystitis and showed that the development of chronic cystitis was dependent upon severe acute inflammation . Furthermore , results from challenge infections following antibiotic therapy indicated that a history of chronic cystitis was a significant risk factor for subsequent severe and chronic infections . These findings suggest that a common mechanism may underlie both chronic and recurrent bacterial infections .
C57BL/6J mice typically resolve acute UPEC UTI by 2 weeks post-infection ( wpi ) , with sterile urines and kidneys and low level bladder colonization ( <104 colony-forming units ( cfu ) UPEC per bladder ) indicative of a latent QIR population [46] . In contrast , we previously demonstrated that a large percentage of C3H/HeN mice have high level bladder colonization ( >104 cfu UPEC per bladder ) at 2 wpi [29] . However , C3H/HeN mice had not been previously reported to be susceptible to chronic UTI [32] , [45] . To determine whether this finding represented an alternative disease outcome or merely a high rate of recurrence , we performed comparative long-term UPEC infection studies investigating the progression of UTI in C57BL/6J and C3H/HeN mice . Female mice were infected at 7–8 weeks of age with 107 cfu of a kanamycin-resistant derivative of the UPEC strain UTI89 , UTI89 KanR , unless otherwise indicated , and then followed by longitudinal urinalysis over 4 wpi . We found that C3H/HeN mice were significantly more susceptible to developing persistent bacteriuria , as defined by the recovery of greater than 104 cfu UTI89 per ml of urine at all time points over 4 wpi , compared to C57BL/6J mice ( Table 1; Fig . 1A–B ) . Urine titers of 104 cfu/ml or more had previously been demonstrated to be a reasonable cutoff for indicating the presence of urinary tract infection in mice when assaying free catch urine samples [46] . Increasing the inoculum to 108 cfu UTI89 significantly increased the incidence of persistent bacteriuria in C3H/HeN mice from 21% to 52% ( P<0 . 01 , Fisher's Exact test ) , while C57BL/6J mice remained resistant to persistent bacteriuria ( Table 1 ) . Resolution of bacteriuria was not observed in C3H/HeN mice after 4 wpi in 12 persistently bacteriuric mice from two independent experiments that were followed for 6–8 months post-infection ( data not shown ) . Therefore , in response to UPEC infection a subset of C3H/HeN mice develop persistent bacteriuria lasting months in an infectious dose-dependent manner . Of the C57BL/6J and C3H/HeN mice that resolved bacteriuria , i . e . their urine titers fell to less than 104 cfu UTI89 per ml at least once during the course of infection , 80 of 81 had bladder bacterial titers <104 cfu at 4 wpi ( Table 1; Fig . 1C–D ) . The presence of low level bladder bacterial burdens after acute infection , with urine and kidney titers at or near the limit of detection , is consistent with a quiescent intracellular reservoir ( QIR ) population comprised of bacteria in a dormant state [48] , [49] . In contrast , C3H/HeN mice with persistent bacteriuria had significantly higher bladder and kidney titers at 4 wpi ( Table 1; Fig . 1D ) . Furthermore , in these mice whole bladder titers exceeded kidney titers on average ( geometric mean ) by 2 . 9 orders of magnitude at the 107 cfu inoculum , and increasing the inoculum did not significantly alter this bias towards bladder infection ( p = 0 . 38 , Mann-Whitney test ) . As a result , the development of persistent bacteriuria had a high positive predictive value ( PPV ) for the presence of high titer bladder infection ( >104 cfu ) at 4 wpi ( Table 1 ) . In contrast , persistent bacteriuria in C57BL/6J mice was rare and when it did occur , severe kidney infection with abscess formation was a consistent finding in those mice ( Table 1 , data not shown ) . Gross examination of the urinary tract organs of C3H/HeN mice at 4 wpi infection with either 107 or 108 cfu UTI89 ( n = 58 ) revealed that all bladders from mice with persistent bacteriuria ( n = 18 ) were enlarged and rigid compared to the bladders from mice that resolved bacteriuria ( n = 40 ) . Gross examination of the kidneys revealed no differences between disease outcomes , and kidney abscesses were not observed . Histopathological analysis of bladder tissue was performed on a random subset of mice ( n = 17 ) . Bladders from mice that had resolved bacteriuria ( n = 9 ) lacked any evidence of cystitis ( inflammation of the bladder ) or noticeable pathology and the bladder epithelium appeared to be fully repaired ( Fig . 1E ) . In contrast , abundant bacterial colonization accompanied by lesions of both acute and chronic inflammation , as defined by the presence of polymorphonuclear leukocytes ( PMN , also called granulocytes ) and mononuclear leukocytes , respectively , was detected in 7 of 8 bladders examined from mice with persistent bacteriuria ( Fig . 1F ) . Acute inflammation was most apparent within the urothelium , with epithelial reactivity and marked infiltration by PMN that resemble neutrophils and other CD45+ cells . The bladder urothelium of persistently bacteriuric mice was hyperplastic and poorly differentiated , lacking the superficial facet cell layer as indicated by diminished uroplakin III expression on the luminal surface ( Fig . 1G–H ) . Unlike the acute phase when UPEC progress through IBCs , bacterial colonization in these bladders appeared to be entirely luminal in nature as IBCs or other intracellular bacteria were not observed within the urothelium by immunofluorescence staining . Chronic inflammation was observed in the lamina propria ( connective tissue layer underlying urothelium ) with accumulations of lymphoid-like CD45+ cells in large follicle-like aggregates ( Fig . 1F , I–J ) . These lymphoid follicles were found in all mice with persistent bacteriuria ( n = 8 mice , 2–5 follicles/tissue section ) , while none were observed in the bladders of mice that resolved bacteriuria . Similar pathology of bacterial colonization coupled with both acute and chronic inflammation were observed in the upper urinary tract of mice with persistent bacteriuria ( data not shown , n = 4 ) . However , these lesions were confined to tissues lined by urothelium: the ureters and renal pelvices , and were minimal or absent in the renal medullae and cortices . Thus , persistent bacteriuria in C3H/HeN mice through 4 wpi is highly indicative of chronic infection and inflammation of all urothelium-lined tissues and will be used in the following analyses as a method for identifying C3H/HeN mice with chronic cystitis . The outcome of UPEC infection in C3H/HeOuJ and C3H/HeJ mice was also investigated , as these two C3H substrains have been reported to be highly susceptible to chronic bladder and kidney infection at 2 wpi with UPEC [45] , [46] . We found that both C3H/HeOuJ and C3H/HeJ mice develop persistent bacteriuria at a much more efficient rate than C3H/HeN mice upon infection with 107 cfu UTI89 KanR ( Table 1 , Fig . 2A ) . However , unexpectedly we found that C3H/HeJ mice , which are incompetent for TLR4 signaling , had significantly lower levels of bladder infection compared to C3H/HeOuJ mice at 4 wpi ( Fig . 2B; Table 1 ) . Yet , these strains were equally and highly susceptible to pyelonephritis compared to C3H/HeN mice ( Fig . 2C , Table 1 ) , often with grossly visible kidney abscess formation . Thus , persistent bacteriuria is a poor predictor of high bladder bacterial burdens ( >104 cfu ) at 4 wpi in C3H/HeJ mice compared to either C3H/HeN or C3H/HeOuJ mice at 107 cfu inoculum ( PPV of 44% versus 100% , P<0 . 05 , Fisher's Exact Test; Table 1 ) . Finally , 5/25 C3H/HeJ mice died between 1 and 7 days post-infection ( dpi ) , demonstrating the contribution of TLR4 signaling in preventing fatal bacterial infection . Among C3H/HeOuJ and C3H/HeJ mice with persistent bacteriuria and high bladder bacterial burdens ( >104 cfu ) at 4 wpi , a number of differences were observed . The bladders from C3H/HeJ mice meeting these criteria ( n = 7 ) weighed significantly less than those from C3H/HeOuJ mice ( n = 22 ) ( Fig . 2D ) . The bladders from C3H/HeOuJ mice with persistent bacteriuria demonstrated lesions of acute and chronic bladder inflammation ( Fig . 3 , panels A , and D ) identical to those described above for C3H/HeN mice with persistent bacteriuria ( see Fig . 1 ) . These included marked luminal bacterial colonization and urothelial hyperplasia accompanied by extensive acute and chronic inflammatory cell infiltrates , highlighted by marked PMN infiltration and numerous lymphoid follicles . In contrast , histological evidence of bladder inflammation was only present in the 5 C3H/HeJ mice with the highest bladder bacterial burdens ( >106 cfu ) . Moreover , this inflammation , when present , was greatly subdued ( Fig . 3 , panels B , and E ) compared to that observed in C3H/HeOuJ and C3H/HeN mice with high bladder bacterial burdens . Specifically , neutrophil infiltration from the mucosal tissue was reduced and lymphoid follicles were absent . For comparison , C3H/HeOuJ mice that resolved bacteriuria and C3H/HeJ mice with bladder colonization below 106 cfu all lacked detectable bladder inflammation ( Fig . 3 , panels C and F , data not shown ) . The urothelial response to high titer chronic UPEC infection of the bladder also differed between C3H/HeOuJ and C3H/HeJ mice . As with C3H/HeN mice , the terminal differentiated facet cell layer was entirely ablated in the C3H/HeOuJ mice with persistent bacteriuria and high titer bladder infections as indicated by the lack of uroplakin staining ( Fig . 3G ) . In contrast , in the C3H/HeJ mice with persistent bacteriuria , high bladder titers , and evidence of bladder inflammation , the bladder mucosae retained a capacity for terminal differentiation , albeit reduced , as evidenced by the intermediate level of uroplakin expression compared to C3H/HeJ mice that lacked bladder inflammation ( Fig . 3H and I ) . Furthermore , fully mature superficial facet cells with abundant uroplakin expression were detected in every bladder section , and a number of these harbored dense bacterial colonies surrounded by uroplakin staining , i . e . apparent IBCs ( Fig . 3H , arrow ) . Neither mature superficial facet cells , nor IBCs were ever observed in the chronic stages of UTI in C3H/HeN or C3H/HeOuJ mice although IBCs were observed normally in the acute stages . We also observed that luminal colonization of the urothelium appeared much more robust in these 5 C3H/HeJ mice than in the other two C3H substrains , resembling biofilms ( Fig . 3 , panels E & H ) . Therefore , in contrast to previous studies [45] , [50] , our data indicate that C3H/HeJ mice have significantly lower bladder bacterial burdens during chronic infection than C3H/HeOuJ mice , despite similar bacterial burdens in the kidneys . Furthermore , when bacterial persistence does occur in the bladders of C3H/HeJ mice , it may be mechanistically distinct from chronic cystitis in TLR4 signaling-competent C3H mice . Given that C3H/HeJ mice are known to be immunodeficient , mounting only very muted acute inflammatory responses to UPEC infection [30] , [53] , these results suggest that the host immune response may contribute to the development of chronic cystitis . We then investigated whether chronic cystitis was specific to UPEC infection in C3H/HeN and C3H/HeOuJ mice . We found that other C3H sub-strains , C3H/HeSnJ and C3HeB/FeJ , were also capable of developing chronic cystitis in response to infection with UTI89 ( Table 1 , Fig . S1A–B ) . However , the relative susceptibilities of these sub-strains to chronic cystitis differed: C3H/HeN mice were among the most resistant ( 18% at 107 cfu ) and C3H/HeOuJ mice were the most susceptible ( 92% at 107 cfu ) . CBA/J mice , which are commonly used in experimental UTI models [54] and like C3H mice are descendents of a DBA by Bagg albino cross , and DBA/2J mice were each also susceptible to chronic cystitis ( Table 1 , Fig . S1C–D ) . In contrast , other common strains such as BALB/cJ and 129S1/SvImJ were not observed to develop chronic cystitis at either 107 or 108 cfu inoculum of UTI89 in this model ( Table 1 , Fig . S2 ) , despite successful acute infections . Conversely , infection of C3H/HeN mice with either J96 , another UPEC isolate , or Top52 , a Klebsiella pneumoniae human cystitis isolate that had previously been demonstrated to form IBCs during acute infection [29] , resulted in varying incidences of chronic cystitis ( Fig . S3 ) . Therefore , chronic cystitis in response to UPEC or Klebsiella pneumoniae infection appears to be restricted to certain strain backgrounds . The relative resistance of TLR4 signaling-deficient C3H/HeJ mice , which have muted acute inflammatory responses [45] , [50] , [55] , to chronic cystitis when compared to C3H/HeOuJ mice , which have very robust acute inflammatory responses to UPEC infection [45] , [50] , suggested a positive association between acute inflammation and chronic cystitis . To investigate this further , we tested whether the severity of the acute bladder inflammatory responses correlated with disease outcome in C3H/HeN mice . We found that mice that proceeded to develop persistent bacteriuria had significantly higher urine bacterial titers at 24 hours post-infection ( hpi ) than their cage mates that resolved bacteriuria ( Fig . 4A ) . These mice also had increased numbers of PMN in the urine ( pyuria ) , and greater weight loss at 24 hours post-infection ( hpi ) than their cage mates that resolved bacteriuria ( Fig . 4B–C ) . However , this weight loss was transient and the mice quickly recovered to gain weight due to aging , indicating that chronic infection is well tolerated through 4 wpi . Severe acute pyuria and weight loss at 24 hpi was also predictive of chronic cystitis in C3H/HeOuJ mice , but was not observed in C3H/HeJ mice ( Fig . S4A–B ) . To identify the nature of this acute systemic inflammatory response , we monitored the serum of mice at 24 hpi by multiplex cytometric bead array for the presence of 23 mouse cytokines . We found that C3H/HeN mice that developed persistent bacteriuria had significantly elevated levels of four serum cytokines at 24 hpi compared to both mock-infected mice and mice that resolved active infection: IL-5 , IL-6 , granulocyte colony-stimulating factor ( G-CSF , encoded by Csf3 ) , and keratinocyte-derived cytokine ( KC , a . k . a . growth-regulated alpha protein , encoded by Cxcl1 ) , an IL-8 analog in mice ( Fig . 4D ) . Elevated levels of these serum cytokines at 24 hpi were also prognostic of chronic cystitis in C3H/HeOuJ , but not C3H/HeJ , mice ( Fig . S4C ) . Another UPEC isolate , J96 , also elicited similar serum responses prognostic of chronic cystitis in C3H/HeN mice ( Fig . S5 ) . Finally , we found that elevated urine IL-6 , G-CSF , and KC , but not IL-5 , were also predictive of the development of chronic cystitis in C3H/HeN mice at 24 hpi ( Fig . S6 ) At 24 hpi , the bladder inflammatory responses to acute UPEC infection vary widely among individual C3H/HeN mice infected with 107 cfu UTI89 ( Fig . 5A–D ) , despite the fact that the bacterial titers recovered from these bladders are fairly uniform at 6 hpi [24] , [29] . Thus , we investigated the potential correlation between the severity of bladder inflammation and serum cytokine levels as prognosticators of chronic cystitis . C3H/HeN mice were infected with 107 cfu UTI89 . At 24 hpi we found that 20% of the mice had severely inflamed urinary bladders with full-thickness urothelial necrosis ( inflammatory scores of 5 ) , marked edema , epithelial necrosis , and inflammatory cell infiltrates ( Fig . 5A , 5B ( right ) and 5D ) , while the other 80% of the bladders had minimal to moderate inflammation at 24 hpi ( Fig . 5A , inflammatory scores of 1–4; 5B ( left ) and 5C ) . The incidence of severe inflammation was infectious dose-dependent , as 80% of mice had bladder inflammatory scores of 5 when infected with 108 cfu ( Fig . 5A ) . The levels of all four serum cytokines prognostic of chronic cystitis strongly and positively correlated with the degree of bladder inflammation at 24 hpi with 107 cfu ( Fig . 5E ) . Collectively , these findings demonstrate a strong association between severe , acute immunopathology and the development of chronic cystitis . To further substantiate whether the immune response is required for the development of chronic cystitis in C3H mice , we investigated the outcome of UPEC infection in C3H background severe combined immunodeficient ( C3Hscid ) mice , which lack mature lymphocytes and may have other immune cell abnormalities . We found that C3Hscid mice are almost entirely resistant to the development of chronic cystitis compared to the congenic wild type strain , C3H/HeSnJ ( Figs . 6A , S7A ) . Even after infection with 108 cfu of UTI89 , only 4% of C3Hscid mice developed persistent bacteriuria and high titer ( >104 cfu ) bladder infection at 4 wpi , compared to 62% of C3H/HeSnJ mice ( 1/23 versus 16/26 , P<0 . 0001 , Fisher's Exact test ) . However , intermittent bacteriuria was common in the C3Hscid mice , perhaps indicating a failure to eliminate colonization of specific niches . Nevertheless , intermittent bacteriuria was not associated with bladder bacterial burdens >104 cfu at 4 wpi . Acute inflammation was greatly muted in the C3Hscid mice compared to their congenic wild type controls . PMN scores and weight loss at 24 hpi were significantly reduced ( P<0 . 01 for each , Mann-Whitney test; Figs . 6B , S7B ) . Furthermore , while the C3Hscid mice had a serum IL-5 response similar to wild type controls , they were completely unable to mount serum IL-6 , G-CSF , and KC responses ( Fig . 6C ) . Examination of the bladder tissue from 5 C3Hscid and 5 wild type control mice at 24 hpi ( Fig . 6D–E ) revealed that these strains had similar levels of bladder edema ( Fig . 6E ) . However , C3Hscid mice entirely lacked the severe cellular inflammatory infiltrates and urothelial exfoliation seen in wild type C3H mice at 24 hpi , despite vigorous acute infection , including numerous apparent IBCs ( Fig . 6E , arrows ) . Thus , the muted inflammatory response does not appear to result from a defect in acute colonization . These data provide further evidence that inflammation is required for the development of chronic cystitis and suggest that lymphocytes may play a necessary role in this process . To test whether severe acute inflammation is specifically required for the development of chronic cystitis , we treated C3H/HeN mice with a single immunosuppressive dose of the glucocorticoid , dexamethasone sodium phosphate , by intraperitoneal injection 2 hours prior to infection . We found that mice treated with dexamethasone were significantly more resistant to the development of chronic cystitis ( Figs . 7A , S8A ) upon infection with 108 cfu UTI89 KanR compared to saline-treated controls ( 2/17 versus 13/18 , p<0 . 001 , Fisher's Exact test ) . Furthermore , dexamethasone treatment significantly reduced the severity of acute inflammation at 24 hpi in reponse to UPEC infection as indicated by pyuria ( Fig . 7B ) , weight loss ( Fig . 7C ) , serum cytokines ( Fig . 7D ) , and bladder inflammation ( Fig . 7E–G ) . Yet , these groups had similar bladder bacterial burdens at 24 hpi ( Fig . S8B ) , including what appears to be robust IBC formation in the dexamethasone treated group ( Fig . 7G , arrows ) . Thus , the development of chronic cystitis appears to require severe acute inflammation , strongly suggesting that bladder immunopathology during acute infection facilitates the establishment of chronic infection . Lymphoid follicles have been reported in humans with persistent bacteriuria , with or without symptoms of UTI , and have been reported to resolve spontaneously after antibiotic therapy to clear infection [56] , [57] . We also found that persistent bacteriuria in C3H/HeN mice was rapidly eliminated by antibiotic therapy initiated at 4 wpi ( n = 46 ) . Bladder tissue sections from a sample of these mice ( n = 5 ) , together with mock-infected controls ( n = 6 ) and mice that had resolved bacteriuria ( n = 9 ) , were examined one month after initiation of antibiotic therapy . This analyisis revealed that urothelial integrity was indistinguishable between these groups , indicating full healing of the urothelium in the previously infected mice ( Fig . 8A–C ) . However , the bladders of mice with a history of persistent bacteriuria weighed significantly more than either of the other two groups ( Fig . 8D ) , and while the lymphoid follicles were largely gone , small to moderately sized clusters of mononuclear cells remained in the bladder lamina propria ( Fig . 8C ) . Similar chronic inflammatory cell infiltrates have been reported after serial UPEC infections of C57BL/6J mice , where their presence coincides with the development of adaptive immunity [9] . These findings suggested that an acquired , bladder-resident immune cell population may modulate the host response to subsequent infection . Therefore , we tested whether the specific disease outcome , persistent bacteriuria or resolved bacteriuria , affected the susceptibility of mice to UTI challenge after clearance of infection with antibiotics , compared to age-matched naïve mice ( Fig . 8D ) . All challenges were performed 4 weeks after initiation of antibiotic therapy , which allowed time for the bladder urothelium to heal fully as described above ( Fig . 8A–C ) . We found that mice with a history of persistent bacteriuria for 4 wpi with UTI89 KanR at 7–8 weeks of age were significantly more prone to developing persistent bacteriuria ( Fig . 8E ) and chronic cystitis ( Table 2 ) upon challenge infection with an isogenic , spectinomycin-resistant UTI89 strain , UTI89 SpcR , at 15–16 weeks of age than either their cage mates who spontaneously resolved bacteriuria or age-matched naïve mice that had previously been mock-infected . Redevelopment of chronic cystitis was significantly associated with the onset of severe pyuria and elevated serum IL-5 , G-CSF , and KC at 24 hpi , compared to mice that resolved the challenge infection , whether or not they had a history of chronic cystitis ( data not shown ) . The rapid clearance of bacteriuria after bacterial challenge in mice that had previously spontaneously resolved infection , compared to naïve mice , suggests that these mice develop some degree of acquired immunity to UPEC . Only 5% ( 2/42 ) of mice previously infected with UTI89 KanR had recurrent bacteriuria with this strain on the day of challenge , 4 weeks after initiation of antibiotic therapy , similar to what was previously reported in C57BL/6J mice [46] . Otherwise , little reemergence of UTI89 KanR was observed upon challenge with UTI89 SpcR ( data not shown ) . As additional controls , a subset of 20 mice , 10 previously infected and 10 age-matched naïve , and all treated with the antibiotic regimen described above , were challenged with PBS and their urines remained free of UPEC for the remainder of the experiment ( data not shown ) . To determine the duration of infection sufficient for modulating the host susceptibility to further infection , we initiated antibiotic treatment at 24 hpi and 2 wpi and then challenged mice with UPEC four weeks later , as described above ( Table 2 , Fig . 8F–G ) . We found that 14 days of chronic infection were sufficient for inducing the hypersusceptibility phenotype to challenge infection , but 1 day of infection was not . This result indicates that , although accurate indicators of the development of chronic cystitis exist at 24 hpi , infection-related events in the course of sub-acute to chronic UTI are necessary for the enhanced susceptibility to further chronic infections . Furthermore , these findings refute the alternative explanation to our data: that we are selecting for a pre-existing subpopulation , either due to the presence of genetic polymorphisms within the breeding colony , or prior exposure to pathogens . Taken together , these data indicate that the development of chronic UPEC cystitis predisposes C3H mice to chronic UTI upon subsequent colonization .
Using a C3H murine model of UTI , we have unveiled an inflammation-dependent checkpoint occurring at the host-pathogen interface within the first 24 hours of infection that has significant ramifications upon the long term fate of disease . Furthermore , subsequent development of chronic cystitis alters the host susceptibility to further bacterial cystitis after antibiotic therapy , resulting in increased susceptibilities to both severe acute symptomatology and chronic infections . We identified elevated IL-5 , IL-6 , G-CSF , and KC as a serum biomarker signature prognostic of the development of chronic cystitis in C3H mice . This signature was accompanied by severe bladder inflammation at 24 hpi as indicated by marked bladder edema and pyuria . While the severity of bladder damage and systemic signs diminish after the early acute stage of infection , these mice remain persistently bacteriuric and develop chronic cystitis . This improvement in clinical signs despite persistent bacteriuria is similar to the natural course of UTI reported in a majority of women from placebo-controlled studies [51] , [52] . The alternate host response was characterized by mild to moderate bladder inflammation in the acute stage with minimal or no systemic signs . These mice typically achieved resolution of acute bacteriuria within 1–2 weeks , though recurrent bacteriuria after this time was common , either because of transient periurethral colonization or rUTI . The fate of disease outcome appears to be determined within the first 24 hpi and is infectious dose-dependent , indicating that this response is at least partially stochastic . Understanding the molecular and cellular mechanisms of this acute checkpoint may have far-reaching implications for how UTI and other chronic and recurrent infections are treated and evaluated . Our findings indicate that a glucocorticoid-sensitive , pro-inflammatory response to early acute infection is required for the efficient development of chronic cystitis . Glucocorticoids are made in the adrenal gland and , among other functions , act as potent anti-inflammatory agents . Specifically , they block the production of acute mediators of inflammation such as TNFα , IL-1 , IL-6 , and arachidonic acid derivatives such as prostaglandins and leukotrienes , in part through blockade of NF-κB activity [58] . For example , dexamethasone is commonly used as an adjunct therapy to treat acute bacterial meningitis , as its anti-inflammatory activity appears to improve survival [59] . The immunosuppressive effects of a single dexamethasone treatment are transient , providing evidence for a mucosal immune checkpoint early in UPEC infection that , when triggered , leads to urinary tract immunopathology and facilitates chronic cystitis . The resistance of C3Hscid mice to chronic cystitis implicates lymphocytes as immune cells necessary for the development of chronic cystitis in C3H mice . Specifically , our data suggest that one or more lymphocyte populations with innate properties may play a necessary role in mediating a number of the early severe inflammatory responses to UPEC that are associated with the development of chronic cystitis . Candidate lymphocytes include γδ T cells , natural killer T ( NKT ) cells , and B1 cells , which make natural antibodies . γδ T cells have previously been implicated in host resistance to UTI in C57BL/6J mouse infection models [34] , [41] , and they are normally found in the bladder lamina propria of naïve mice [60] . Acute host responses that were muted or absent in C3Hscid mice included severe pyuria , weight loss , and elevation of serum IL-6 , G-CSF , and KC . Induction of both IL-6 , an acute phase pro-inflammatory cytokine that causes fever and malaise if systemic , and IL-8 , a chemotactic cytokine for granulocytes , are strongly associated with UTI in both humans and mice [36] , [37] , [61] , [62] , [63] . The combination of elevated IL-6 , KC , an IL-8 analog , and G-CSF , a growth factor that promotes granulocyte development and release from the bone marrow , in both the serum and urine would explain the weight loss and severe acute PMN response associated with the development of chronic cystitis in mice [64] . Yet , these acute molecular and cellular responses are inadequate for resolving infection in C3H mice . Furthermore , a recent study suggests that such a response may contribute to disease pathogenesis , as depletion of G-CSF surprisingly resulted in lower UPEC titers in the bladders of C57Bl/6J mice at 48 hpi , despite reduced neutrophil infiltration into the bladder [33] . Similarly , treatment with G-CSF exacerbates disease in a murine model of Klebsiella pneumoniae respiratory tract infection [65] . Further studies are needed to identify the immune cell populations required for the development of chronic cystitis , and to clarify the role of these early innate responses in disease pathogenesis . We have found a strong association between elevated serum IL-5 at 24 hpi and the development of chronic cystitis . IL-5 promotes eosinophil and B1 lymphocyte development and IgA class switching and has been reported to enhance production of IL-6 in kidney cell lines [66] , [67] . However , the role of IL-5 in UTI pathogenesis is unclear . While “low level” IL-5 mRNA expression has been reported in the mouse bladder early in UPEC infection of C3H/HeN mice , IL-5 has not otherwise been implicated in UTI pathogenesis [68] . T helper cells polarized to produce type 2 cytokines ( TH2 cells ) have long been considered the primary source of IL-5 , but the acute serum IL-5 response to UPEC infection in this study was still present in C3Hscid mice . This finding is consistent with a previous study in mice , which found that non-lymphoid cells , such as mast cells or basophils , are the major source of IL-5 in the peripheral organs [69] . Activation of either of these cell types could mediate the bladder edema seen in C3Hscid mice . However , if IL-5 is originating from the bladder , it does not appear to be able to access the bladder lumen as IL-5 was not elevated in the urines of mice at 24 hpi , regardless of disease outcome . C3H background mice are known to have differing susceptibilities to various bacterial infections [70] , [71] , [72] , [73] . Currently , the specific genetic basis for susceptibility to chronic cystitis and chronic pyelonephritis are unknown . However , a recent mouse genetics study by Hopkins and colleagues has made significant progress in identifying genetic loci associated with susceptibility to cystitis and pyelonephritis at 10 dpi in second generation crosses of C3H/HeJ ( reported as susceptible to each ) and Balb/c ( resistant ) mice [74] . In those studies , the authors utilized quantitative trait loci ( QTL ) analyses to discover that the genetic basis for chronic cystitis and pyelonephritis differed . This is consistent with our data comparing experimental UPEC infection in the various C3H sub-strains through 4 wpi , which demonstrate that chronic cystitis and chronic pyelonephritis can be independent disease outcomes . In C3H/HeN , C3H/HeSnJ , and C3HeB/FeJ mice with chronic cystitis , concurrent kidney infection was an inconsistent finding and , when present , was largely limited to the renal pelvis ( pyelitis ) . Equally , the majority of C3H/HeJ mice with chronic pyelonephritis did not have cystitis at sacrifice , as defined by the presence of bladder inflammation and bacterial titers greater than 104 cfu . Only in C3H/HeOuJ mice was there a strong association between chronic cystitis and pyelonephritis . Since previous work had demonstrated that UPEC infection of both C3H/HeJ and C3H/HeOuJ mice resulted in similarly high bladder titers at 14 dpi , it had been hypothesized that chronic cystitis occurred independently of TLR4 signaling [45] , [50] . However , our studies have now demonstrated that these two strains differ in their susceptibility to chronic cystitis . We suggest that differences in experimental methodology , including the use of ten-fold higher inocula and study durations limited to 2 weeks , likely prevented resolution of the differing susceptibilities of C3H/HeOuJ and C3H/HeJ mice to chronic cystitis in previous studies [45] , [50] . In support of the hypothesis that enhanced TLR4 signaling contributes to the development of chronic cystitis , which in turn enhances host susceptibility to recurrent symptomatic infection , a TLR4 polymorphism that results in diminished responses to LPS has recently been associated with reduced susceptibility to rUTI in humans [75] . We also found that a subset of C3H/HeJ mice developed a persistent biofilm-like bladder colonization in the absence of robust inflammation . Bacterial biofilms and reduced TLR4 expression on neutrophils have each been associated with asymptomatic bacteriuria in humans [40] , [76] . Thus , the innate host responses in the urinary tract must be fine-tuned to successfully eliminate UPEC infection while also maintaining tissue integrity . Intracellular bacterial communities ( IBC ) had not been previously observed during the chronic stage of UTI and thus their role in bacterial persistence was unknown . However , IBC formation is known to be required during the acute stages of infection [17] , [23] , [24] , [25] , [26] , [27] , [28] , [29] which is a prerequisite for subsequent chronic infection . For example , we recently identified residues in FimH , the mannose-binding tip adhesin subunit of type 1 pili , that are under positive selection in UPEC strains isolated from human patients with UTI and that function in IBC formation [26] . A double mutation in two of these residues abolished IBC formation despite retaining the ability to bind to mannose , bind to the urothelium , and invade the urothelium . This mutant was severely defective in a mouse model of UTI , and behaved similarly to a fimH knockout or a FimH receptor binding mutant . Therefore , this mutant separated the phenotypes of ( i ) mannose-inhibitable binding and invasion of the urothelium and ( ii ) IBC formation [26] . Its phenotype in a mouse model of UTI indicates that both of these phenotypes are critical for acute stages of UTI , which we have now shown are a prerequisite for persistence . However , the lack of development of chronic cystitis in C3Hscid mice and in C3H mice treated with dexamethasone was not due to the lack of IBC formation in the acute stages of infection since IBCs developed normally in the acute stages of infection in these mice . Furthermore , we found evidence of IBC formation in the chronic stage of infection in C3H/HeJ mice , which have muted inflammatory responses to UPEC infection . This finding implicates mucosal inflammation , including urothelial reactivity and exfoliation , in restricting continuous IBC formation in immunocompetent mice . It also raises the hypothesis that one mechanism of asymptomatic bacteriuria , for which UPEC infection of C3H/HeJ mice has been proposed as a model [77] , may be continuous IBC formation in hosts with muted urothelial inflammatory responses to UPEC . In summary , we have discovered a new basis for understanding UTI that provides a possible mechanism for both chronic and recurrent infection . We propose that , in females that are genetically predisposed to enhanced mucosal TLR4 signaling , initial episodes of UTI due to UPEC or Klebsiella pneumoniae may be particularly severe in both degree of symptomatology and duration . If allowed to progress past the early acute stage before initiation of antibiotic therapy , these individuals could then develop altered bladder mucosal responses to gram-negative uropathogens . Upon repeated exposure to gram-negative uropathogens , these individuals would then be at increased risk for developing severe , symptomatic rUTI . It is also possible that a quiescent intracellular reservoir ( QIR ) state may exist in humans , similar to what has been described in mice [47] , [49] , not only serving as seeds for recurrent UTI , but also potentially modulating the mucosal response to UPEC due to the chronic persistence of bacteria within mucosal cells . This hypothetical model not only provides a logical basis for further , intensive murine and human UTI studies , but also extends our general understanding of chronic and recurrent gram-negative infections of mucosae .
All animal experimentation was conducted following the National Institutes of Health guidelines for housing and care of laboratory animals and performed in accordance with institutional regulations after pertinent review and approval by the Animal Studies Committee at Washington University School of Medicine . The UPEC strains used in this study were the human cystitis isolate , UTI89 [17] and derivatives thereof: UTI89 attHK022::COM-GFP ( kanamycin-resistant , KanR ) and UTI89 attλ::PSSH10-1 ( spectinomycin-resistant , SpcR ) [78]; and the human pyelonephritis isolate , J96 [79] . The Klebsiella pneumoniae strain used , TOP52 , is a human cystitis isolate [29] . Bacteria were routinely cultured in Luria-Bertani ( LB ) broth . C3H/HeN mice were obtained from Harlan Sprague Dawley , Inc . ( Indianapolis , IN ) . 129S1/SvImJ , Balb/cJ , C3HeB/FeJ , C3H/HeJ , C3H/HeOuJ , C3H/HeSnJ , C3Smn . CB17-Prkdcscid/J , C57BL/6J , CBA/J and DBA/2J mice were all obtained from the Jackson Laboratory ( Bar Harbor , ME ) . Bacterial strains were inoculated into 20 mL of LB broth directly from freezer stock , grown statically at 37°C overnight , and subcultured 1∶1000 into 20 ml of fresh media and again grown statically at 37°C for 18 hr . These cultures were centrifuged for 10 min at 3000×g , resuspended in 10 ml PBS , and then diluted to approximately 2–4×108colony forming units ( cfu ) /ml ( OD600 = 0 . 35 ) . 50µL of this suspension ( ∼1–2×107 cfu ) or one concentrated 10-fold ( ∼1–2×108 cfu ) was inoculated into the bladders of 7–8 week old female mice by transurethral catheterization . In most cases urines were collected prior to infection , at 1 , 3 , 7 , 10 , and 14 dpi , and then weekly thereafter by applying suprapubic pressure with proper restraint and collecting the urine stream in sterile 1 . 5ml eppendorf tubes . Urines were then serially diluted in PBS and 50 µL total of each dilution was spotted onto LB , and LB with 25 µg/ml kanamycin ( LB/Kan25 ) where appropriate . In experiments where a marked strain was used , urine titers were always reported from the plate containing the relevant antibiotic . Urine sediments were obtained by cytocentrifuging 80 µL of a 1∶10 dilution of the collected urine onto poly-L-lysine-coated glass slides and stained as described [28] . Stained urine sediments were examined by light microscopy on an Olympus BX51 light microscope ( Olympus America ) , and the average number of polymorphonuclear leukocytes ( PMN ) per 400× magnification field ( hpf ) calculated from counting 5 fields . A semi-quantitative scoring sytem was created to facilitate analysis: 0 , less than 1 PMN/hpf; 1 , 1–5 PMN/hpf; 2 , 6–10 PMN/hpf; 3 , 11–20 PMN/hpf , and 4 , >20 PMN/hpf . To quantify the bacteria present in urinary tract tissues at the time of sacrifice , bladders and kidneys were aseptically harvested at the indicated time point and homogenized in PBS . Homogenates were then serially diluted and spotted as described above , duplicate plating on LB and LB/Kan25 where appropriate . Tissue homogenates to be analyzed later for the presence of soluble cytokines were centrifuged at high speed for 5′ at 4°C and the supernatant removed for storage at −80°C . Venous blood was collected by submandibular puncture using 5 mm steel lancets ( Medipoint , Inc . , Mineola , NY ) into BD Microtainer serum separation 400 µl tubes . Blood tubes were allowed to clot at room temperature for 1–2 hours and , after centrifugation at 15 , 000×g for 5′ , were stored at −20°C . Tissues were either fixed in methacarn ( 60% methanol , 30% chloroform , 10% glacial acetic acid ) or embedded in OCT and frozen on dry ice before long term storage at −80°C . Methacarn fixed tissues were embedded in paraffin , sectioned , and stained with hemotoxylin and eosin . Bladder inflammatory scores were determined in a blinded fashion by two investigators ( T . J . H . and C . S . H . ) and an average score calculated , as previously described [45] . 7µm thick frozen sections were cut and fixed in acetone at −20°C for 10 minutes . All sections were hydrated and blocked in 1% BSA , 0 . 3% triton X-100 in PBS . After incubation with primary and secondary antibodies and associated washes , slides were stained with bis-benzimide ( Sigma ) . Stained tissues were examined by epifluorescence microscopy on a ZEISS Axioskop 2 MOT Plus microscope . The presence of 23 mouse cytokines was analyzed in specimens by a Luminex-based multiplex cytometric bead array platform ( Bioplex , Bio-Rad , Hercules , CA ) . These cytokines were IL-1α , IL-1β , IL-2 , IL-3 , IL-4 , IL-5 , IL-6 , IL-9 , IL-10 , IL-12 ( p40 subunit ) , IL-12 ( p70 subunit ) , IL-13 , IL-17 , Eotaxin , G-CSF , granulocyte-macrophage colony-stimulating factor ( GM-CSF ) , gamma interferon , KC , monocyte chemotactic protein 1 ( MCP-1 ) , macrophage inflammatory protein ( MIP ) 1α , MIP-1β , Regulated upon Activation , Normal T-cell Expressed and Secreted ( RANTES ) , and tumor necrosis factor α . Individual samples were run in duplicate and the mean values used in all graphs . Dexamethasone sodium phosphate ( Dexaject SP , Butler Animal Health Supply , Dublin , OH ) was diluted to 1 mg/ml in sterile saline and one group of mice were given 200 uL of this solution intraperitoneally ( i . p . ) , corresponding to a dose of 20 µg ( 10 mg/kg ) , two hours prior to UPEC infection , as previously described [80] , [81] . Another group of control mice were mock-treated i . p . with sterile saline alone . At 1 , 14 , or 28 dpi with either PBS or 108 cfu of KanR , all mice were treated with trimethoprim and sulfamethoxazole in the drinking water daily for 10 days at concentrations of 54 and 270 µg/ml , respectively [46] . During this time , longitudinal urinalysis was continued weekly to confirm clearance of bacteriuria . Four weeks after the initiation of antibiotic therapy , mice from each test group ( previously infected or naïve ) were challenged with either PBS or 107 cfu of UTI89 SpcR . Longitudinal urinalysis was then performed as for the primary infection ( except now triplicate plating on LB , LB/Kan25 and McConkey agar with 50 µg/ml spectinomycin ( McC/Spc50 ) to identify mice with persistent bacteriuria and the responsible strain . Mice were sacrificed 4 weeks after challenge and tissue titers determined as above , triplicate plating on LB , LB/Kan25 , and LB/Spc50 . Statistical analyses were performed using GraphPad Prism and InStat ( GraphPad Software ) and significance was defined by attaining P values<0 . 05 , in two-tailed tests where appropriate . Mouse Interleukin 5: P04401; Mouse Interleukin 6: P08505; Mouse Granulocyte Colony Stimulating Factor: P09920; Mouse Keratinocyte-derived Cytokine ( Growth-regulated alpha protein ) : P12850 . | The natural history of urinary tract infection ( UTI ) with uropathogenic E . coli in humans frequently includes persistent bacterial shedding in the urine for weeks after the initial infection , despite varying degrees of improvement of symptoms . Although antibiotic treatment has been very successful in treating UTI , antibiotic resistance is rising and recurrent infections are a common and costly problem affecting millions of women . Here , we examine an experimental mouse model of chronic bladder infection with uropathogenic E . coli , the most common cause of UTI . We found that the development of chronic bladder infection was preceded by severe bladder tissue inflammation that results in injury to the lining of the bladder . Furthermore , immunodeficient mice that lacked these acute inflammatory responses were protected from chronic bladder infection , suggesting that the development of chronic bladder infection requires immune-mediated tissue damage during acute infection . Finally , we demonstrate that mice with a history of chronic bladder infection that was subsequently cleared with antibiotic treatment are highly susceptible to further UTI . Thus , this study has discovered possible overlap in the mechanisms underlying the development of chronic and recurrent UTI . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"immunology/cellular",
"microbiology",
"and",
"pathogenesis",
"infectious",
"diseases/urological",
"infections",
"infectious",
"diseases",
"microbiology/immunity",
"to",
"infections",
"immunology/immune",
"response",
"microbiology/innate",
"immunity",
"immunology/innate",
"immunity",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis",
"immunology",
"microbiology",
"infectious",
"diseases/bacterial",
"infections",
"immunology/immunity",
"to",
"infections"
]
| 2010 | Early Severe Inflammatory Responses to Uropathogenic E. coli Predispose to Chronic and Recurrent Urinary Tract Infection |
Identification of novel cellular proteins as substrates to viral proteases would provide a new insight into the mechanism of cell–virus interplay . Eight nuclear proteins as potential targets for enterovirus 71 ( EV71 ) 3C protease ( 3Cpro ) cleavages were identified by 2D electrophoresis and MALDI-TOF analysis . Of these proteins , CstF-64 , which is a critical factor for 3′ pre-mRNA processing in a cell nucleus , was selected for further study . A time-course study to monitor the expression levels of CstF-64 in EV71-infected cells also revealed that the reduction of CstF-64 during virus infection was correlated with the production of viral 3Cpro . CstF-64 was cleaved in vitro by 3Cpro but neither by mutant 3Cpro ( in which the catalytic site was inactivated ) nor by another EV71 protease 2Apro . Serial mutagenesis was performed in CstF-64 , revealing that the 3Cpro cleavage sites are located at position 251 in the N-terminal P/G-rich domain and at multiple positions close to the C-terminus of CstF-64 ( around position 500 ) . An accumulation of unprocessed pre-mRNA and the depression of mature mRNA were observed in EV71-infected cells . An in vitro assay revealed the inhibition of the 3′-end pre-mRNA processing and polyadenylation in 3Cpro-treated nuclear extract , and this impairment was rescued by adding purified recombinant CstF-64 protein . In summing up the above results , we suggest that 3Cpro cleavage inactivates CstF-64 and impairs the host cell polyadenylation in vitro , as well as in virus-infected cells . This finding is , to our knowledge , the first to demonstrate that a picornavirus protein affects the polyadenylation of host mRNA .
Enterovirus 71 ( EV71 ) belongs to the family of Picornaviridae , to which poliovirus also belongs . EV71 infection usually causes childhood herpagina or exanthema , which is also called hand , foot , and mouth disease ( HFMD ) . Acute EV71 infection may also cause a severe polio-like neurological disease and significant mortality . EV71-related neurological complications include aseptic meningitis , brainstem and/or cerebellar encephalitis , acute flaccid paralysis ( AFP ) , myocarditis , and rapid fatal pulmonary edema and hemorrhage , which have been observed during outbreaks in Taiwan , mainland China , Malaysia , Brunei , Singapore , western Australia , the United States , and Europe [1]–[10] . Numerous host machineries are affected by picornaviral infection , including host cap-dependent translation [11] , [12] and transcription [13] , [14] . Viral proteases are responsible for inhibitory effects . Picornaviruses typically encode two viral proteases , 2Apro and 3Cpro , which are important for viral polypeptide processing . The major catalytic sites of type I poliovirus and EV71 3Cpro are His40 , Glu71 , and Cys147 [15]–[20] . In general , enterovirus 3C proteinases cleave at Gln/Gly scissile pairs [21] . Picornaviral 3Cpro can enter nuclei through its precursor 3CD′ or 3CD , which contains a nuclear localization sequence ( NLS ) [22] , [23] , and can cleave some cellular transcriptional factors or regulators , such as TATA-box binding protein , p53 , Histone H3 and transcription factor IIIC [13] , [14] , [24]–[27] , offering insight into the effects of these factors or regulators on the transcriptional machinery of the host . Therefore , the identification of other cellular proteins that are cleaved by EV71 3Cpro in nuclei is of interest , as it may help us to understand unknown viral-host interactions during EV71 infection . In this work , a novel nuclear factor , CstF-64 , that is cleaved by EV71 3Cpro , is identified . Poly ( A ) tails are important for both cellular mRNA and picornaviral mRNA . The poly ( A ) tail in eukaryotic cells is suggested to confer mRNA stability , promote the translational efficiency of mRNA , and the transport mRNA from the nucleus to the cytoplasm [28]–[31] . The poly ( A ) tail of poliovirus is critical to viral replication [32] , [33] . However , poly ( A ) for picornaviruses is obtained differently from that for host cells . The poly ( A ) tail of picornaviruses is encoded by the genome , copied into a 5′ poly ( U ) tract in the negative sense form , then back into a poly ( A ) tail during the virus RNA replication cycle [34] . Polyadenylation of most eukaryotes involves a series of steps . Prior to poly ( A ) synthesis , a polyadenylation signal located at the 3′ end of precursor mRNA ( pre-mRNA ) must be recognized by multiple factors such as cleavage factors I and II ( CF I and II ) , cleavage/polyadenylation specificity factor ( CPSF ) and cleavage stimulation factor ( CstF ) complex , which causes endonucleolytical cleavage at the site of polyadenylation [35] , [36] . One of these factors , cleavage stimulation factor , 3′ pre-RNA , subunit 2 , 64 kDa ( CstF-64 ) is responsible for recognizing the second polyadenylation sequence element , a G/U-rich motif which is down-stream of the polyadenylation sites on pre-mRNA [37] . In vitro and in vivo blockage of CstF-64 impairs the 3′-end pre-mRNA processing and polyadenylation [38] , [39] . This work identified eight cellular nuclear proteins as potential targets for EV71 3Cpro cleavage by two-dimensional ( 2D ) electrophoresis and MALDI-TOF analysis . CstF-64 , the target selected from these eight proteins , is further examined . CstF-64 in the nuclear extract was verified by western blot to be degraded upon EV71 3Cpro treatment . The reduction and cleavage pattern of CstF-64 in EV71-infected cells were studied . The EV71 3Cpro cleavage sites in CstF-64 were also mapped by performing an in vitro cleavage assay . The 3′-end pre-mRNA processing in host cells was monitored during EV71 infection . An in vitro assay was also conducted to test the effect of 3Cpro on the cellular 3′-end pre-mRNA processing and polyadenylation . The findings herein are the first to demonstrate that a picornaviral protease interferes with host gene expression at the machinery of polyadenylation .
Picornavirus 3Cpro proteins can enter the nuclei of host cells [22] , [23] . To identify potential substrates of EV71 3Cpro in nuclei , in vitro 3Cpro cleavage assay was conducted and the proteomic technique was utilized to identify the cellular targets . First , the catalytic activities of purified recombinant EV71 wild-type 3Cpro or mutant 3Cpro ( C147S ) were verified using a [35S]-labeled peptide substrate , which contains a 3Cpro cleavage site , as described elsewhere [19] . The result shows that the wild-type , but not the C147S mutant of 3Cpro cleaves the viral substrate ( Fig . 1A ) . The recombinant EV71 wild-type 3Cpro or mutant 3Cpro were then added to nuclear extracts from SF268 cells . Following incubation for 4 hours at 37°C , the reactions were subjected to 2D electrophoresis . The experiments were performed six times and the results were analyzed using PDquest 7 . 0 ( Bio-rad ) . Fig . 1B presents one of these six 2D electrophoresis experiments . Proteins that appeared in the mutant 3Cpro-treated reactants but in at least half of the amounts of the wild-type 3Cpro-treated reactants , based on silver staining , were regarded as potential targets . From the six pairs of gels in these 2D gel experiments , eight proteins that yielded similar results at least three times were selected . These proteins were then identified by MALDI-TOF mass spectrometry and the results were summarized in Table 1 . In contrast with that of mutant 3C-treated nuclear extract , CstF-64 of the wild-type 3Cpro-treated nuclear extract was decreased to nearly undetectable levels in five individual 2D electrophoresis gel ( Fig . 1C ) . Arrows in Figs . 1B and 1C indicate the locations of CstF-64 on 2D gels . These results suggest that CstF-64 is a potential substrate for EV71 3Cpro . The reduction of CstF-64 by proteomic assay was confirmed by using an anti-CstF-64 antibody [40] to detect the level of CstF-64 in nuclear extracts from SF268 cells , which were incubated with wild-type or mutant recombinant 3Cpro . The amount of CstF-64 in SF268 nuclear extract dramatically declined following incubation with wild-type 3Cpro ( Fig . 2A , lane 3 ) , and the intact form of CstF-64 was clearly detected in mutant 3Cpro-treated nuclear extract ( lane 2 ) . One potential cleavage product of 55 kDa was also detected using CstF-64 antibody in wild-type 3Cpro-incubated nuclear extract ( lanes 3 , 6 and 9 , indicated by * ) . CstF-64 levels in nuclear extracts from other cell lines , RD and HeLa cells , also declined upon incubation with wild-type 3Cpro and the cleavage pattern is similar to that in SF268 nuclear extracts . ( Compare lanes 6 and 9 with results for mutant 3Cpro-treated nuclear extracts in lanes 5 and 8 ) . The wild-type 3Cpro-induced CstF-64 reduction in nuclear extract depends on dose of recombinant 3Cpro ( Fig . 2B , lanes 3–7 ) . The intensities between the full-length CstF-64 and its 55 kDa cleavage product were not the same as expected . This discrepancy is attributed to that later in the experiments , the cleavage sites were mapped to position 251 and multiple positions close to the C-terminus ( around position 500 ) of CstF-64 . The cleavage of CstF-64 at positions around 500 produce a product of 55 kDa , which contains another 3Cpro cleavage site at its Gln251 . When position 251 was mutated ( Gln to Ala ) , the mutant CstF-64 was cleaved into a 55 kDa product with the same intensity as that of full-length CsfF-64 . Fig . 2A shows a band faster than 24 KDa ( lane 2 , 5 and 8 ) , which we speculate is a cleavage product from a degraded form of CstF-64 . The degraded form of CstF-64 might contain the cleavage sites for 3Cpro as well as the CstF-64 antibody recognition sites . Therefore , western blot analysis detected the cleaved product from the degraded form only in mutant 3Cpro , but not in wild-type 3Cpro-treated nuclear extract because the cleavage destroyed the antibody recognition of CstF-64 . The other EV71 viral protease , 2Apro was also tested to determine whether it could degrade CstF-64 in vitro . The catalytic activities of recombinant EV71 2Apro were firstly elucidated by treatment with RD cell lysate and were verified by western blot for a known 2Apro substrate , eIF-4GI , to be able to be cleaved in vitro ( Fig . 2C , lane 1–3 ) . The western blot for CstF-64 in the nuclear extracts that were treated by purified recombinant 2Apro demonstrated the inability of 2Apro to induce CstF-64 degradation in vitro ( Fig . 2C , lane 4–6 ) . These results indicate that viral 3Cpro , but not 2Apro is responsible for reducing CstF-64 in vitro . In vitro assays demonstrated that EV71 3Cpro can cleave CstF-64; the degradation of CstF-64 in EV71-infected cells was , therefore , examined . A time-course study was conducted to monitor the levels of CstF-64 in RD cells infected with EV71 at a multiplicity of infection ( m . o . i . ) of 40 . Total cell lysates harvested from EV71-infected RD cells at various hours post-infection ( h . p . i . ) were analyzed by western blot assay . Infection reduced the amount of host CstF-64 proteins from 6 to 10 hours post-infection ( h . p . i . ) ( Fig . 3A , lanes 4 , 6 and 8 ) . One potential CstF-64 cleavage product of 55 kDa ( indicated by * ) which resembles that obtained by in vitro cleavage , began to be detected at 6 h . p . i . ( lane 4 ) . EV71 3Cpro in the same infected cells was also observed at 6 h . p . i . ( Fig . 3A , lane 4 ) , and its level increased markedly from 8 to 10 h . p . i . ( Fig . 3A , lane 6–8 ) , in a manner that is related to the degradation times of CstF-64 in infected cells . The reduction of CstF-64 and the appearance of 55-kDa product were also observed in cells infected with a lower titer ( m . o . i of 1 ) of EV71 ( Fig . S1 ) , but with a 2 hour delay . To detect other cleavage products from CstF-64 , a CstF-64 that was fused with FLAG at its N-terminal was overexpressed in EV71-infected cells and detected using FLAG antibody at 6 and 8 h . p . i . . During EV71 infection , FLAG-CstF-64 was cleaved into products of size 55 kDa and 30 kDa ( CP1 and CP2 in Fig . 3B ) . The 55 kDa products from FLAG-CstF-64 reduced at late point of EV71 infection ( Fig . 3B , CP1 , compare lanes 2 and 4 ) , similar to the result from endogenous CstF-64 in EV71-infected cells ( Fig . 3A , the bands marked as * in lane 4 and 6 ) . On the other hand , another cleavage product of about 30 kDa was increased during virus infection ( Fig . 3B , CP2 compare lanes 2 and 4 ) . The 55 kDa intermediate cleavage products suggest that CstF-64 may be cleaved at more than one cleavage site in EV71-infected cells . Picornavirus infection can relocalize numerous nuclear proteins to the cytoplasm [41]–[43] . To determine whether CstF-64 also undergoes changes in sub-cellular localization , the location of CstF-64 in EV71-infected was monitored by immunohistostaining using a confocal microscope . The result demonstrated that most CstF-64 remains in the nucleus of the infected cells at 6 to 10 h . p . i . ( Fig . 3D ) , whereas the other nuclear factor , heterogeneous nuclear ribonucleoprotein K ( hnRNP K ) , is redistributed to the cytoplasm ( Fig . S2 ) . Our 3C Ab was unable to be used in confocal experiment; therefore to determine whether EV71 3Cpro was also present in the nucleus , EV71-infected cells were fractionated into cytoplasmic and nucleus fractions . Western blot demonstrated that EV71 3Cpro is equally distributed across cytoplasmic and nuclear fractions ( Fig . 3C , lanes 2 and 4 ) , indicating that EV71 3Cpro can enter into a host nucleus during virus infection , as can other picornaviruses [44] . CstF-64 was detected to be localized primarily in the nuclei of uninfected cells ( Fig . 3C , lane 3 compared to lane 1 ) , consistent with previous findings [38] and confocal data presented herein ( Fig . 3D ) . However , CstF-64 in both nuclei and cytoplasm were cleaved during virus infection ( Fig . 3C , lanes 2 and 4 ) . The results presented above suggest that the CstF-64 was reduced in wild-type 3Cpro-treated nuclear extracts but not in mutant 3Cpro-treated nuclear extracts ( Fig . 2A ) . Therefore , 3Cpro is reasonably hypothesized to cause the reduction of CstF-64 by proteolytic cleavage . Details of the cleavage were elucidated using a [35S]-labeled CstF-64 generated by in vitro transcription and translation ( TNT ) as a substrate in the in vitro 3Cpro cleavage assay . After it had incubated with recombinant 3Cpro , [35S]-labeled CstF-64 can be cleaved into fragments of approximately 25 and 30 kDa ( Fig . 4A , lane 3 ) . Since the size of these two peptides do not correlate with the full-length size of CstF-64 ( 64 kDa compared to 25 kDa+30 kDa = 55 kDa ) , at least one cleavage product smaller than 10 kDa that could not be detected in PAGEs was expected to exist . To identify the 3Cpro cleavage sites on CstF-64 , [35S]-labeled CstF-64 and a series of truncated CstF-64 peptides that were designed according to functional domains of CstF-64 ( Fig . 4B ) were generated in vitro by TNT for 3Cpro cleavage assay . Fig . 4B summarizes the size of these cleavage products , estimated from Fig . 4A . The peptides 221–409 and 410–557 were both cleaved suggesting that CstF-64 included at least two cleavage sites . The 20 kDa peptide 410–557 was cleaved into the product of 10 kDa , suggesting that a cleavage site was located close to the C-terminus of CstF-64 . Peptide 1−220 ( 25 kDa ) was not cleaved but peptides 221−409 and 110−409 were both cleaved into a product of about 17 kDa , suggesting that one cleavage site can produce a product of 30 kDa from the N-terminal of CstF-64 . The cleavage of peptide 221−469 supported this conclusion . In summary , these cleavage results suggest the existence at least two potential cleavage sites around the amino acid position 250 and 500 of the CstF-64 protein . A high amount of 3Cpro ( 5 µg ) was used in above in vitro cleavage assay , and the incubation time was fixed at 3 hours ( Fig . 4A ) . An attempt was made to detect the cleavage intermediates by performing in vitro kinetics analysis of [35S]-labeled CstF-64 cleavage by recombinant 3Cpro . According to Fig . 4C , three cleavage products began to appear at 15 min . A 55 kDa product resulted from cleavage at positions around 500 , while two products of 30 ( front ) and 35 ( rear ) kDa both resulted from cleavage at position around 250 . A new 25 kDa product began to appear at 60 min , in addition to the three previously detected products . This 25 kDa segment is attributed to the further cleaving of either the aforementioned 35 ( rear ) kDa product at positions around 500 or the 55 kDa product at position around 250 . As the incubation time exceeded 2 hours , only the 25 and 30 kDa products dominated and became the final cleavage products . Various concentrations of 3Cpro were added to the reactions ( Fig . S3 ) , and similar cleavage products were observed , as shown in Fig . 4C . Picornavirus 3Cpro generally cleaves peptides at the Gln/Gly junction [21] . Based on the cleavage results presented above , an analysis of the amino acid sequence of CstF-64 was conducted , revealing that numerous Gln/Gly junctions may have potential 3Cpro cleavage sites on CstF-64 ( Fig . 5A ) . They are Gln251 , corresponding to the predicted cleavage site in the region of amino acid 250 , and residues Gln483 , 496 , 505 , 510 , 515 at around the amino acid position 500 on CstF-64 . To determine which Gln/Gly junction on CstF-64 is the actual 3Cpro cleavage site , an attempt was made to produce mutant [35S]-labeled CstF-64 at putative Gln/X cleavage sites . Since two cleavage sites around the amino acids positions 250 and 500 of CstF-64 were estimated , cleavage at position around 250 is expected to produce a product of ∼55 KDa . However , as is predicted , a CstF-64 in which both 250 and 500 sites are mutated would be fully withstand 3Cpro cleavage ( Fig . 5B ) . Following 3Cpro-treatment , the CstF-64 with the mutation of Gln251 into Ala ( Q251A ) yielded a cleavage product of 55 KDa ( Fig . 5C , lanes 1–3 ) , which is consistent with the predicted product when the cleavage site around amino acid position 250 is blocked ( Fig . 5B ) . This result suggests that the Gln251 of CstF-64 is the only 3Cpro cleavage site around the position 250 in vitro . To identify the other cleavage sites around the amino acid position 500 of CstF-64 , CstF-64 ( Q251A ) was further mutated at the Gln483 , 496 , 505 , 510 and 515 amino acids into Ala ( Q483A , Q496A , Q505A , Q510A and Q515A , respectively ) . The results indicate that CstF-64 ( Q251A ) peptides with single mutations on these potential cleavage sites are not completely resistant to wild-type ( WT ) 3Cpro treatment ( Fig . 5C , lanes 4–18 ) . A mutant CstF-64 ( Q251A ) with Q483A , Q496A , Q505A , Q510A , and Q515A is fully resistant to 3Cpro treatment ( Fig . 5D , lanes 4–6 ) , suggesting that the existence of multiple cleavage sites around the amino acid position 500 of CstF-64 in the five residues , Gln483 , 496 , 505 , 510 and 515 . Moreover , the cleavage pattern was examined in a mutant CstF-64 ( CstF-64-5 m ) , in which all Gln/Gly sites around position 500 were mutated . This observation indicates that CstF-64-5 m was cleaved into two detectable products ( Fig . S4 ) : a ) 30 kDa product , i . e . the N-terminal part of CstF-64 after 3Cpro cleaves it at Gln/Gly of position 251 and b ) the 35 KDa product , i . e . the C-terminal part of CstF-64 when 3Cpro cleaves it at Gln/Gly of position 251 , but not at Gln/Gly sites near position 500 . Taken together , above results demonstrate that Gln to Ala mutations at position 251 and multiple sites around position 500 let CstF-64 become fully resistant to EV71 3Cpro cleavage . CstF-64 is an important factor for host 3′ pre-mRNA processing and RNA polyadenylation [35] , [38] . Results presented above demonstrate the reduction of CstF-64 protein in EV71-infected cells; therefore , whether the machinery of host cell polyadenylation is affected by EV71 infection was further examined . Initially , the pre-mRNA processing in EV71-infected cells was assessed using an artificial exogenous expressed GFP RNA . The pEGFP plasmid contained a SV40 polyadenylation signal sequence , which was commonly used in the study of cellular mRNA processing and polyadenylation . Primers that target GFP pre-mRNA and an oligo-dT primer that targets GFP poly ( A ) mRNA were used in RT-PCR for detecting RNA species ( pre-mRNA and mature mRNA with poly ( A ) tail ) in EV71-infected cells ( Fig . 6A ) . A set of primers that target the GFP coding region , which can theoretically detect both pre-mRNA and mRNA with a poly ( A ) tail , were also utilized as controls to demonstrate the total amount of GFP RNA in cells . RD cells were infected with EV71 after they were transfected with eGFP plasmid . Additionally , by adopting those primers designed in Fig . 6A , RT-PCR analysis was performed for semi-quantifying different GFP RNA species in these infected cells ( Fig . 6B ) . After they were calibrated to the total amount of GFP , the results revealed the accumulation of unprocessed pre-mRNA and a reduction in polyadenylated mRNA in EV71-infected cells . An attempt was made to identify if EV71 infection affected endogeneous pre-mRNA processing . Interlekin 10 receptor beta ( IL-10RB ) was selected as the target because , in an earlier study , the poly ( A ) mRNA of IL-10RB was found to be decreased in EV71-infected cells by cDNA microarray [45] . The relative amounts of IL-10RB pre-mRNA and poly ( A ) mRNA were monitored by real-time RT-PCR . According to those results , the IL-10RB pre-mRNA in EV71-infected cells decreased to 62 . 95% of that in mock-infected cells , which correlates with our previous micorarray results . Additionally , the relative amount of IL-10RB pre-mRNA increased to 156% of that in mock-infected cells ( Fig . 6C ) . Above results suggest that EV71 infection impairs the CstF-64-related 3′ pre-mRNA processing mechanism . The results herein suggest that EV71 3Cpro cleaves CstF-64 , potentially inhibiting cellular pre-mRNA 3′-end formation and polyadenylation . An in vitro assay was conducted to test the effect of EV71 3Cpro on HeLa nuclear extract . The substrate employed in this assay is a capped pre-mRNA , which contains an SV40 late gene polyadenylational cleavage site [38] . A comparison with the mutant 3Cpro-treated nuclear extracts ( Fig . 7A , lane 5 and Fig . 7B , lane 5 ) or untreated nuclear extracts ( Fig . 7A , lanes 2 , 3 and Fig . 7B , lanes 2 , 3 ) shows that the cleavage of pre-mRNA and polyadenylation proceeded efficiently . However , upon wild-type 3Cpro treatment , the treated nuclear extracts lost the ability to perform pre-mRNA cleavage ( Fig . 7A , lane 4 ) and polyadenylation ( Fig . 7B , lane 4 ) . Moreover , the impairments were rescued by adding purified recombinant CstF-64 protein ( Fig . 7A , lane 6 and Fig . 7B , lane 6 ) . The results suggest that EV71 3Cpro targeting to CstF-64 is a factor that inhibits cellular 3′-end pre-mRNA processing and polyadenylation .
Picornaviral 3Cpro reportedly influences numerous cellular functions by cleaving various host proteins , which mechanism is important in shutting off cellular gene expression in transcriptional and translational levels [13] , [14] , [24]–[27] . Here we show that EV71 3Cpro may affect another cellular mechanism , polyadenylation , since a novel substrate , CstF-64 , was identified using the proteomic approach . CstF-64 is a crucial element in cellular mRNA maturation [38] . This 64 kDa component plays a role in the CstF complex that binds to the polyadenylation signal region of pre-mRNA [37] . Our in vitro cleavage studies revealed that 3Cpro cleaves CstF-64 at amino acid position 251 and around 500 close to the C-terminal domain ( Figs . 4 and 5 ) . The in vitro data also demonstrated the blockage of pre-mRNA 3′-end processing and the polyadenylayion activities of the nuclear extract upon 3Cpro-treatment ( Fig . 7 ) . The inhibitory effect of EV-71 3Cpro resembles that of the blockage of CstF-64 by the specific antibody [38] . The results together suggest that EV71 3Cpro cleaves CstF-64 and thus inhibits the 3′-end processing of cellular pre-mRNA . Our semi-quantitative RT-PCR results support this claim because the cellular 3′-end pre-mRNA processing was inhibited in EV71-infected cells ( Fig . 6B ) . Pre-mRNA and mature poly ( A ) mRNA in Fig . 6B were detected individually by using two pairs of primers . Different primers have varying sensitivities , making it difficult to achieve a quantitative correlation between the increase in precursor and decrease in polyadenylated RNA . However , a quantivitive real-time RT-PCR result ( Fig . 6C ) provides further evidence to demonstrate that the pre-mRNA processing or polyadenylation ( either an exogenous RNA or an endogenous RNA ) was impaired in EV71-infected cells . Other proteins than CstF-64 may be involved in EV71-induced inhibition of pre-mRNA processing in virus-infected cells even though the levels of poly ( A ) polymerase in EV71-infected cells were found to be similar to those of the mock-infected cells ( Fig . S5 ) . However , our in vitro results in Fig . 7 demonstrate that CstF-64 is the target factor that caused 3Cpro to affect 3′-end processing and polyadenylation . Adding recombinant CstF-64 can restore the function of 3C-incubated nuclear extracts on 3′-end pre-mRNA processing and polyadenylation ( Lane 6 in both Fig . 7A and B ) . Before the surge of the proteomic method , a substrate for a protease is commonly identified using western blot analysis . The proteomic approach has the advantages of the ability to identify multiple candidates simultaneously with no limitation by antibody availability . However , 2D electrophoresis and the silver staining are occasionally unable to detect proteins visible in western blot analysis [46] . This study identified eight novel proteins as the substrates for EV71 3Cpro via the proteomic approach , in which no previously known substrates for other picornaviral 3Cpro were available , such as TATA-box binding protein , p53 , Histone H3 and transcription factor IIIC [13] , [14] , [26] , [27] . This might be due to that silver staining and western blot differ in sensitivity . Moreover , the strategy adopted in Fig . 1 is set only in a single condition , i . e . the recombinant 3Cpro was incubated with nuclear extract from SF268 cells for 4 hours . More potential substrates could be identified if several interaction conditions were tested , e . g . , varying in incubation time and nuclear extracts from other cell lines . The in vitro cleavage assay revealed that 3Cpro cleaves [35S]-labeled CstF-64 protein into detectable products of 25 kDa and 30 kDa ( Fig . 4B ) . However , CstF-64 in 3Cpro-treated nuclear extract yielded 55 kDa products detected by western blot ( Fig . 2 ) . A product similar to that of 55 kDa was also present in EV71-infected cells ( Fig . 3A ) . The size of the 55 kDa products above is correlated to that of the 3Cpro cleavage product when the cleavage site of the amino acid position 251 on CstF-64 is mutated ( Fig . 5C ) . The detection of FLAG in FLAG-CstF-64 overexpressed cells revealed that the proportion of 30 kDa increased and that of 55 kDa declined upon EV71-infection ( Fig . 3B ) . A FLAG-CstF-64 with Q251A mutation was transfected into EV71-infected cells and the results indicated the accumulation of 55 kDa products but no 30 KDa product was detected ( data not shown ) , suggesting that the 55 kDa component was an intermediate cleavage product when 3Cpro cleaved CstF-64 at an amino acid position ∼500 . The production of multiple 55 kDa bands in infected cells ( Figs . 3A and B ) was also observed , explaining the multiple cleavage sites around the amino acid position 500 of CstF-64 . In addition , the results of in vitro kinetics of CstF-64 cleavage by 3Cpro ( Fig . 4C and Fig . S3 ) showed a cleavage intermediate of 55 kDa . In summary , these results showed the similarity between the cleavage patterns obtained in the in vitro and in vivo assays suggests that 3Cpro may have reduced CstF-64 levels via proteolytic activity in infected cells . Single mutation at position around 500 in CstF-64 still made it susceptible to EV71 3Cpro cleavage , while multiple mutations cause CstF-64 become resistant to the cleavage . Analytical results indicate the C-terminal of CstF-64 may contain multiple cleavage sites; otherwise , these mutations could lead to a loss of exposure of the cleavage site . This study also attempted to examine whether C147S mutation in EV71 impairs the viral inhibition of cellular pre-mRNA processing and polyadenylation . However , C147S mutation rendered EV71 lethal . Alternatively , transcripts derived from EV71 infectious clones , i . e . wild-type versus C147S mutant , were transfected into cells to examine how they affect cellular pre-mRNA processing and polyadenylation . In wild-type transcript -transfected cells , CstF-64 was cleaved and cellular pre-mRNA processing was inhibited ( data not shown ) , the same phenomenon was observed in EV71 virus-infected cells . However , in C147S mutant transcript -transfected cells , no cleavage occurred in CstF-64 and no effect on cellular pre-mRNA processing was observed , which is similar to the results in mock-infected cells . Still , no detectable viral protein ( mature 3Cpro ) was found . We believe that phenomenon is owing to that C147S mutation destroys the proteolytic activity of 3Cpro and inhibits viral protein processing and maturation . Therefore , we can not conclude that the C147S mutant in EV71-infected cells impairs the viral inhibition of cellular pre-mRNA processing . CstF-64 reportedly contains at its N-terminal an RNA-binding domain and a hinge domain , which are responsible for the reorganization of RNA and its binding to CstF-77 or other factors [47]–[50] . No 3Cpro cleavage was detected in this region ( 1–220 ) , suggesting that the cleaved CstF-64 maintains functions RNA binding and interaction with CstF-77 ( Fig . 4B ) . However , the Pro/Gly-rich domains of the C-terminal region of CstF-64 form an extended α-helix structure and are suspected to be able to interact with several transcription factors [47] , including binding of the transcription factor PC4 via the 100 amino acids of C-terminal CstF-64 [51] . One study noted that the truncation of the C-terminal domain ( 529–577 ) of CstF-64 inhibited cellular 3′ pre-mRNA processing [52] . Accordingly , multiple cleavage by 3Cpro in C-terminal domain ( around 500th a . a . ) of CstF-64 is expected to destroy the cellular pre-mRNA processing functions of CstF-64 . The RNA virus-induced inhibition of 3′-end pre-mRNA processing has been observed in the influenza A virus . NS1A of influenza A reportedly inhibits the cellular 3′-end pre-mRNA processing by interacting with CPSF-30 which is within the same machinery co-factor as CstF-64 [53] . Similar to those in EV71-infected cells , the inhibition of the cellular 3′-end pre-mRNA processing by influenza A virus causes the loss of poly ( A ) mRNA and the accumulation of un-poly ( A ) pre-mRNA . This inhibitory effect of NS1A has also been suggested to be involved in the countering of cellular antiviral responses , including virus-induced interferon β production [54] . The capacity of NS1A to bind to CPSF-30 may also be crucial to viral replication because of high interferon-β mRNA production in NS1A mutant virus-infected cells [55] . Picornaviruses influence host-cell gene expression by the inhibition of cellular transcription and cap-dependent mRNA translation [11]–[13] . However , several cellular genes could escape the inhibition of gene expression by picornavirus infection . Our previous cDNA microarray analysis for total cellular RNA demonstrated that the level of some RNAs , related to chemokines , protein degradation , complement proteins and proapoptosis proteins increased upon EV-71 infection [45] , suggesting leakage from the inhibition of transcription by EV71 . Translations of of c-myc , Bip , and eIF4G mRNA have been found to be increased in poliovirus-infected cells as the cap-dependent translation shuts down , because of the presence of internal ribosome entry sites ( IRES ) [56]–[58] . This work proposes a novel mechanism by which picornavirus inhibits cellular gene expression in the 3′-end pre-mRNA processing step , which would probably inhibit polyadenylation of the surviving host RNAs . It has been reported that the poly ( A ) tail of poliovirus is essential to viral replication [32] , [33] , which suggests several host poly ( A ) -associated factors are important for viral growth . Inhibition of cellular poly ( A ) RNA synthesis may provide yet an additional advantage for virus replication . In conclusion , a novel mechanism by which picornavirus inhibits cellular function was identified . CstF-64 was identified as an EV71 3Cpro substrate . The 3Cpro cleavage sites were mapped onto amino acid position 251 and the C-terminal region of CstF-64 . The cleavage of CstF-64 impairs the cellular 3′-end pre-mRNA processing and polyadenylaytion . EV71 utilizes 3Cpro to interfere polyadenylation of host cellular RNA; however , its own poly ( A ) synthesis is not affected since picornaviral poly ( A ) tail is genetically encoded [32] , [33] .
RD and SF268 cells were maintained in DMEM that contained 10% FBS ( Glibco ) at 37°C . Cells were split in a ratio of 1∶10 in fresh medium every three days . After the RD cells had grown to 80% confluence , they were washed in PBS . EV71/4643/TW viruses with a multiplicity of infection ( m . o . i . ) of 40 were inoculated into the cells in DMEM without FBS at room temperature for one hour to infect the cells with the virus . Following one hour of incubation , viruses and the medium were replaced by DMEM that contained 2% fetal bovine serum . Infected cells were cultured at 35°C . pET-23-EV71-3C and pET-23-EV71-m3C-C147S plasmid constructs , which contain wild-type 3Cpro or mutant 3Cpro ( with Cys147 replaced by Ser ) cDNA in pET-23a ( + ) vector for recombinant EV71 3Cpro production have been constructed elsewhere [19] . For CstF-64 recombinant production , cDNA of CstF-64 was cloned from the Cst-64 coding region that contained the plasmid , pZ64–18 , which was a gift from Dr . James L . Manely , and then inserted into vector pET23a ( + ) using EcoRV and NotI . For the production of [35S]-labeled substrates by TNT assay , full-length and partial cDNA of CstF-64 were cloned and inserted into pcDNA3 . 1 ( + ) using EcoRV and NotI . For the production of FLAG-CstF-64 in transfected cells , cDNA of CstF-64 was cloned into pFLAG-CMV-2 using EcoRV and NotI . The mutant CstF-64 construct was produced using a QuikChange Site-Directed Mutagenesis Kit ( Stratagene ) . Plasmid pG3SVL-A , which contained sequence -142 to +138 of SV40 late gene poly ( A ) cleavage sites was kindly provided by Dr . James L . Manely and Dr . Yoshio Takagaki , for the in vitro 3′-end pre-mRNA processing and polyadenylation assay . PVDF membrane was blocked wih Tris-buffered saline/0 . 1% ( vol/vol ) Tween 20 that contained 5% non-fat dry milk and probed with the indicated antibody . Antibodies against CstF-64 ( 1∶200; from Dr . Clinton C . MacDonald or 1∶2000; Santa Cruz ) , FLAG ( 1∶2000; SIGMA ) , PCNA ( 1∶2000; Santa Cruz ) , HDAC ( 1∶2000; Santa Cruz ) , Poly ( A ) polymerase ( 1∶2000; Santa Cruz ) , actin ( 1∶4000; Chemicon ) and EV71 3Cpro monoclonal antibody which generated from recombinant 3Cpro protein by our lab ( 1∶50 ) were used . Following washing , the membranes were incubated with HRP-conjugated anti-mouse or HRP-conjugated anti-rabbit ( 1∶2000 ) . HRP was detected using a Lighting Chemiluminescence reagent ( Amersham Pharmacia ) . To perform the cleavage assay in a cell-free system , SF268 , HeLa or RD cells were washed in PBS and scraped out . Packed cells were re-suspended in hypotonic buffer ( 10 mM HEPES pH 7 . 9 , 1 . 5 mM MgCl2 , 10 mM KCl and 0 . 5 mM DTT ) and homogenized using a 25G needle . Homogenized cell fragments were centrifuged at 3300 g to remove cytoplasmic proteins . The pellet was washed twice in hypotonic buffer and the soluble nuclear proteins were extracted by adding buffer with a graded conc . of KCl ( 20 mM HEPES , 25% glycerol , 1 . 5 mM MgCl2 , 0 . 2 mM EDTA , 0 . 5 mM DTT , 0 . 02 to 1 . 6 M KCl at pH = 7 . 9 ) . After centrifuged at 25000 g , the extracted nuclear proteins in supernatant were harvested and dialyzed in buffer ( 20 m M HEPES , 20% glycerol , 100 mM KCl , 0 . 2 mM EDTA and 0 . 5 mM DTT at pH = 7 . 9 ) and then stored at −80°C . Commercial reagents for nuclear extraction ( CMN compartment protein extraction kit; Biochain ) were applied to prepare the nuclear extract from EV71-infected cells . Commercial HeLa nuclear extract ( Santa Cruz ) was used in in vitro 3′-end pre-mRNA processing and polyadenylation assay . To produce recombinant 3Cpro and CstF-64 proteins , plasmid pET-23-EV71-3C , pET-23-EV71-m3C-C147S and pET-23-CstF-64 , were introduced into competent E . coli BL21 ( DE3 pLysS ) and protein expression was induced using 40 µM isopropyl b-D-thiogalactopyranoside . 3C-His fusion proteins were purified using a HiTrap kit ( Pharmacia ) . To produce [35S]-labeled protein , TNT Coupled Reticulocyte Lysate Systems ( Promega ) was used in the in vitro transcription and translation reaction . In the proteomic approach to screening potential 3Cpro substrates of , 30 µg of 3Cpro was incubated with 250 µg nuclear extracts in digestion buffer ( 20 mM HEPES , 20% glycerol and 100 mM KCl at pH = 7 . 9 ) with total volume of 100 µl at 37°C for four hours . In western blot assay , 10 µg of 3Cpro was incubated with 50 µg nuclear extract under the aforementioned conditions . To cleave recombinant CstF-64 , 10 µg recombinant CstF-64 proteins interacted with 10 µg of 3Cpro for four hours in the same buffer as in the earlier experiments . To cleave the [35S]-labeled substrate , 4 µl of labeled protein from one TNT assay reaction was incubated with 5 µg of 3Cpro in buffer ( 20 mM HEPES , 20% glycerol and 100 mM KCl at pH = 7 . 9 ) with a total volume of 15 µL at 37°C for two hours . 2D gel electrophoresis was conducted using the IPGphor system ( Amersham Biosciences ) . After the nuclear extracts had incubated with 3Cpro and precipitated using 20% trichloroacetic acid ( TCA ) , the sample was dissolved in 8 M Urea , 2% CHAPS , 0 . 5% IPG buffer ( Amersham Bioscience ) and bromophenol blue . These samples were loaded on a cup-loading system ( Amersham Biosciences ) . Isoelectric focusing was performed at 45 , 000 Vh in a stepwise fashion ( 1500 V , gradient for 2 hours; 4000 V , gradient for 3500 V . hours; 8000 V , step-on-hold until the end ) in Immobiline Drystrip IEF gels with the range of pH 3–10 ( Amersham Biosciences ) . For SDS-PAGE , the IEF gels were loaded on gels with graded acryamide conc . of 9% to 16% . Proteins on 2D electrophoresis gels were visualized by silver staining and the results were scanned as TIF files . Protein spots on scanned gels were analyzed using PDquest v7 . 0 ( Bio-rad ) software . Six pairs of gels with a similarity of over 60% were selected for further analysis . Spots that appeared in the mutant 3Cpro-treated nuclear extracts disappeared or became at least 50% smaller upon silver staining of the wild-type 3Cpro-treated nuclear extracts were selected as the potential target substrates for wild-type 3Cpro . Following this selection strategy , proteins that yielded similar results to each other more than three times were identified by in-gel digestion and analyzed by Bruker Ultraflex MALDI-TOF mass spectrometry . Mass lists were performed peptide mass fingerprinting by Biotool 2 . 0 software and the algorithm of Masscot ( http://www . matrixscience . com ) . RD cells were seeded onto 22 mm diameter coverslips to 80% confluence and were infected with enterovirus 71 ( strain 4643/TW/1998 ) at a multiplicity of infection of 40 . At each time point , the culture medium was removed and the cells were washed with PBS , and fixed with 3 . 7% formaldehyde for 20 min at room temperature . The cells were then washed with PBS and permeabilized using 0 . 3% Triton X-100 for 5 min at room temperature . Cells were washed once using PBS containing 2% FBS , incubated in blocking solution for 1 hr at room temperature . In CstF-64 , hnRNP K and EV71 2B immunostaining , cells were incubated with anti-CstF-64 ( diluted 1∶100; Santa Cruz ) , anti-EV71 2B ( diluted 1∶200; Provided by Dr . Jim-Tong Horng ) for 8 hours at 4°C or anti-hnRNP K ( diluted 1∶200; Santa Cruz ) for 1 . 5 hr at room temperature . After they had been washed three times with PBS , coverslips were incubated with FITC-conjugated goat anti-mouse IgG or goat anti-rabbit lgG , and Alexa Fluor 568 goat anti-rat IgG ( Invitrogen ) for 1 hr at room temperature . The coverslips were then washed once with PBS; treated with nuclear stain Hoechst 33258 ( diluted 1∶500 ) for 15 min; washed three times with PBS , and mounted on glass slides with mounting fluid ( 75% glycerol in PBS ) . The images were obtained under a confocal laser-scanning microscope ( Zeiss; LSM 510 NLO ) . After they had been transfected with pEGFP-N1 for 1 hour , RD cells were infected with EV71 virus . The total RNA of these infected cells were isolated by TRIZOL Reagent ( Invitrogen ) and purified by phenol/chloroform extraction . To detect EGFP pre-mRNA , a set of primers 5′- CCGGAATTCTGAGCAAAGACCCCAACGAG -3′ and 5′- CCCAAGCTTAAAATATTAACGCTTACAAT-3′ , which target the sequence that was removed from pre-mRNA following polyadenylation was used . To detect poly ( A ) EGFP mRNA , a set of primers 5′- CCGGAATTCTGAGCAAAGACCCCAACGAG -3′ and 5′- TTTTTTTTTTTTTTTGCAGT-3′ , which target the poly ( A ) tail and polyadenylation site of EGFP mRNA , were used . To detect total EGFP RNA , a set of primers 5′-ATGGTGAGCAAGGGCGAGGA-3′ and 5′-CTTGTACAGCTCGTCCATGC-3′ , which target the coding region of EGFP was used . The RT-PCR products were placed in 2% agarose gel with ethidium bromide , and semi-quantified using software FUJIFILM Science Lab 2005 ( Fuji ) . Capped RNA substrate with a poly ( A ) signal of the SV40 late gene was synthesized using a Megascript SP6 kit ( Ambion ) and Cap analog ( Ambion ) . Plasmid pG3SVL-6 , flanked by DraI , was utilized as the template in the in vitro transcription reaction . The synthesized capped RNA substrate was purified by phenol/chloroform extraction before it was used in in vitro 3′-end pre-mRNA processing and polyadenylation assay . The nuclear extract was treated with recombinant 3Cpro at 37°C for two hours before it was used in 3′ pre-mRNA processing and polyadenylation assay . After the treatment , the 3Cpro activity was stopped by adding protease inhibit cocktail ( Roche ) . The volume of each 3′ pre-mRNA processing reaction with 20 µl contained 25% ( v/v ) nuclear extract ( 25 µg in total ) , and 2 ng of RNA subtrates with added 1 mM EDTA , 0 . 5 mM dATP , 10 mM creatine phosphate , 2 . 5% ( v/v ) of polyvinyl alcohol . In in vitro polyadenylation , dATP was replaced by ATP , and the conc . of EDTA declined to 0 . 08 mM with the addition of 1 mM MgCl2 . 4 µg of recombinant CstF-64 was added into one of the 3Cpro-treated reaction . After incubation for two hours at 37°C , the assayed RNA was purified phenol/chloroform extraction . The purified RNA was analyzed in 1 . 5% agarose gel and denatured using formaldehyde . Real-time RT-PCR was performed to detect levels of IL-10RB RNA . The High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) was applied for reverse transcription . To differentiate pre-mRNA and polyadenylated IL-10RB RNA , different primers were used in reverse transcription step . A specific primer 5′-CTGAGAGCTCCCAGATGACTGA-3′ that targets the region that contains polyadenylation cleavage signal on IL-10RB was used to detect the pre-mRNA . On the other hand , an oligo-d ( T ) was used as the primer to detect poly ( A ) -mRNA . For total mRNA ( including pre-mRNA and poly ( A ) -mRNA ) , a random hexomer was used an the primer in reverse transcription step . For PCR step , 5′-GAGGGATCAGGGCAGCAA-3′ and 5′-CAGGGTCTGGGAGTTCTAGATGTG-3′ designed by Primer Express software ( Applied Biosystems ) targeting IL-10RB coding region were used . The PCR reaction was performed on 7500 Real-Time PCR System ( Applied Biosystem ) using SYBR Green Core Kit ( Applied Biosystem ) . | Many viruses contain specific proteases that are essential for processing their own viral proteins . For an efficient replication within their hosts , on the other hand , viruses also utilize these proteases to cleave a number of key host proteins and hijack cellular machineries . In this study , host proteins are identified as the substrates for enterovirus 71 viral protease by adopting a proteomic strategy . Enterovirus 71 infection is highly associated with neurological complication and mortality in an era when poliovirus has been controlled by vaccination . Investigating the host substrates for enterovirus 71 protease may shed light on the viral–host interaction , ultimately providing further insight into the pathogenesis of EV71 infection . We found that 3C protease ( 3Cpro ) cleaved CstF-64 , the latter a cleavage stimulation factor in host polyadenylation machinery . 3Cpro is known to enter host nucleus , whereas the replication of the virus occurs in cytoplasm . While its role in nucleus has been thought to inhibit host transcription , we found that 3Cpro may inhibit host-cell gene expression at host 3′-end pre-mRNA processing and polyadenylation steps . Consequently , less polyadenylated host mRNA is synthesized and more cellular resources , such as the translation factors , would be available for viral RNA expression . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"virology/effects",
"of",
"virus",
"infection",
"on",
"host",
"gene",
"expression",
"virology"
]
| 2009 | Enterovirus 71 3C Protease Cleaves a Novel Target CstF-64 and Inhibits Cellular Polyadenylation |
Gastropod-borne parasites may cause debilitating clinical conditions in animals and humans following the consumption of infected intermediate or paratenic hosts . However , the ingestion of fresh vegetables contaminated by snail mucus and/or water has also been proposed as a source of the infection for some zoonotic metastrongyloids ( e . g . , Angiostrongylus cantonensis ) . In the meantime , the feline lungworms Aelurostrongylus abstrusus and Troglostrongylus brevior are increasingly spreading among cat populations , along with their gastropod intermediate hosts . The aim of this study was to assess the potential of alternative transmission pathways for A . abstrusus and T . brevior L3 via the mucus of infected Helix aspersa snails and the water where gastropods died . In addition , the histological examination of snail specimens provided information on the larval localization and inflammatory reactions in the intermediate host . Twenty-four specimens of H . aspersa received ~500 L1 of A . abstrusus and T . brevior , and were assigned to six study groups . Snails were subjected to different mechanical and chemical stimuli throughout 20 days in order to elicit the production of mucus . At the end of the study , gastropods were submerged in tap water and the sediment was observed for lungworm larvae for three consecutive days . Finally , snails were artificially digested and recovered larvae were counted and morphologically and molecularly identified . The anatomical localization of A . abstrusus and T . brevior larvae within snail tissues was investigated by histology . L3 were detected in the snail mucus ( i . e . , 37 A . abstrusus and 19 T . brevior ) and in the sediment of submerged specimens ( 172 A . abstrusus and 39 T . brevior ) . Following the artificial digestion of H . aspersa snails , a mean number of 127 . 8 A . abstrusus and 60 . 3 T . brevior larvae were recovered . The number of snail sections positive for A . abstrusus was higher than those for T . brevior . Results of this study indicate that A . abstrusus and T . brevior infective L3 are shed in the mucus of H . aspersa or in water where infected gastropods had died submerged . Both elimination pathways may represent alternative route ( s ) of environmental contamination and source of the infection for these nematodes under field conditions and may significantly affect the epidemiology of feline lungworms . Considering that snails may act as intermediate hosts for other metastrongyloid species , the environmental contamination by mucus-released larvae is discussed in a broader context .
Gastropod-borne agents infect approximately 300 million people worldwide , causing major debilitating conditions , as well as adversely affecting quality of life and healthcare . For example , some ailments may contribute to the development of neoplastic tumours , such as cholangiocarcinoma during opisthorchiosis/clonorchiosis [1] or may induce debilitating diseases ( e . g . , schistosomiosis or fascioliosis ) [2 , 3] . In addition , human beings may act as dead-end hosts for zoonotic metastrongyloids of rats , such as Angiostrongylus cantonensis or Angiostrongylus costaricensis ( Strongylida , Angiostrongylidae ) , which cause eosinophilic meningitis and ileitis , respectively [4 , 5] . The main pathway for human infection is represented by the consumption of undercooked intermediate hosts ( i . e . , snails ) or raw paratenic hosts ( e . g . , prawns and freshwater shrimps ) [6 , 7] . In veterinary medicine , the lungworm Aelurostrongylus abstrusus ( Strongylida , Angiostrongylidae ) , along with the less-known Troglostrongylus brevior ( Strongylida , Crenosomatidae ) are of increasing concern due to their spreading in cat populations [8–10] . Indeed , these parasites share a similar biology and ecological niches , and may infect the same individual [10 , 11] . Although gastropod molluscs are recognized as the intermediate hosts for feline lungworms [12] , the role of snails in the transmission of the infection is debatable because they are not preferred preys for felids [13 , 14] . As a result , paratenic hosts ( i . e . , rodents , birds , amphibians and reptiles ) may play a crucial role in the epidemiology of A . abstrusus [15 , 16] . Metastrongyloid nematodes may , however , use additional routes of the infection , as it has been hypothesized for the vertical transmission of T . brevior in cats [17] . Land snails excrete a thin layer of pedal mucus , consisting of water and mucin-like carbohydrate-protein complexes , which acts as glue and lubricant during the locomotion [18] . Again , the production of the mucous trail facilitates several functions of the snail , such as the homing behaviour , aggregation , and protection against desiccation and predation [18] . Interestingly , gastropods may be able to release mucous secretions of variable quality , depending on the nature of the external stimuli ( i . e . , mechanical or chemical ) [19] . The ingestion of fresh vegetables , contaminated by water and/or snail mucus harbouring A . cantonensis third-stage larvae ( L3 ) , has been previously suggested in a recent outbreak of human eosinophilic meningitis in Jamaica [20] . Accordingly , the presence of A . cantonensis and A . costaricensis in the mucus of slugs or snails and in contaminated water has been investigated , but their role in the epidemiology of the infection is yet to be confirmed [7] . This study sought to establish whether A . abstrusus and T . brevior L3 could be found in the mucus of infected Helix aspersa snails and in the water where specimens had died . Larval localization in snails and inflammatory reactions in the intermediate hosts has also been investigated by histology .
Adult H . aspersa ( n = 200 ) were sourced from a commercial farming centre in Barletta ( southern Italy ) . The absence of natural metastrongyloid infections was confirmed by artificially digesting and examining a representative number ( i . e . , 10% ) of snails . During the maintenance , gastropods were placed in plastic boxes filled with natural soil , covered with a net and kept at 23±1°C . Snails were fed every two days with lettuce ad libitum . First-stage larvae of A . abstrusus and T . brevior used for infection were obtained from cats artificially infected with pure isolates , originating from Hungary and Italy , respectively . Two groups of 50 snails each were prepared , being infected with ~500 L1 of A . abstrusus and T . brevior , respectively , as described in [21] . Snails were then placed into two different boxes , according to lungworm species , and two specimens from each group were artificially digested at +6 , +12 and +18 days post infection ( dpi ) to evaluate the success of the infection . In order to assess whether larvae are released in the mucus of snails , six groups ( G1–G6 ) of 4 snails each were formed 20 dpi for each metastrongyloid species . Snails were placed into a 1 l plastic box for bacteriological use , together with 25 ml of tap water , and subjected to different stimuli . Specimens in G1 and G4 were left with water , those in G2 and G5 , and G3 and G6 were fed with lettuce ( 3 g ) or commercial cat food ( 0 . 8 g ) , respectively . Group 1 , G2 and G3 were placed on a thermostat shaker , applying approximately 50 oscillations per minute overnight , while G4 , G5 and G6 were constantly kept without mechanical stimuli . All snails were housed under controlled room temperature ( 23±1°C ) . Once a day in the morning for 20 days , snails were taken out from their plastic boxes , which were rinsed with 20 ml of tap water ( 30°C ) , and the solution was filtered through a 180-μm meshed sieve . The filtrate was centrifuged at 600 g for 5 min and the pellet examined under a light microscope . Finally , snails were returned to the same experimental group boxes . Larvae recovered were quantified and identified using appropriate morphological keys [15 , 21] . In order to evaluate the emergence of larvae from dead submerged snails , at the end of the experiment described above , gastropods were identified with a letter , placed into a 50 ml tube , submerged in tap water and left at room temperature ( 23±1°C ) . Every 24 h for three consecutive days , the sediment was collected and centrifuged at 600 g for 5 min , and the pellet was examined under a light microscope . Nematode larvae were considered motile if they were not damaged and were moving within 10 s [22] . Each snail specimen was artificially digested in a solution of 1% HCl ( 100 ml ) and 1 . 2 g powdered pepsin ( Sigma-Aldrich , St . Louis , Missouri , United States ) , as described elsewhere [21] . The solution was collected in plastic tubes and centrifuged at 600 g for 5 min . The whole suspension ( 5 ml ) was examined under a light microscope; larvae were morphologically identified according to species and developmental stage and counted . Ten representative larval specimens of each of A . abstrusus and T . brevior were isolated from the sediment of digested snail tissues and stored in plastic vials , containing phosphate buffer saline ( PBS ) at −20°C to be analysed by PCR . DNA was extracted using the DNeasy Blood & Tissue Kit ( Qiagen , GmbH , Hilden , Germany ) , in accordance with the manufacturer instructions , and a duplex-PCR was performed , using species-specific forward primer sets , amplifying ITS-2 region of different size ( i . e . , A . abstrusus: 220 bp; T . brevior: 370 bp ) [11] . Sequences were determined from both strands and compared with those available in the GenBank database by Basic Local Alignment Search Tool ( BLAST ) . The arithmetic mean of L3 counted in the mucus , in water and in sediment following artificial digestion was calculated for each group . Larval counts obtained from these three experiments were totalled for each snail and the infection rate was calculated using the following formula: infection percentage = ( 500—T ) /500 , where T was the total number of L3 recovered . Accordingly , the values were compared between A . abstrusus and T . brevior infected snails using the Mann-Whitney U-Test . The statistical analysis was two-sided , at the significance level p = 0 . 05 . To assess the anatomical localization of metastrongyloid larvae , two infected snails were examined every three days , from +1 to +30 dpi ( i . e . , 11 time points ) . Snails were anesthetized with menthol steam in a plastic box for 3–5 hours and , as soon as the foot was completely extended and insensitive to touch , they were deprived of their shells and fixed in a 50 ml vial with 10% neutral buffered formalin solution , to be histologically examined . Longitudinal sections across the middle of the body and parasagittal sections through the coiled part of the snail were routinely processed , embedded in paraffin and 5 μm slices were stained with haematoxylin and eosin ( H&E ) . The presence of a tissue inflammatory response around larvae was also recorded .
No larval nematodes were recovered from H . aspersa specimens digested before the infection . Conversely , larval nematodes were recovered from all snails experimentally infected with either A . abstrusus or T . brevior ( n = 6 for each parasite ) at each sampling point . Third-stage larvae of both feline lungworm species were detected in the mucus of snails in all groups , except G1; only a single L2 of T . brevior was recovered from G4 at 20 dpi ( Table 1 ) . In particular , out of 37 A . abstrusus ( mean 6 . 2±12 . 4 ) and 19 T . brevior ( mean 3 . 2±7 . 3 ) L3 recovered , the highest number of larvae was observed in G2 and G5 for both lungworm species . Overall , 7 . 8% and 1 . 8% of A . abstrusus and T . brevior L3 , respectively , were recovered in the mucus without significant difference recorded in the total number of L3 between A . abstrusus and T . brevior groups ( p>0 . 05 ) . In snails that had died through submersion , larvae of A . abstrusus ( n = 172; mean 7 . 2±12 . 2 ) and T . brevior ( n = 39; mean 1 . 6±2 . 6 ) were detected in the sediment of 21/24 ( 87 . 5% ) and 10/24 ( 41 . 7% ) specimens , respectively ( Tables 1 and 2 ) ; the mean number of A . abstrusus L3 was significantly higher ( p = 0 . 00076 ) . In total , 5 . 7% A . abstrusus and 2 . 7% T . brevior L3 were found in the water ( Table 1 ) . Following the artificial digestion , all H . aspersa snails scored positive for metastrongyloid larvae . A total of 2858 A . abstrusus and 1390 T . brevior L3 were recovered and individual larval counts are reported in Table 2 . The mean number of L3 was 127 . 8 ( min 14–max 386 ) for A . abstrusus and 60 . 3 ( min 10–max 170 L3 ) for T . brevior . Taking into account the original 500 L1 administered to each snail and the total number of larvae recovered , the proportion of L1 that had moulted to L3 was 25 . 6% for A . abstrusus and 12 . 1% for T . brevior ( see statistical analysis paragraph ) . All nematodes detected were actively motile , and all specimens were morphologically identified as infective L3 , since they had lost their outer sheaths and measured 442 . 7±17 . 8 μm ( i . e . , T . brevior and Fig 1A ) and 548 . 6±30 . 3 μm ( i . e . , A . abstrusus and Fig 1B ) . This identification was confirmed by molecular amplification and sequencing of partial ITS-2 gene . Nucleotide sequences , examined by BLAST , displayed a 100% homology with those of A . abstrusus and T . brevior in GenBank ( accession numbers KF751656 and KF751655 ) . During the histological examination , larvae of metastrongyloids were found at different time points , starting from 1 dpi and 9 dpi for A . abstrusus and T . brevior , respectively . Larval transverse sections were recognized as belonging to A . abstrusus or T . brevior according to their mean diameter , detected on round bodies , measuring ~25 and 20 μm , respectively . The number of snail sections positive for A . abstrusus ( n = 103 ) was higher than that for T . brevior ( n = 16 ) , with a mean number of 6 . 4 for the former and 1 . 0 for the latter species . For both metastrongyloids , larvae were mainly observed in the anterior and posterior parts of fibro-muscular tissue ( Fig 2A ) of the foot and in the skirt , close to the pedal and oesophageal glands . Larvae were randomly detected in other organs , such as the kidney parenchyma , the wall of the pallial cavity and the connective sub-epithelial layer of the intestine ( Fig 2B ) . Nonetheless , free larvae were rarely found in infected snails at different time points ( i . e . , at 1 , 12 and 21 dpi for A . abstrusus and at 9 dpi for T . brevior ) . Specimens were localized in the fibro-muscular tissue of the foot , with some of them near vessels , and one specimen in the coelom . All individuals were separated from the surrounding tissue by a thin optically empty space . The tissue response to nematode larvae included: i ) cell-poor ( 3 dpi; Fig 3A ) and cell-rich ( 9 dpi; Fig 3B ) granuloma-like formations , composed of non-vacuolated or vacuolated epithelioid amebocytes; ii ) small necrotic granulomas ( 15 dpi; Fig 3C ) ; iii ) fibroblast-like encapsulations ( 27 dpi; Fig 3D ) . The severity of the inflammatory pattern ranged from mild reactions , featured by vessel dilatation , mild increase of the cellularity and small granulomas ( Fig 4A ) , to strong focal reactivity . In the latter case , large necrotic granulomas were characterized by nodular aggregates of amebocytes in the periphery and their debrided remnants in the centre ( Fig 4B ) . This last response was mainly observed in T . brevior infected snails , which were often featured by enlarged ventral surfaces of the foot and prominent vessel dilatation , along with few larval granulomas ( Fig 4C ) . The occurrence of large amount of amebocytes was also seen in the kidney of T . brevior snails at 3 dpi ( Fig 4D ) .
Further large-scale epidemiological studies that will help identify risk factors associated with the spread of A . abstrusus and T . brevior are needed . Based on data herein reported , owners should pay attention to their cats living outdoor and clean the water and food bowls if left outside , considering that snails and slugs may shed L3 . In addition , since H . aspersa snails may act as intermediate hosts for other metastrongyloid nematodes , such as Oslerus rostratus ( Strongylida , Filaridae ) [38] or A . vasorum [12] , the potential shedding of both lungworms within gastropod mucus should be investigated . Therefore , the identification of alternative pathways for parasite transmission is important particularly since the geographic range of gastropod-borne diseases is expanding due to inter alia introduction of allochthonous infected snails [39 , 40] . | Gastropod-borne parasites may cause debilitating clinical conditions in animals and humans . The infection occurs by consumption of intermediate hosts ( i . e . , snails ) . However , the ingestion of fresh vegetables contaminated by water and/or snail mucus has been proposed as a transmission source for some zoonotic metastrongyloids ( e . g . , Angiostrongylus cantonensis ) . Aelurostrongylus abstrusus and Troglostrongylus brevior cause severe clinical conditions in cats and their prevalence is increasing in feline populations . The epidemiology of feline lungworm infections is poorly understood and the role of gastropod molluscs , as vectors of the disease , is questioned . Since metastrongyloid nematodes may use other little-studied infection pathways , the presence of A . abstrusus and T . brevior larvae was herein assessed in the mucus of Helix aspersa snails and in the water where specimens died . The finding of A . abstrusus and T . brevior infective L3 in the mucus shed by H . aspersa or when infected gastropods die submerged in water indicates that alternative pathways for the transmission of feline lungworms may occur . Accordingly , these results should spur the interest of the scientific community towards the delineation of further infection routes for other metastrongyloids of major human concern , such as Angiostrongylus costaricensis and A . cantonenis . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | Release of Lungworm Larvae from Snails in the Environment: Potential for Alternative Transmission Pathways |
Trypanosomatids are unicellular protists that include the human pathogens Leishmania spp . ( leishmaniasis ) , Trypanosoma brucei ( sleeping sickness ) , and Trypanosoma cruzi ( Chagas disease ) . Analysis of their recently completed genomes confirmed the presence of non–long-terminal repeat retrotransposons , also called retroposons . Using the 79-bp signature sequence common to all trypanosomatid retroposons as bait , we identified in the Leishmania major genome two new large families of small elements—LmSIDER1 ( 785 copies ) and LmSIDER2 ( 1 , 073 copies ) —that fulfill all the characteristics of extinct trypanosomatid retroposons . LmSIDERs are ∼70 times more abundant in L . major compared to T . brucei and are found almost exclusively within the 3′-untranslated regions ( 3′UTRs ) of L . major mRNAs . We provide experimental evidence that LmSIDER2 act as mRNA instability elements and that LmSIDER2-containing mRNAs are generally expressed at lower levels compared to the non-LmSIDER2 mRNAs . The considerable expansion of LmSIDERs within 3′UTRs in an organism lacking transcriptional control and their role in regulating mRNA stability indicate that Leishmania have probably recycled these short retroposons to globally modulate the expression of a number of genes . To our knowledge , this is the first example in eukaryotes of the domestication and expansion of a family of mobile elements that have evolved to fulfill a critical cellular function .
Trypanosomatids are members of the kinetoplastid family of unicellular protists , which includes human pathogens responsible for Chagas disease ( Trypanosoma cruzi ) , African sleeping sickness ( Trypanosoma brucei ) , and leishmaniasis ( Leishmania spp . ) . T . brucei and T . cruzi belong to the monophyletic Trypanosoma group , which is distantly related to all the other trypanosomatids , including Leishmania spp . [1] . Kinetoplastid protein-coding genes are often organized as large directional gene clusters ( DGCs ) that form polycistronic units [2–4] . Individual mRNAs with a 39-nt 5′ capped spliced leader sequence and 3′ poly ( A ) tail are generated from the polycistronic pre-mRNAs via 5′ trans-splicing and 3′ cleavage-polyadenylation reactions [5] . Several lines of evidence raise the intriguing possibility that in trypanosomatids poly ( A ) addition is coupled to trans-splicing of the downstream gene [6 , 7] . trans-splicing signals are often U-rich polypyrimidine tracts , which precede AG acceptor sites on average 50–100 nt upstream of the translational start site . There is no consensus polyadenylation signal in trypanosomatid mRNA , and evidence obtained from a small number of loci suggests that polyadenylation occurs within a short region 100–400 nt upstream of the next polypyrimidine trans-splicing signal [7 , 8] . It was recently reported that in 89% of all available cDNA sequences from T . brucei , polyadenylation usually occurs at an A residue located between 80 and 300 nt from a downstream polypyrimidine tract [9] . The aforementioned polycistronic transcription , and the absence of pol II promoters in all known protein-coding genes , necessitate that gene expression be controlled post-transcriptionally . Indeed , numerous examples in kinetoplastids , including Leishmania , show that sequences predominantly located in the 3′-untranslated regions ( 3′UTRs ) control mRNA stability and translation [10–18] . Transposable elements ( TEs ) are DNA sequences capable of moving from one chromosomal region to another . They are classified into two major groups based on the mechanisms used for their transposition . Class I TEs , or retroelements , transpose via reverse transcription of an RNA intermediate and are further divided into the long-terminal repeat ( LTR ) retrotransposons with LTRs and the non-LTR retrotransposons , also called retroposons . Class II TEs , or DNA transposons , move strictly through a DNA intermediate . A considerable fraction of higher eukaryote genomes comprises TEs , as exemplified in human ( over 40% of the genome ) [19] and maize ( over 50% of the genome ) [20] . There is now a growing body of evidence to suggest that TEs can be functionally important and not just “junk , ” “selfish , ” or “parasitic” DNA sequences that make as many copies of themselves as possible [21–23] . For example , there is a considerable number of domesticated TE copies that act as transcriptional regulatory elements or contribute to protein-coding regions of cellular genes ( for review see [24–26] ) . The recent completion of the Tritryp genome projects confirmed the presence of LTR retrotransposons and non-LTR retrotransposons ( transposons ) but no DNA transposons [2–4] . Retroposons constitute the most abundant TEs described in the genome of T . cruzi and T . brucei ( ∼3% of nuclear genome ) , while no potentially active TEs have been characterized to date in L . major [3] . The most abundant retroposons , ingi and ribosomal mobile element ( RIME ) in T . brucei [27–29] and L1Tc and NARTc in T . cruzi [30 , 31] , are distributed across their respective genomes , although they do show a relative site specificity for insertion [32 , 33] . The T . brucei RIME ( 0 . 5 kb ) appears as a truncated version of the T . brucei ingi ( 5 . 25 kb ) , in which the central 4 . 7 kb fragment has been deleted ( Figure 1 ) . Similarly , the T . cruzi NARTc ( 0 . 25 kb ) element was derived from L1Tc ( 4 . 9 kb ) by a 3′ deletion [30] . The potentially functional ingi and L1Tc each encode a large single multifunctional protein that is probably responsible for their retrotransposition and that of the short non-autonomous RIME and NARTc , respectively [32 , 33] . Consequently , ingi/RIME and L1Tc/NARTc are considered as pairs of retroposons , as previously described for the human long interspersed element 1 ( LINE1 ) /Alu , the eel UnaL2/UnaSINE1 , and the plant LINE/S1 pairs [34–37] . Until now , potentially active or short non-autonomous TEs have not been detected in the L . major genome [3 , 4] . However , the genome does contain degenerated retroelements ( L . major degenerated ingi/L1Tc-related elements [LmDIREs] ) corresponding to remnants of extinct ingi/L1Tc-like retroposons [38] . Interestingly , the ingi/RIME and L1Tc/NARTc pairs and DIREs share the first 78–79 nucleotides even though they are otherwise unrelated to each other [30 , 38] ( Figure 1 ) . This “79 bp signature , ” therefore , constitutes the hallmark of trypanosomatid retroposons . Using the “79 bp signature” for BLASTN searches , we identified in the L . major genome 1 , 858 short ( ∼550 bp ) , noncoding and degenerated retroposons that belong to two new large families of relatively conserved repetitive DNA elements ( L . major short interspersed degenerated retroposon 1 [LmSIDER1] and LmSIDER2 ) , which display all the hallmarks of trypanosomatid retroposons . LmSIDER1 and LmSIDER2 are predominantly located in the 3′UTR of L . major mRNAs and represent the most abundant TEs now characterized in trypanosomatid genomes . Considering that regulation of gene expression in Leishmania is mediated almost exclusively by sequences within 3′UTRs , we hypothesized that LmSIDERs may play a role in the regulation of gene expression . In the present study , we provide experimental evidence that members of the second retroposon subfamily in L . major , LmSIDER2 , promote mRNA destabilization . We conclude that Leishmania spp . , but not trypanosomes , have recycled and probably expanded an extinct family of short retroposons that participate in the maintenance of an essential cellular function , i . e . , the regulation of gene expression .
All ingi/RIME , L1Tc/NARTc , and DIRE present in the T . brucei , T . cruzi , and L . major genomes have been identified and annotated [3] . These different retroposon families contain at their 5′-extremity a 79-bp conserved motif ( called “79 bp signature” ) , which constitutes the hallmark of trypanosomatid retroposons [38] . In order to identify other repeated sequences containing the “79 bp signature , ” we surveyed the L . major and T . brucei genomes for the presence of the first 79 bp of ingi and 78 bp of L1Tc . BLASTN searches initially detected 108 significant matches in the L . major genome , in addition to identifying the LmDIRE sequences . Comparison of the sequences located downstream of these 108 “79 bp signature” matches revealed two heterogeneous groups of sequences , which we named LmSIDER1 and LmSIDER2 . After several rounds of BLASTN searches with complete LmSIDER1 and LmSIDER2 sequences , we identified 1 , 858 related sequences ( 785 LmSIDER1 and 1 , 073 LmSIDER2 ) in the L . major genome ( Figure 1 ) . Coordinates for these elements on each L . major chromosome are listed in Table S1 . A phylogenetic analysis of 789 LmSIDER sequences confirmed their division into two distinct subfamilies , LmSIDER1 and LmSIDER2 ( Figure 2 ) . A similar BLASTN analysis of the T . brucei genome revealed 22 sequences forming two groups of relatively conserved sequences ranging from 558 to 587 bp , named T . brucei short interspersed degenerated retroposon 1 ( TbSIDER1 ) ( ten sequences ) and TbSIDER2 ( 12 sequences ) ( Figure 1 ) . The work reported will hereafter primarily focus on the LmSIDER2 family . One thousand thirteen LmSIDER2s were aligned with the inclusion of numerous gaps to maximize the alignments ( Figure S1 ) . The aligned LmSIDER2 sequences ranged between 178 bp and 702 bp , with a mean of 545 bp . Although LmSIDER2 sequences are highly heterogeneous in composition and size ( the alignment comprises 1 , 612 positions ) , a conserved core sequence was identified ( 538 bp ) by removing insertions ( 1 , 074 positions ) ( Figure S2 ) . The removed positions ( 66 . 6% of the positions in the original alignment ) account for 14 . 5% of the aligned nucleotides . The defined core sequence was used to perform all the subsequent bioinformatics analyses . To determine whether the LmSIDER2 sequences are significantly conserved , we performed a chi-square ( χ2 ) test on the LmSIDER2 core and the flanking sequences ( 200 bp upstream and 160 bp downstream ) ( Figure 3 ) . All positions of the LmSIDER2 core show a χ2 score far above the threshold line corresponding to significant levels ( using three degrees of freedom , a χ2 value of 16 . 3 corresponds to a significance level of p < 0 . 001 ) , indicating that the LmSIDER2 core is conserved . The flanking regions are not conserved , except for a thymidine-rich stretch ( 18 residues ) starting at 15 bp upstream from the LmSIDER2 ( unpublished data ) . Several lines of evidence demonstrate that members of LmSIDER2 are clearly related to retroposons identified in trypanosomes ( ingi/RIME , L1Tc/NARTc , and DIRE ) . ( i ) Two tandemly arranged “79 bp signatures” are found at the 5′-extremity of the LmSIDER2 core ( Figures 3 and 4 ) . These are 68% and 62% identical with the first 79 bp residues of the T . brucei ingi/RIME . ( ii ) The 3′-extremity of the LmSIDER2 core sequence is composed of an adenosine-rich stretch , which is a hallmark of retroelements due to the requirement of an RNA intermediate during retrotransposition [39] ( Figures 3 and 5 ) . ( iii ) The LmSIDER2 sequences show a high GC content ( 65 . 3% ) , similar to the one seen in LmDIRE ( 64 . 5% ) , as compared to the rest of the L . major genome ( 59 . 7% ) ( Table 1 ) . The GC content is also higher for the T . brucei RIMEs ( 53 . 8% ) , ingis ( 52 . 3% ) , and TbDIREs ( 48 . 7% ) , as compared to the rest of the T . brucei genome ( 41% ) . The relative lower GC content of the TbDIREs compared to the ingi/RIME sequences is probably due to the accumulation of point mutations in TbDIREs , as previously observed for extinct retroposons [24] . This interpretation may also explain the relative lower GC content bias observed in the degenerated LmSIDER2 and LmDIRE sequences , compared to the potentially active ingi and RIME elements . ( iv ) As previously observed for the T . brucei ingi/RIME and T . cruzi L1Tc/NARTc retroposons , an 18-bp thymidine-rich motif is conserved upstream of LmSIDER2 ( Figures 3 and 5 ) . According to the current model of retrotransposition , this sequence motif corresponds probably to the recognition site of the endonuclease encoded by ingi/L1Tc-related elements [32 , 33] . ( v ) During retrotransposition , the retroposon-encoded endonuclease performs two assymetrical single-strand cleavages , leading to a duplication of the residues between both cleavages . The duplicated motif , flanking the newly inserted retroposons , is called target site duplication ( TSD ) ( Figure 5 ) . One hundred ninety-one LmSIDER2 sequences ( 18 . 9% of the aligned LmSIDER2 ) are flanked by a conserved motif ( >75% identity ) ranging from 11 bp to 19 bp ( 69 of them being 13 bp long ) , which resemble vestiges of TSDs . For three of them , the 11–13-bp TSD is conserved without mismatch ( Figure 5A ) . Interestingly , the size of TSD flanking LmSIDER2 and the ingi/RIME/L1Tc/NARTc elements is similar ( ∼13 bp versus 12 bp ) [32 , 33] . ( vi ) 90% of the identified LmDIRE sequences ( 47 out of 52 ) overlap with a LmSIDER2 sequence at their 5′- and/or 3′-extremities ( unpublished data ) , suggesting that LmDIRE ( previously characterized as retroelement vestiges related to ingi and L1Tc [38] ) and LmSIDER are related . This last observation suggests that LmSIDER was derived from LmDIRE by deletion , as observed for the T . brucei ( ingi/RIME ) and T . cruzi ( L1Tc/NARTc ) autonomous/non-autonomous pairs of retroposons [30] ( Figure 1 ) . Similarly , both TbSIDER groups show hallmarks of trypanosomatid retroposons , including the presence of the “79 bp signature” ( Figure 4 ) and an adenosine-rich stretch ( Figure 5B and 5C ) at their 5′- and 3′-extremity , respectively . In addition , one member each of the TbSIDER1 and TbSIDER2 groups is flanked by a degenerated TSD sequence ( Figure 5B and 5C ) . The difficulties encountered in performing the LmSIDER2 alignment reflect the high level of divergence of this TE family . To study the extent of this divergence and gain better insight into the evolutionary dynamics of the LmSIDER2 family , we calculated the percentage of divergence between the consensus LmSIDER2 core sequence deduced from the alignment and each LmSIDER2 core sequence . Since the consensus sequence is assumed to approximate the element's original sequence at the time of insertion , the percentage of substitution from the consensus sequence is correlated to the age of a given element ( the age corresponds to the time of retrotransposition ) . The divergence ranged between 12% and 40% , with median and mean values of 20% and 17% , respectively ( Figure 6 ) . The high level of divergence between the consensus and the most conserved LmSIDER2 sequence ( 12% ) implies that LmSIDER became extinct a long time ago . The same analysis was carried out on the T . brucei RIME/TbSIDER and T . cruzi NARTc elements , which are the only short retroposons characterized so far in the trypanosome genomes [40] . For TbSIDERs , the percentage of divergence from the consensus TbSIDER1 and TbSIDER2 core sequences ranged between 11 . 6% and 18% and 8% and 13 . 7% , with median values of 16% and 11% , respectively ( Figure 6 ) . This indicates that TbSIDERs are also extinct TEs , as observed for LmSIDERs . In contrast , RIME and NARTc are far more conserved compared to SIDERs ( median divergence value of 4% and 2% , respectively ) ( Figure 6 ) . In addition , 13 . 8% and 22 . 5% of the analyzed RIME and NARTc sequences are over 99% identical with the consensus sequence , respectively , indicating recent retrotransposition activities in the trypanosome genomes . The T . brucei and L . major genomes are highly syntenic , with approximately 70% of all genes remaining in the same genomic context [40] . This large-scale synteny enables a comparative analysis of TE distribution in these two completed trypanosomatid genomes . The trypanosomatid genomes are characterized by their unique arrangement of DGCs , which are separated by short ( 0 . 9–14 kb ) divergent or convergent strand-switch regions . For example , the L . major genome ( 32 . 6 Mb ) has 36 pairs of chromosomes ( 0 . 25–2 . 7 Mb ) that are organized into 133 DGCs of tens to hundreds of protein-coding genes ( up to 1 . 26 Mb per DGC ) [4] . The T . brucei genome is more compact ( 26 Mb ) , with 11 pairs of megachromosomes ( 1 . 1–5 . 5 Mb ) containing subtelomeric genes at both extremities , which account for ∼20% of the genome ( ∼5 . 2 Mb ) [2] , while L . major chromosomes do not contain large subtelomeric regions [4] . Interestingly , retroposons do not show the same distribution in the L . major and T . brucei genomes ( Tables 2 and 3 ) . Indeed , almost all of LmSIDERs and LmDIREs in L . major are located in DGCs ( 95 . 4% of the TE ) , while the ingi , RIME , TbDIRE , and TbSIDER retroposons in T . brucei are primarily located in subtelomeric regions ( 60 . 1% of the TE ) . Tables 2 and 3 also show that strand-switch regions display the highest TE richness in both T . brucei and L . major , i . e . , over 110 TE per Mb , which corresponds to 23 . 4% ( 54 TE ) and 4 . 6% ( 88 TE ) of the retroposons , respectively . The most striking observation is that retroposons are ∼50 times more abundant in L . major DGCs compared to T . brucei DGCs ( 1 , 821 versus 38 ) , despite the high level of synteny observed between these regions , which contain an equivalent number of protein-coding genes [40] . This extraordinary difference is the consequence of the unusual distribution and high copy number of LmSIDERs , as exemplified by the comparative analysis of T . brucei Chromosome 6 and L . major Chromosome 30 , which are almost completely syntenic ( Figure 7 ) ( see Figures S3 and S4 for the other chromosomes ) . Since most LmSIDERs are present in the intergenic regions of DGCs , it was important to determine where they are located in regards to the pre-mRNA processing sites . Individual mature mRNAs in trypanosomatids are generated from polycistronic precursors by 5′ trans-splicing of a 39-nt capped leader RNA and 3′ polyadenylation [41] . To determine the putative position of polyadenylation sites in L . major , we used the prediction algorithm previously developed for trypanosome mRNA processing sites [9] . There are 8 , 162 genes annotated in version 4 . 0 of the L . major genome . The algorithm could predict the vast majority of the 5′UTRs and 3′UTRs of those genes with the exception of 121 5′UTRs ( 1 . 5% ) and 569 3′UTRs ( 7% ) . Of the 1 , 858 LmSIDERs characterized in the L . major genome , 1 , 356 were found to overlap with a 3′UTR , and 494 have at least one 3′UTR upstream , including 85 LmSIDERs found in strand-switch regions . Conversely , 1 , 852 have at least one 5′UTR downstream , including 50 LmSIDERs overlapping with the 5′UTR of a gene . Because 73% of the LmSIDERs are found within 3′UTRs , we calculated the median distance of these elements to the upstream stop codon ( 680 bp ) and the downstream ATG ( 978 bp ) , as well as the distances from the polypyrimidine tract ( 833 bp ) and putative polyadenylation site ( 734 bp ) ( Figure 8 ) . The average location of LmSIDER2s is in the middle of the in silico–predicted 3′UTRs ( at almost equal distance from the upstream stop codon and the downstream polyadenylation site ) , which clearly demonstrates that most LmSIDER2s are located in the 3′UTR of mRNAs . 3′UTRs are known to play a key role in regulating gene expression in Leishmania [13 , 15 , 18 , 42–46] . The widespread distribution of LmSIDER elements within the Leishmania genome and their predominant localization in 3′UTRs , therefore , support the hypothesis that LmSIDER2 may contribute to the regulation of gene expression in this organism . To test this hypothesis , we used custom-designed low density DNA oligonucleotide microarrays to determine expression profiles of LmSIDER2-containing mRNAs in L . major promastigotes and L . major lesion amastigotes isolated from BALB/c mice . Oligonucleotide microarrays were designed to represent 154 L . major genes , from which only 38 bear LmSIDER2 in their 3′UTR . Four independent hybridization experiments were scanned and analyzed using recommended statistic parameters for low spot density arrays in the GeneSpring software . The overall pattern of gene expression for L . major promastigotes and amastigotes is shown in the scatterplot of normalized data in Figure 9A . Approximately 50% of the LmSIDER2-containing transcripts are developmentally regulated in either L . major promastigotes or amastigotes , without any bias towards a particular life stage ( 24% amastigotes versus 26% promastigotes ) and with the majority of genes being constitutively expressed ( Figure 9A; Table S2 ) . Interestingly , from these LmSIDER2-containing transcripts , more than 75% have signal intensities that are lower than the mean intensity of all the spots , as compared to 40% for the non-LmSIDER2 transcripts ( Figure 9A ) . The minority of LmSIDER2-containing more abundant transcripts ( ∼25% ) may be explained by a higher degeneracy of LmSIDER2 that results in a nonfunctional element or by the presence of additional elements within the 3′UTR . To gain independent evidence for the relatively lower expression of LmSIDER2 mRNAs , a randomly selected number of L . major transcripts containing or lacking LmSIDER2 that are most likely clustered within the same transcription unit on three distinct chromosomes were analyzed by quantitative northern blotting . LmjF13 . 0440 , LmjF24 . 1260 , LmjF24 . 1360 , and LmjF36 . 3810 transcripts harbor LmSIDER2 in their 3′UTR , whereas LmjF13 . 0430 , LmjF24 . 1250 , LmjF24 . 1280 , and LmjF36 . 3910 do not . LmjF13 . 0430/LmjF13 . 0440 and LmjF24 . 1250/LmjF24 . 1260 are tandemly linked , whereas LmjF24 . 1280/LmjF24 . 1360 and LmjF36 . 3810/LmjF36 . 3910 are part of the same transcription unit but are separated by seven to eight genes ( Figure 9B ) . Figure 9B demonstrates that LmSIDER2-containing mRNAs are systematically expressed at much lower levels compared to their co-transcribed genes lacking LmSIDER2 . Taken together , these results argue for a more general role of LmSIDER2 in downregulating mRNA expression . We have recently identified conserved regulatory elements within the 3′UTR of a large set of developmentally regulated transcripts in Leishmania and showed that these elements operate principally at the translational level [16 , 17] . While characterizing the LmSIDER families , we found that these regulatory elements are part of the LmSIDER1 subfamily . We next wanted to obtain direct evidence for the role of LmSIDER2 elements in the regulation of gene expression using luciferase ( LUC ) as a reporter mRNA . For this , two members of the LmSIDER2 subfamily were selected for further analysis . LmjF08 . 1270 encodes a hypothetical protein of unknown function [47] and LmjF36 . 3810 encodes an aminomethyltransferase . Both harbor LmSIDER2 in their 3′UTR . The LmSIDER2 in the LmjF08 . 1270 transcript ( LmSIDER2–1270 ) is 563 nt long and is located at the end of a 1 , 531-nt-long 3′UTR ( 53 nt upstream from the mapped polyadenylation site , unpublished data ) . In the case of LmjF36 . 3810 , LmSIDER2 ( LmSIDER2–3810 ) is 610 nt long and is located within a 1 , 831-nt 3′UTR , at 534 nt from the 3′end of the mRNA ( see Figure 10A ) . The sequence identity between the two LmSIDER2 is 60% . The full-length 3′UTR of either LmjF08 . 1270 or LmjF36 . 3810 mRNAs was cloned downstream of the LUC reporter gene . LUC reporter constructs with the whole 3′UTR lacking LmSIDER2 or the LmSIDER2 alone were also made ( Figure 10A ) . Each construct was transfected into L . major promastigotes , and stable recombinant parasites were analyzed for LUC activity . Relative LUC activity was calculated by comparing the values obtained with either SIDER2-expressing or SIDER2-lacking recombinant parasites to the LUC control [16] . Figure 9B demonstrates that the LmjF36 . 3810 3′UTR ( LUC-3′UTR3810 ) results in a 3 . 1-fold decrease in LUC activity in comparison to the LUC control . A similar decrease ( 2 . 7-fold ) was obtained with the LmjF36 . 3810 LmSIDER2 alone ( LUC-SIDER3810 ) . Contrasting with this , deletion of SIDER3810 in L . major LUC-ΔSIDER3810 promastigotes caused a 3 . 5-fold increase in LUC activity with respect to the LUC-3′UTR3810 and LUC-SIDER3810 recombinant parasites . In the case of LUC-3′UTR1270 and LUC-SIDER1270 promastigote cultures , the presence of LmSIDER2 had only a slight effect on LUC activity; however , the deletion of LmSIDER2 in LUC-ΔSIDER1270 resulted in a 2 . 1-fold increase in LUC activity ( Figure 10B ) , which is consistent with a putative role of LmSIDER2 in regulating LmjF08 . 1270 gene expression . To investigate the basis of the differences observed in LUC activity between LmSIDER2-bearing and LmSIDER2-lacking LUC chimeric constructs , we first tested the effect of LmSIDER2 on LUC mRNA abundance by northern blotting . RNA loading on the gel was monitored by hybridization to the 18S rRNA–specific probe . The LmjF36 . 3810 or LmjF08 . 1270 LmSIDER2 reduces the levels of LUC chimeric mRNAs by an average of 5-fold with respect to the LUC control mRNA levels ( Figure 10C ) . In contrast to this , deletion of LmSIDER2–3810 or LmSIDER2–1270 retroposons causes a marked increase in LUC mRNA accumulation ( 3 . 45- to 3 . 8-fold ) . These findings indicate that LmSIDER2 could downregulate mRNA abundance . To determine the relative contribution of mRNA abundance to the observed LUC activity , we evaluated the level of LUC protein expression derived from the LmSIDER2-containing 3′UTRs by western blotting ( Figure 10D ) . In the case of LUC-3′UTR3810 and LUC-SIDER3810 transfectants , the amount of LUC mRNA dictates the amount of LUC protein . A linear correlation was also observed between LUC-ΔSIDER3810 mRNA accumulation and LUC-ΔSIDER3810 protein levels ( Figure 10C and 10D ) . These findings establish that LmSIDER2–3810 does not alter translational regulation in L . major promastigotes , but rather confers lower mRNA levels . However , although LmSIDER2–1270 clearly contributes to lower steady-state RNA levels , the decrease in mRNA ( 2 . 7-fold to 4 . 54-fold ) does not perfectly correlate with LUC protein levels ( 1 . 6-fold to 1 . 8-fold decrease ) , and LUC activity remained practically unchanged between LUC-3′UTR1270 and LUC-SIDER1270 recombinant parasites in comparison to the LUC control ( Figure 10B–10D ) . These data suggest that in the context of LmjF08 . 1270 , other sequences might compensate for the downregulation effect of LmSIDER2 on mRNA abundance , probably by increasing translation rates . As regulation of gene expression in Leishmania is known not to occur at the transcriptional level , and as there is virtually no evidence for differential splicing [10] , the most likely mechanism for lower abundance of LmSIDER2 mRNAs is through altered mRNA stability . To examine whether lower accumulation of LmSIDER2–3810- and LmSIDER2–1270-containing LUC chimeric transcripts in L . major promastigotes could be due to mRNA destabilization , we measured half-lives of the LUC transcripts that bear or lack LmSIDER2 using actinomycin D treatment to block de novo transcription and northern blot hybridization to visualize mRNAs . Analysis of the data revealed that LUC-3′UTR1270 and LUC-3′UTR3810 transcripts have half-lives of 45 min and 80 min , respectively ( Figure 11A and 11B ) . LmSIDER2 deletion resulted in a marked increase of the half-life of the LUC transcript by 3 . 0- to 5 . 5-fold , respectively ( Figure 11A and 11B ) . We also evaluated the half-lives of the single copy endogenous LmjF36 . 3810 and LmjF08 . 1270 mRNAs , which are very short ( ∼16 and 14 min , respectively ) ( Figure 11C and 11D ) . The differences in the half-lives observed between the endogenous and the episomal LmSIDER2-containing transcripts can be explained by the higher copy number ( ∼35 ) of the latter compared to that of the former .
The genomes of higher eukaryotes contain pairs of autonomous/non-autonomous retroposons composed of small noncoding elements , which use for their own mobility the retrotransposition machinery encoded by autonomous elements ( for review see [48] ) . This is exemplified by the retroposon pairs described in human ( LINE1/Alu , LINE2/MIR , and LINE2/Ther-1 ) [24 , 34 , 37] , fish ( UnaL2/UnaSINE1 ) [36] , reptiles ( CR1-like LINE/SINE ) [49] , and plants ( Bali1/S1 ) [35] . In these examples , the small noncoding elements , called small interspersed elements ( SINEs ) , are tRNA- , 5S RNA– , or 7SL RNA–related sequences [50–52] . In contrast , the small noncoding partners ( RIME and NARTc ) of the trypanosome ingi/RIME ( T . brucei ) and L1Tc/NARTc ( T . cruzi ) pairs are derived from the autonomous retroposons ( ingi and L1Tc ) by deletion of the coding sequence [27–31] . The truncated RIME and NARTc elements became fixed in the trypanosome genome with copy numbers equivalent to that of the autonomous ingi and L1Tc retroposons ( see Table 1 ) [32 , 33] . In addition , all trypanosomatid genomes analyzed so far contain degenerated retroposons related to ingi and L1Tc ( DIRE ) [38] . The majority of LmDIRE sequences identified in the L . major genome ( 90% ) overlaps with a subset of LmSIDER , suggesting that the latter are derived from the former by deletion . This indicates the existence of an LmDIRE/LmSIDER pair comparable to the trypanosome ingi/RIME and L1Tc/NARTc pairs . The trypanosome ingi/RIME and L1Tc/NARTc pairs are considered active , since the very low level of sequence divergence observed is consistent with recent retrotransposition activities . The T . brucei and T . cruzi genomes contain several potentially active ingi/L1Tc , which encode a single long and conserved protein [3 , 32 , 33] . This contrasts with the L . major LmDIRE/LmSIDER pair , which has lost its retrotransposition activity . Indeed , the only reverse transcriptase domains identified in the completed L . major genome belong to LmDIREs , which have accumulated numerous point mutations after their extinction [38] . In the absence of functional retroposons , the noncoding LmSIDER families can be considered extinct as well , since their members need enzymes produced in trans by autonomous retroposons for their mobilization . Consequently , the LmSIDER and LmDIRE families probably became extinct simultaneously , when the last active LmDIRE disappeared from the L . major genome , as proposed for the extinct human LINE2/MIR and rodent LINE1/B1 pairs [24 , 53] . The simultaneous extinction of the human autonomous LINE2 and non-autonomous MIR retroposons is illustrated by their similar nucleotide substitution level [24] . However , this comparative analysis cannot be done for the LmDIRE/LmSIDER pair because of the inappropriately low number of LmDIRE sequences available for such a statistical analysis [38] . The high level of divergence ( 12% ) between the consensus and the most conserved LmSIDER2 sequence suggests that LmSIDER became extinct a long time ago . The rise and fall of TE families has been well documented in several genomes [19 , 24] . For example , it was estimated that the human LINE2 retroposons , which show at least 18% divergence with the consensus LINE2 sequence , became extinct 50–100 million years ago [19] . In the absence of trypanosomatid fossil records and thus of a molecular clock , the date of LmSIDER extinction cannot be estimated with accuracy . It probably occurred after the speciation of the Trypanosoma and Leishmania genus 200–500 million years ago [54] , since trypanosomes still contain putative active elements [38] . The recent completion and comparative analysis of eukaryotic genomes provides evidence that several superfamilies of short non-autonomous retroposons ( e . g . , SINE ) have been conserved and distributed among a wide range of species [55–57] . These conservations suggest that numerous extinct retroposons were domesticated hundreds of million years ago and are still functional in several species . While superfamilies of retroposons are conserved and shown to be functional , exaptation of a TE family to the extent described here has not been reported so far . Here , we provide evidence that Leishmania spp . have recycled a whole family of short retroposons ( LmSIDER2 ) , which have evolved to fulfill important biological pathways such as the regulation of gene expression , whereas its close relative T . brucei developed other approaches to maintain similar cellular functions . Retroposon-mediated regulation at transcriptional or post-transcriptional levels [23 , 48 , 58–61] remains a relatively rare event in other eukaryotes and is not thought to be an intrinsic function of retroposons . Most LmSIDERs ( 95 . 4% ) are located within intergenic regions of DGCs , mainly in 3′UTRs , while 95 . 5% of the TbSIDERs are located outside DGCs; the retroposon density in DGCs being ∼50 times higher in L . major than T . brucei . This contrasting SIDER distribution can also be correlated with the difference in the average size of intergenic regions between L . major and T . brucei ( 1 , 432 bp versus 721 bp ) [40] , in part due to the presence of LmSIDERs in the 3′UTRs . We have previously identified a conserved 450–550-bp element located in the 3′UTR of several Leishmania amastigote–specific transcripts that is implicated in stage-specific translational control [16 , 17] . Interestingly , this element belongs to the LmSIDER1 subfamily of retroposons , which comprises at least 785 sequences across the Leishmania genome ( A . Rochette , M . Smith , P . Padmanbhan , B . Papadopoulou , unpublished data ) . In this study , we presented several lines of evidence showing that LmSIDER2 promotes mRNA destabilization . This conclusion stems from a comprehensive microarray analysis , from northern blotting data , and from a more direct reporter gene analysis of selected mRNAs . The functional distinction between LmSIDER1 and LmSIDER2 is consistent with the way they clustered in a phylogenetic tree . The ability of LmSIDER2 to destabilize mRNA seems to be intrinsic and context independent , since it can be functional at different distances from the poly ( A ) tail and even outside the context of the endogenous 3′UTRs ( see Figure 10 ) . LmSIDER2-containing mRNAs are generally expressed at lower levels compared to non-SIDER2-bearing transcripts and are short-lived ( half-lives of ∼15 min ) . Taken together , these observations suggest that LmSIDER2 are cis-acting components of a regulatory pathway that generally downregulates gene expression to ensure rapid turnover of a specific subset of Leishmania mRNAs . Throughout its complex life cycle , Leishmania is subjected to a variety of rapidly changing environmental conditions , and rapid mRNA turnover can permit the parasite to adapt its pattern of protein synthesis to continuously changing physiological needs . We hypothesize that the mRNA-destabilizing function of LmSIDER2 can be enhanced or blocked as needed due to their particular sequence or structure ( LmSIDER2 elements are highly heterogeneous ) , and/or the presence of other elements in the 3′UTR of Leishmania transcripts . This is in agreement with our preliminary results in L . infantum amastigotes , where the 36 . 3810 SIDER2 becomes inactive due to the presence of a downstream element ( M . Müller , B . Papadopoulou , unpublished data ) , and with the observation that none of the highly expressed housekeeping genes harbor LmSIDER2 ( unpublished data ) . In the case of LmjF36 . 3810 and LmjF08 . 1270 transcripts , which are both constitutively expressed in L . major , the LmSIDER2 destabilizing element works as efficiently in amastigotes as it does in promastigotes ( unpublished data ) . Our microarray data on 38 LmSIDER2-containing transcripts are also consistent with these observations . However , other stages of the parasite , irrespective of whether they are morphologically distinct ( e . g . , metacyclics ) or not , exist where the function of these elements might be more crucial . Indeed , we also found that several transcripts reported to be upregulated in the metacyclic stage of L . major [62] contain LmSIDER2 ( unpublished data ) . Likewise , the role of these elements might be more evident as the parasite experiences a specific environmental challenge , particularly in the rather dynamic ecological niche inside its insect host . Indeed , a number of short-lived mRNAs are known to be responsive to specific extracellular environmental stimuli in other systems where expression is regulated by sequences in 3′UTRs ( e . g . , the AU-rich elements of inflammatory cytokines and growth factors ) [63 , 64] . Alternatively , the role of LmSIDER2 might be to negatively modulate gene expression and thereby check that mRNAs , stage-specific or constitutively expressed , are maintained at nontoxic levels ( for instance , mRNAs encoding structural proteins are generally expected to be more abundant than those encoding regulatory proteins ) . Comparison of the L . major and T . brucei genomes showed that SIDERs are ∼70 times more abundant in L . major compared to T . brucei [38] . Considering that the majority of LmSIDERs is co-transcribed with coding genes and that members of the LmSIDER families are shown to play a role in the regulation of gene expression , whereas most of the very few TbSIDERs are distributed in the relatively silent subtelomeric regions , it is tempting to propose that Leishmania , but not trypanosomes , have exapted and expanded the SIDER retroposons . The reasons behind this extraordinary LmSIDER expansion are currently unknown . The widespread genomic distribution of LmSIDER2 and our functional data on both LmSIDER1 and LmSIDER2 members raises the interesting possibility that numerous Leishmania transcripts encoding a wide repertoire of functionally diverse proteins may be regulated by a similar mechanism in response to specific environmental stimuli and/or growth conditions . The involvement of TE in coordinated expression of genes was already proposed in the seventies [65] . We propose that Leishmania , an organism with no known control at the level of transcription initiation , has acquired the ability to post-transcriptionally coordinate gene regulation via short retroposons ( LmSIDERs ) in the 3′UTR . This is consistent with the prevailing notion that retroelements likely emerged as genomic parasites and gradually invaded the genomes of most eukaryotic cells , but later became an integral part of their genome and were used for the benefit of these organisms .
A BLASTN search of the L . major genome with the first 79 residues of the T . brucei ingi/RIME ( “79 bp signatures” ) revealed 108 homologous sequences , corresponding to the 5′-extremity of degenerated retroposons , subsequently called LmSIDERs . A multiple alignment ( ClustalW [66] ) of the sequences located downstream from these 108 “79 bp signatures” ( 1 kb ) was then done to define six groups of related but very heterogeneous sequences , ranging from 450 to 790 bp in length . The 3′-extremity of most of these relatively conserved sequences was composed of an adenosine-rich stretch , as generally observed for retroposons . In order to identify other LmSIDER in the L . major genome , a second BLASTN search was performed with one representative from each group of sequences . About 1 , 500 matches were retained . A third BLASTN search conducted with a subset of very divergent LmSIDER identified new sequences . Some of these newly identified LmSIDER were used for a fourth BLASTN search . We stopped this reiterative BLASTN search approach after two additional runs , since no more sequences were detected , with a total of 1 , 858 identified LmSIDER elements . The BLASTN analysis also revealed that LmSIDER could be separated into two groups composed of 785 ( LmSIDER1 ) and 1 , 073 ( LmSIDER2 ) sequences ( see Figure 2 ) . The first 79 residues of the T . brucei ingi/RIME ( “79 bp signatures” ) were used to perform a BLASTN search of the T . brucei genome database ( version 3 . 0 of The Institute for Genomic Research's [TIGR] T . brucei assembly ) . For this BLAST analysis , the annotated RIME , ingi , and DIRE sequences were masked using Repeat Masker ( http://www . repeatmasker . org/ ) . A multiple sequence alignment ( ClustalW [66] ) of the regions located downstream of 51 identified “79 bp signatures” ( 1 kb ) and defined two groups of related sequences , named TbSIDER1 ( ten sequences ) and TbSIDER2 ( 12 sequences ) , while the other 29 sequences were unique and appeared not to be related to retroposons . We used ClustalW ( http://www . ebi . ac . uk/tools/clustalw/ ) , MUSCLE ( http://www . drive5 . com/muscle/ ) , and 3DCoffee ( http://igs-server . cnrs-mrs . fr/Tcoffee/tcoffee_cgi/index . cgi ) programs to perform a multiple sequence alignment of all ( 1 , 073 sequences ) or different subsets of ( from 50 sequences ) LmSIDER2 . None of these attempts produced in and of themselves a satisfactory alignment , probably because of the high degree of divergence and size polymorphism . The MUSCLE program produced a workable alignment from a selection of 50 full-length and relatively closely related LmSIDER2 sequences . This multiple alignment was manually refined to generate a framework used to manually align , one by one , the LmSIDER2 sequences . The final alignment contained 1 , 013 LmSIDER2 sequences ( Figure S1 ) The LmSIDER2 core sequence was generated by deleting all positions showing a gap for at least 50% of the aligned sequences , which represents 66 . 6% of the positions ( 1 , 074 positions out of 1 , 612 in the original alignment ) ( Figure S2 ) . The statistical and comparative analyses were performed using this LmSIDER2 core sequence . LmSIDERs were extracted from the most recent L . major genome annotation ( http://www . genedb . org/ ) for phylogenetic analysis . We extracted SIDER ( formerly named LmRIME ) sequence regions between 400 and 700 nucleotides long using Artemis [67] . An automated multiple sequence alignment was generated by comparing individual sequences to a Hidden Markov Model ( HMM ) using HMMER 1 . 8 . 5 ( http://hmmer . janelia . org/ ) . The HMM profile used to align the LmSIDERs was generated using 15 representative sequences selected from the manual alignment shown in Figure S1 ( 24 . 0477 , 36 . 1076 , 29 . 0524 , 31 . 0641 , 33 . 0760 , 36 . 1128 , 36 . 1087 , 35 . 1074 , 34 . 0878 , 31 . 0653 , 25 . 0573 , 38 . 0225 , 34 . 0863 , 14 . 0386 , 28 . 0581 ) . Limiting the amount of sequences in the profile minimizes position-specific base composition bias . To facilitate visualization of the subsequent tree , we removed additional LmSIDERs displaying >95% identity to at least one other aligned sequence using an ad-hoc JAVA script ( http://java . sun . com/ ) . The final alignment contains 785 LmSIDER sequences ( 140 LmSIDER1 and 645 LmSIDER2 ) . The 785 resulting LmSIDERs were submitted to a Minimum Evolution phylogenetic analysis based upon the number of differences using the MEGA3 program [68] . Furthermore , only parsimonious informative sites were considered . The phylogenetic tree was displayed using HyperTree JAVA program [69] . TbSIDER1 ( ten ) , TbSIDER2 ( 12 ) , RIME ( 70 ) , and NARTc ( 115 ) sequences were separately aligned using ClustalW ( http://www . ebi . ac . uk/tools/clustalw/ ) , whereas LmSIDER2 sequences ( 1 , 013 ) were manually aligned as described above . The core sequences , deduced from these alignments , were defined as described above for the LmSIDER2 core sequence . 21 , 2 , 44 , and 19 positions were removed from the original TbSIDER2 , TbSIDER1 , RIME , and NARTc alignments , which corresponds to 4% , 0 . 4% , 8 . 1% , and 6 . 7% of the positions , respectively . The core consensus sequences were reconstituted by considering the most conserved residue at each position of the alignment . Then , the percentage of substitution from the consensus was determined for each sequence aligned by calculating the sequence identity of each sequence with the consensus . The consensus sequence was created with BioEdit ( http://www . mbio . ncsu . edu/BioEdit/bioedit . html ) using a threshold frequency for inclusion of 26% . Gaps were treated like residues . To quantify the degree of conservation at each column in the core sequence multialignments , a chi-square ( χ2 ) score was computed comparing the observed distribution of ACGTs in the column to the distribution in the entire genome . The background ACGT distribution for the genome was obtained by counting the occurrences of each base in the set of all assembled chromosomes . Then , in each of the four multiple-alignments at each column , the chi-square score was computed as where oi is the observed number of occurrences of character i in the given column , and ei is the expected number of occurrences of character i computed as the proportion of character i in all assemblies multiplied by the number of sequences in that column of the multialignment . Using three degrees of freedom , a χ2 value of 16 . 3 corresponds to a significance level of p < 0 . 001 . The chromosomes and genomic coordinates of all L . major coding sequences were retrieved from version 4 . 0 of the assembly and annotation database hosted at TIGR . Using the predictive algorithm developed by Benz et al . [9] , we scanned all L . major chromosomes to locate the putative polypyrimidine tract and splice acceptor and polydenylation sites for each gene , thus delimiting the coordinates of the putative 5′UTR and 3′UTR . We selected the splice acceptor signal nearest to the start codon . This choice was based on what was observed in T . brucei , where EST mapping validated that 66% of the genes primarily used the closest site [9] . The distance between each LmSIDER element and its closest downstream and upstream gene on each chromosome strand was computed , disregarding the strand on which the element was located . The distance between each LmSIDER and the first methionine codon of the nearest downstream gene was calculated to determine a list of LmSIDERs that overlapped with the in silico–predicted 3′UTRs or 5′UTRs . Then , the distance between 3′UTR overlapping LmSIDERs and polypyrimidine and polyadenylation sites of the overlapping gene was calculated . The L . major LV39 strain used in this study was described previously [70] . Promastigotes were cultured at pH 7 . 0 and 25 °C in SDM-79 medium supplemented with 10% heat-inactivated FCS ( Wisent , http://www . wisent . ca/ ) and 5 μg/ml hemin . Intracellular L . major amastigotes were isolated from footpad lesions of infected BALB/c mice as previously described [71] . The expression vector pSPYNEOαLUC was described previously [16] and is referred to as LUC-control in the present study . The LUC-chimeric mRNAs transcribed from this vector are processed in Leishmania using sequences within the alpha-tubulin intergenic region cloned at the 5′-end . The different LUC-chimeric constructs listed in Figure 10 were made as follows . The full-length 3′UTR of LmjF36 . 3810 and LmjF08 . 1270 transcripts from the termination codon to 434 bp beyond the poly ( A ) site in the case of LmjF36 . 3810 , and to 84 bp beyond the poly ( A ) site in the case of LmjF08 . 1270 , or the LmSIDER2 element or the 3′UTR lacking LmSIDER2 , were amplified by PCR using Taq DNA polymerase ( Qiagen , http://www . qiagen . com/ ) and primers with inserted BamHI or PstI restriction sites ( see Table S3 ) . PCR products were cloned into vector pCR2 . 1 ( Invitrogen , http://www . invitrogen . com/ ) , digested with BamHI or PstI ( New England Biolabs , http://www . neb . com/ ) and subcloned into the BamHI site downstream of the LUC gene in vector pSPYNEOαLUC [16] . All constructs have been verified by sequencing . Purified plasmid vector DNA ( 10–20 μg , Qiagen ) were transfected into Leishmania by electroporation as described previously [72] . Stable transfectants were selected with 0 . 04 mg/ml G-418 ( Sigma , http://www . sigmaaldrich . com/ ) . The LUC activity of the recombinant parasites was determined as described previously [17] . Briefly , mid-log-phase promastigotes were diluted 1:100 in SDM-79 supplemented with 10% glycerol and counted in a Neubauer counting chamber . Equivalents of 4 × 107 and 2 × 107 parasites were spun , the pellet resuspended in 5× luciferase lysis ( Promega , http://www . promega . com/ ) buffer and frozen at −80°C . Twenty μl of each lysate was then mixed with an assay buffer ( Promega ) containing D-luciferin potassium salt , and LUC activity was measured in a luminometer ( Dynex MLX , http://www . dynextechnologies . com/ ) . Total RNA of L . major promastigotes was isolated using the TRIzol reagent ( Gibco BRL , http://www . invitrogen . com/ ) following manufacturer instructions . Northern blot hybridizations were performed following standard procedures [73] . To prepare soluble protein lysates , Leishmania cells were harvested by centrifugation , washed with ice-cold phosphate-buffered saline ( PBS ) , resuspended in Laemmli buffer , and syringed with a microsyringe ( ten times ) . Proteins were quantified using Amido Black 10B ( Bio-Rad , http://www . bio-rad . com/ ) , and 50 μg of total protein extracts were loaded onto 10% SDS-PAGE gels . The gels were transferred on a polyvinylidene difluoride membrane ( Immobilon-P; Millipore , http://www . millipore . com/ ) and the membranes were incubated for 90 min in blocking buffer ( PBS with 0 . 1% Tween 20 and 5% nonfat dry milk ) . The first antibody , a goat anti-luciferase pAB ( Promega ) diluted 1:10 , 000 in blocking buffer , was incubated with the membrane for 90 min with agitation . Following three washes with PBST ( PBS supplemented with 0 . 1% Tween 20 ) , a second antibody , a donkey anti-goat ( Santa Cruz Biotechnology , http://www . scbt . com/ ) diluted 1:10 , 000 in blocking buffer , was incubated for 45 min with the membrane . After additional washes , the blot was visualized by chemiluminescence with a Renaissance kit ( New Life Science Products , http://las . perkinelmer . com/ ) . RNA and protein levels were estimated by densitometric analyses using a PhosphorImager with ImageQuant 5 . 2 software . To determine the half-life of LmSIDER2-containing transcripts , mid-log phase L . major promastigote cultures were incubated with 10 μg/mL of actinomycin D ( Sigma ) , an inhibitor of de novo transcription . At specific times post-addition of the drug , 10-ml culture aliquots were pelleted by centrifugation , washed once with Hepes-NaCl buffer , and lysed in 1 ml TRIzol reagent ( Gibco BRL ) . Total RNA was extracted from these samples and subjected to northern blot hybridization . Quantitation of the different transcripts was done by densitometric analysis using a PhosphorImager with the ImageQuant 5 . 2 software . Thirty-eight L . major genes predicted to harbor LmSIDER2 in their 3′UTR were chosen for DNA microarray analysis , as part of a previously described 70-mer oligonucleotide array comprising a total of 154 selected genes [47] . Total RNA from L . major promastigotes and lesion amastigotes isolated from infected BALB/c mice was prepared using the TRIzol reagent ( Gibco BRL ) and purified using the RNAeasy kit ( Qiagen ) . Quality and quantity of the RNA was assessed by RNA 6000 Nano Assay Chips ( Agilent Technologies , http://www . home . agilent . com/ ) and a Bioanalyzer ( Agilent Technologies ) . Probes for microarray hybridization were prepared using the indirect Micromax TSA labeling and detection kit ( Perkin Elmer , http://las . perkinelmer . com/ ) . For each labeling reaction , 2 μg of purified RNA was spiked with two exogenous mRNAs ( NAC1 and CAB1 from Arabidopsis thaliana at 2 . 5 pg/μl; Stratagene , http://www . stratagene . com/ ) to adjust for variations in the incorporation efficiency of the modified nucleotides and differences in first-strand cDNA synthesis reactions . Hybridization , washes , and detection of fluorescence were done as described previously [47] . Four independent microarray experiments including dye swapping were scanned , and signal intensities for each spot were exported into GeneSpring software ( Agilent ) for further analysis . Local background was subtracted from each spot on the array , and intensity-dependent normalization was carried out within arrays . Cy5/Cy3 ratio for each spot was normalized with Cy5/Cy3 ratio for the A . thaliana NAC1 spike . Genes were only considered as statistically different in their expression if they satisfied a p-value cutoff of 0 . 05 . Expression ratios of three LmSIDER2-containing genes ( LmjF31 . 1890 , LmjF33 . 2550 , LmjF08 . 1270 ) and one non-LmSIDER2 gene ( LmjF16 . 1430 ) were confirmed by quantitative real-time RT-PCR as described previously [47] . These ratios were normalized using the GAPDH ratio to give a fold difference of expression . To exclude eventual amplification of mouse transcripts , cDNA from mouse macrophages served as negative control in each experiment . | Transposable elements ( TEs ) are DNA sequences capable of moving from one chromosomal region to another . A considerable fraction of higher eukaryote genomes is comprised of TEs , as exemplified in human ( over 40% of the genome ) and maize ( over 50% of the genome ) . There is now a growing body of evidence to suggest that TEs can be functionally important and not just “junk , ” “selfish , ” or “parasitic” DNA sequences that make as many copies of themselves as possible . Indeed , during the past ten years , a considerable number of TE copies have been described as domesticated or exapted elements playing a cellular function , such as transcriptional regulation and contribution to protein-coding regions . TE domestication has been described for only a few copies of TE families , and exaption of a whole TE family has not been reported so far . We provide evidence that Leishmania spp . , unicellular protists responsible for human diseases , have recycled and expanded a whole family of short and extinct TEs ( retroposons ) that have evolved to fulfill an important biological pathway , i . e . , regulation of gene expression . We also observed that Trypanosoma brucei ( a close relative of Leishmania spp . ) developed other approaches to maintain the same cellular function . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"infectious",
"diseases",
"trypanosoma)",
"microbiology",
"computational",
"biology",
"evolutionary",
"biology",
"molecular",
"biology",
"kinetoplastid",
"(leishmania"
]
| 2007 | Members of a Large Retroposon Family Are Determinants of Post-Transcriptional Gene Expression in Leishmania |
Immunization vectors based on cytomegalovirus ( CMV ) have attracted a lot of interest in recent years because of their high efficacy in the simian immunodeficiency virus ( SIV ) macaque model , which has been attributed to their ability to induce strong , unusually broad , and unconventionally restricted CD8+ T cell responses . To evaluate the ability of CMV-based vectors to mediate protection by other immune mechanisms , we evaluated a mouse CMV ( MCMV ) -based vector encoding Friend virus ( FV ) envelope ( Env ) , which lacks any known CD8+ T cell epitopes , for its protective efficacy in the FV mouse model . When we immunized highly FV-susceptible mice with the Env-encoding MCMV vector ( MCMV . env ) , we could detect high frequencies of Env-specific CD4+ T cells after a single immunization . While the control of an early FV challenge infection was highly variable , an FV infection applied later after immunization was tightly controlled by almost all immunized mice . Protection of mice correlated with their ability to mount a robust anamnestic neutralizing antibody response upon FV infection , but Env-specific CD4+ T cells also produced appreciable levels of interferon γ . Depletion and transfer experiments underlined the important role of antibodies for control of FV infection but also showed that while no Env-specific CD8+ T cells were induced by the MCMV . env vaccine , the presence of CD8+ T cells at the time of FV challenge was required . The immunity induced by MCMV . env immunization was long-lasting , but was restricted to MCMV naïve animals . Taken together , our results demonstrate a novel mode of action of a CMV-based vaccine for anti-retrovirus immunization that confers strong protection from retrovirus challenge , which is conferred by CD4+ T cells and antibodies .
In the last two decades , vector-based immunization approaches for the development of an HIV vaccine have been pursued intensively , and recently vectors based on cytomegalovirus ( CMV ) have drawn a lot of interest . At first glance , CMV is not an obvious choice as basis for a vaccine vector: as a β-herpes virus it carries a large and highly complex genome [1] that encodes numerous immune evasion proteins interfering with many aspects of immunity [2] , and CMV infection is associated with severe illness in immune compromised or immature patients [3] . However , after a long period of productive replication following the primary infection , CMV establishes latency from which repetitive episodes of virus reactivation can occur , leading to recurrent rounds of immunogen expression and creating a self-boosting vaccine . Furthermore , the natural CMV infection can induce inflationary T cell responses , which do not contract after the effector phase but keep expanding and can reach very high frequencies ( reviewed in [4 , 5] ) , maybe a desirable feature of vaccine-induced immunity . In recent years , CMV-based vectors for immunization have drawn increasing interest . There have been a number of approaches evaluating the murine CMV ( MCMV ) as a vaccine vector in mice . For the induction of CD8+ T cell based immunity , epitope-based vaccines have been constructed using epitopes from influenza virus [6] , lymphocytic choriomeningitis virus [6] or Ebola virus [7] as sole immunogens , which induced strong immune responses and protection in the respective challenge models . For immunization against Mycobacterium tuberculosis , an MCMV vector encoding a tetanus toxin fragment was tested in a mouse model and was found to induce an antibody-dominated response [8] . Similarly , a rhesus CMV ( RhCMV ) based vaccine encoding an Ebola virus glycoprotein conferred protection to macaques from Ebola virus challenge but induced mainly antibody and not cellular immune responses [9] . Finally , RhCMV-based vectors were developed in the simian immunodeficiency virus ( SIV ) infection model in non-human primates and were shown to confer very strong protection in half of the vaccinated monkeys [10] . Interestingly , RhCMV-based immunization induced very broad CD8+ T cell responses to epitopes presented on major histocompatibility complex ( MHC ) type II and MHC-I E [11 , 12] , which was caused by deletion of multiple genes in this RhCMV vector [11 , 13] . To evaluate the potential of CMV-based immunization when neither vector design nor immunogen choice are targeted at the induction of CD8+ T cell responses , we constructed a vaccine vector based on MCMV that we employed in a mouse retrovirus model . Friend retrovirus ( FV ) is a murine retrovirus complex consisting of the apathogenic , replication-competent Friend murine leukemia virus ( F-MuLV ) and the replication-defective but pathogenic spleen focus forming virus ( SFFV; [14] ) . Infection of susceptible mice results in the rapid development of splenomegaly and erythroleukemia due to an aberrant activation of the erythropoietin receptor by the SFFV envelope protein gp55 , whereas mice that are genetically resistant to FV-induced disease develop a chronic infection ( reviewed in [15] ) . FV is regarded as a useful retrovirus mouse model that allows for insights into immunological control of retrovirus infections in general , and it shows similarities to HIV infection with regard to the establishment of persistent reservoirs in immunologically privileged sites [16] as well as immunosuppression driven by regulatory T cells [17 , 18] . As low doses of FV are sufficient to rapidly induce disease in susceptible mice , FV infection is a very stringent mouse model for the development and evaluation of immunization strategies . We and others have employed the FV model extensively in the past to develop improved immunization strategies and to analyse mechanisms underlying protection conferred by different vaccines against retroviruses . Vaccines based on attenuated F-MuLV or FV [19–23] , inactivated F-MuLV [24] , protein or peptide vaccines [25–28] , nanoparticle-based vaccines [29] and vector-based vaccines [30–40] have been tested and shown to confer widely different degrees of protection . The most potent vaccine in the FV model described until now is live-attenuated F-MuLV , which completely protects even highly susceptible mice from FV infection [21] . It has been demonstrated that a complex immune response comprising antibodies as well as CD4+ and CD8+ T cells is necessary for this protection [19 , 20] . Interestingly , we could demonstrate in adenovirus-based vaccine studies that very potent , albeit not sterile , protection can also be conferred if only individual immune components are induced , as the induction of strong CD8+ T cell responses will also protect highly susceptible mice from FV-induced disease and allow them to control FV infection at a very low level [35] . In a side-by-side comparison we showed that the refined employment of adenovirus-based vaccines mediated protection from FV infection that was almost as strong as that conferred by immunization with attenuated F-MuLV . However , the mechanisms underlying protection conferred by the two vaccines differed significantly , as the adenovirus-based vector induced T cell dominated responses , whereas the F-MuLV immunization induced little T cell responses but highly superior antibody responses [40] . To evaluate the general potency of MCMV-based immunization in the FV model , we constructed an MCMV vector encoding F-MuLV envelope , without introducing any modifications into the vector aiming at unconventionally restricted CD8+ T cell induction . In the work presented here , we show that this envelope-encoding MCMV-based vaccine confers strong protection in the FV mouse model , without any evidence for a contribution of vaccine-induced CD8+ T cells .
To analyse the potential of CMV-based vectors in the FV mouse model , we constructed a vector based on mouse cytomegalovirus ( MCMV ) encoding the envelope protein ( Env ) of Friend murine leukemia virus ( F-MuLV ) and tested its efficacy at preventing FV infection of highly susceptible CB6F1 mice . To prevent rapid control of the MCMV vector by natural killer cells , the m157 coding sequence was partially deleted , leaving the neighbouring open reading frames intact [41] , and replaced by the Env transgene expression cassette . An MCMV vector on the m157-deleted background without any transgene was used as a control vector . The expression of F-MuLV Env in MCMV . env infected cells was verified by immunoblot analysis ( Fig 1A ) . The presentation of immunogens on the vaccine vector particles can be beneficial for the induction of immunogen-specific antibody responses [32] , therefore we also analysed if Env is incorporated into the MCMV . env virions . However , the analysis of purified MCMV . env particles by immunoblot gave no indication of incorporation of significant amounts of Env protein ( Fig 1B ) . A comparison of in vitro replication showed comparable progeny titers for wildtype ( wt ) MCMV , the m157-deleted control vector and MCMV . env ( Fig 1C ) . To compare the in vivo replication rates , we infected CB6F1 mice with 2 x 105 PFU of the m157-deleted control vector or MCMV . env , or wt MCMV , and analysed MCMV titers in spleen , kidney and salivary gland tissue on day 3 and day 21 after infection . While the titers of the m157-deleted vectors were slightly higher on day 3 in spleen and kidney tissues compared to wt MCMV , they were mostly cleared from these tissues on day 21 p . i . ( Fig 1D and 1E ) . The titers in salivary glands on day 21 p . i . showed higher variability but seemed comparable for wt MCMV and the m157-deleted MCMV . env , whereas 2 of 3 samples showed higher titers for the m157-deleted control vector ( Fig 1F ) . Taken together , these data suggest that the deletion of m157 does not severely influence pathogenicity of MCMV in CB6F1 mice that we used in the subsequent vaccination experiments . When we analysed the MCMV-specific immune response , we found that mice infected with either of the m157-deleted vectors mounted comparable MCMV-specific CD8+ T cell responses ( Fig 2A ) . An analysis of CD8+ T cells specific for conventional ( M45985-993 , M57816-824 , M102446-455 and m14115-23; Fig 2B ) or inflationary epitopes ( m139419-426; Fig 2C ) 14 days or 77 days after a single or repeated infection with MCMV . env revealed a sustained MCMV-specific CD8+ T cell response , with no significant difference between the single or the repeat administration . Interestingly , the MCMV-specific CD8+ T cell response was fairly low overall , which may be characteristic for this particular mouse strain . To analyse the protective potential of the new MCMV . env construct , we immunized CB6F1 mice with 2 x 105 PFU MCMV . env , or the MCMV vector without transgene as a control , and infected them with FV three weeks later to analyse the protection from FV challenge . However , the immunized mice did not show an improved control over FV induced splenomegaly compared to unvaccinated mice ( Fig 3A ) , and we did not observe any significant differences in overall viremia levels on day 10 after FV infection ( Fig 3B ) or in spleen weights ( Fig 3C ) or spleen viral loads on day 21 after FV challenge infection ( Fig 3D ) , even though individual mice were able to control the FV infection . Seeking to improve the vaccination efficacy , we evaluated the application of a second immunization , and repeated the MCMV . env immunization three weeks after the first immunization ( MCMV . envtwice ) . Since productive MCMV replication is observed for a long period after primary infection ( > 21 days; Fig 1F ) that should lead to long-lasting immunogen expression , we also included a group of mice that was immunized only once , but challenged with FV at the delayed time point of 6 weeks post immunization ( MCMV . envsingle , late ) . Surprisingly , when mice were infected with FV six weeks after the start of the immunization regimen , both groups of MCMV . env immunized mice controlled the infection equally well . All MCMV . env immunized mice had significantly smaller spleens than unvaccinated mice or mice vaccinated with the control MCMV throughout the observation period ( Fig 4A ) . While unvaccinated mice or the control MCMV immunized mice had high viremia levels at day 10 after FV challenge infection , most MCMV . env immunized mice of either group had no detectable viremia ( Fig 4B ) . Similarly , when the spleens were collected three weeks after FV infection , the MCMV . env immunized mice had significantly smaller spleens than unvaccinated mice , with the spleens of most of these mice exhibiting a normal weight ( Fig 4C ) . The viral loads in spleens of unvaccinated mice and of mice immunized with the control MCMV were equally high , whereas the spleen viral loads of mice from either MCMV . env immunized group were significantly reduced in comparison to both unvaccinated and control MCMV immunized mice , and FV was actually undetectable in many mice ( Fig 4D ) . These results demonstrate that a single immunization with MCMV . env confers strong protection against a delayed FV challenge infection , and that in this delayed challenge setting , the single immunization regimen is not inferior to a repeat immunization schedule . Seeking for an explanation for the improved protection after delayed FV challenge infection , we analysed the immune responses to the different vaccination regimens . Analysing the antibody responses to MCMV . env immunization , we found that two weeks after a single immunization , binding antibody levels were rather low and at the limit of detection in about half of the mice; only two mice showed higher antibody levels ( Fig 5A ) . When mice were challenged early after the single immunization , the anamnestic neutralizing antibody response ten days after FV challenge was low , albeit significantly improved compared to unvaccinated mice ( Fig 5B ) . In the prolonged vaccination regimen , mice that were immunized once and mice that had received a repeat immunization had comparable binding antibody levels five weeks after initiation of the immunization , but they were still rather low with median values at the detection limit ( Fig 5C ) . Only very low neutralizing antibody responses were detected in few immunized mice at this time point ( Fig 5D ) . However , most mice of both immunized groups were able to mount a robust neutralizing antibody response ten days after FV challenge infection that was significantly higher than in unvaccinated mice or in mice that were immunized with the control MCMV ( Fig 5E ) . These findings suggest that the ability of mice to mount a neutralizing antibody response after FV infection may be crucial for the MCMV . env mediated protection , and that the short time period to the early FV challenge is insufficient for the maturation of the antibody response . In fact , a Spearman ranked correlation analysis of the data obtained for the FV challenge 3 weeks or 6 weeks after MCMV . env immunization suggests an inverse correlation of neutralizing antibody levels 10 days after FV challenge with spleen viral loads 21 days after FV challenge ( r = -0 . 5751; P = 0 . 0005; Fig 5F ) . To determine the cellular immune responses underlying the MCMV . env mediated protection , we analysed the induction of Env123-141-specific CD4+ T cells by MHC II tetramer staining . Already two weeks after the first immunization , MCMV . env immunized mice had mounted a clearly detectable Env123-141-specific CD4+ T cell response ( Fig 6A ) , that was absent in the control MCMV immunized mice . When CD4+ T cell responses were analysed three weeks later , we could still detect significant levels of Env123-141-specific CD4+ T cells after the single immunization ( Fig 6B ) , and there was no significant difference in the Env123-141-specific CD4+ T cell response in mice that had received a second immunization three weeks after the first immunization . Interestingly , when we performed a Spearman ranked correlation analysis , we found only a poor correlation between the frequency of Env-specific CD4+ T cells determined one week before FV challenge infection and the viral load in spleens 3 weeks after FV challenge infection ( r = 0 . 1076; P = 0 . 5931; Fig 6C ) . We also analysed the cytokine production of Env-specific CD4+ T cells after in vitro restimulation ( Fig 6D ) ; while responses were not very strong , we observed a trend to a higher frequency of IFNγ and IL17 producing CD4+ T cells in MCMV . env immunized mice compared to control MCMV immunized mice , suggesting that apart from providing help for the strong anamnestic antibody response , Env-specific CD4+ T cells may also exhibit some direct antiviral effector functions and provide help for CD8+ T cell induction . To analyse the localization of Env-specific CD4+ T cells , we collected lymph nodes , spleens and peripheral blood mononuclear cells ( PBMC ) from MCMV . env immunized or unvaccinated mice 14 days after MCMV . env immunization or 21 days after FV challenge infection and subjected the cells to MHC II tetramer staining to detect Env123-141-specific CD4+ T cells . After MCMV . env immunization , we detected an appreciable frequency of Env123-141-specific CD4+ T cells in PBMC of most MCMV . env immunized mice , as described above , but very low frequencies in lymph nodes and spleens ( Fig 7A ) . We also stained the cells for expression of the chemokine receptor CXCR5 as a surrogate marker for a follicular helper phenotype and found that on average half of the Env-specific CD4+ T cells expressed CXCR5 . After FV infection , the MHC II tetramer staining revealed similar frequencies of Env-specific CD4+ T cells in unvaccinated and MCMV . env immunized mice in lymph nodes and PBMC , but a significantly higher frequency in spleens ( Fig 7B ) . Again , about half of the Env-specific CD4+ T cells expressed CXCR5 in MCMV . env immunized as well as in unvaccinated mice . The frequency of Env-specific CD4+ T cells after FV infection was lowest in lymph nodes , which carry only relatively low viral loads in FV infection [16] . In immunization studies using RhCMV for vaccination of rhesus macaques against SIV , unconventionally restricted CD8+ T cell responses were demonstrated: the RhCMV-induced CD8+ T cells recognized a very high number of epitopes derived from the vaccine immunogens , and they were unusual in recognising peptides presented by MHC type II or E [11 , 12] . We had not introduced any modifications into the MCMV vector that have been described to be required for the induction of unconventionally restricted CD8+ T cell responses by RhCMV vectors [11 , 13] , still it was of great interest to analyse the CD8+ T cell response after MCMV . env immunization in more detail . In the natural FV infection , a role of CD8+ T cells recognising F-MuLV Env is questionable: while Ruan et al . have described an H2-Db restricted CD8+ T cell epitope in F-MuLV Env [42] it has never been confirmed by later work , and the CD8+ T cell response in FV infection has been shown to be dominated by cells recognizing the Leader-Gag derived epitope GagL85-93 [43] , which is not part of our MCMV . env vaccine . To analyse if the MCMV . env vaccine induces Env-specific CD8+ T cells , we performed an in vitro stimulation assay . We isolated spleen cells from MCMV . env or control MCMV immunized mice six weeks after immunization , or from MCMV . env immunized , FV infected mice 21 days after FV challenge , depleted the spleen cells of CD4+ cells , and stimulated them with Env-derived peptide pools in an IFNγ ELISpot assay , or with a pool of CMV-derived peptides as control . While the MCMV-specific CD8+ T cell response was readily detectable , stimulation of spleen cells with Env-derived peptide pools did not result in a significant number of IFNγ spots ( Fig 8A ) . Of note , no response was detected to pool 8 that contains the previously described putative CD8+ T cell epitope [42] . These results indicate that , as intended , immunization with MCMV . env did not induce any appreciable Env-specific CD8+ T cell response . Similar results were obtained when whole spleen cells from MCMV or MCMV . env immunized mice were restimulated with Env-derived peptide pools and production of various cytokines was analysed by flow cytometry . There was no appreciable production of IFNγ , TNFα or interleukin 2 ( IL2 ) by CD8+ T cells of MCMV . env immunized mice either before or after FV infection ( Fig 8B ) . To analyse if CD8+ T cells are actually dispensable for control of the FV challenge after MCMV . env immunization , we performed depletion experiments where CD8+ T cells , or CD4+ T cells as control , were depleted from MCMV . env immunized mice starting one week before FV challenge infection . Surprisingly , depletion of either CD4+ or CD8+ cells resulted in loss of control over splenomegaly , with severely enlarged spleens from day 10 in contrast to non-enlarged or minimally enlarged spleens in MCMV . env immunized , non-depleted mice ( Fig 9A ) . Similarly , when we analysed viral loads in plasma on day 10 after FV challenge and spleen viral loads on day 21 after FV challenge , we found that depletion of CD4+ cells as well as CD8+ cells resulted in high viral loads in most mice ( Fig 9B and 9C ) . The depletion of CD8+ T cell had a detrimental effect on MCMV . env-induced protection from FV challenge , even though the peptide stimulations gave no indication for the presence of Env-specific CD8+ T cells . This counterintuitive result could be explained by the fact that after FV challenge infection , also MCMV . env immunized mice mount a CD8+ T cell response against the GagL85-93 epitope , which is not part of the MCMV . env vaccine . When we analysed the GagL85-93-specific CD8+ T cell response ten days after FV infection , we found indeed that unvaccinated mice and MCMV . env immunized mice mounted a comparable GagL85-93-specific CD8+ T cell response ( Fig 9D ) . The development of disease in most CD4+ cell depleted mice underlines the importance of CD4+ T cells for the MCMV . env-mediated protection . To confirm the mechanistic role of MCMV . env-induced CD4+ T cells in providing help for the induction of potent antibody responses upon FV challenge infection , we compared the levels of neutralizing antibodies in MCMV . env immunized mice that were challenged with FV without further intervention , or after depletion of CD4+ cells , and found significantly reduced levels of neutralizing antibodies on day 10 after FV infection in mice depleted of CD4+ cells compared to MCMV . env immunized mice ( Fig 9E ) . Only few of the CD4+ cell depleted MCMV . env immunized mice mounted a detectable neutralizing antibody response; importantly , those were the few CD4+ cell-depleted mice that were able to control the FV challenge ( compare Fig 9B and 9C ) and sera of these mice collected before FV challenge had also shown some neutralizing activity . While the overall correlation of neutralizing antibody titers before FV infection and viremia levels after FV infection was moderate ( r = -0 . 3876; P = 0 . 2081; Fig 9F ) , the two mice that controlled the FV infection did indeed exhibit the highest neutralizing antibody levels before FV challenge , supporting the important role of antibodies in MCMV . env-mediated protection . To corroborate our findings that showed the important role of antibodies in FV control after MCMV . env immunization and to confirm the absence of protective CD8+ T cells after MCMV . env immunization , we performed transfer experiments using CD8+ T cells isolated from MCMV . env immunized mice either before or after FV challenge infection , or using plasma isolated from control MCMV or MCMV . env immunized mice 14 days after FV challenge infection . As expected , transfer of CD8+ T cells did not result in significant control of FV infection , with viral loads in plasma ( Fig 10A ) and spleens ( Fig 10C ) that were not significantly reduced compared to unvaccinated mice . While transfer of plasma from control MCMV immunized mice collected after FV challenge had no significant impact on FV control , transfer of plasma from MCMV . env immunized mice on the other hand conferred complete protection to recipient mice , with undetectable viral loads in plasma and spleens ( Fig 10A and 10C ) and very low spleen weights ( Fig 10B ) . An analysis of binding ( Fig 10D ) and neutralizing antibodies ( Fig 10E ) revealed similar levels of binding antibodies , and reduced levels of neutralizing antibodies , in recipients of MCMV . env immunized mouse-derived plasma compared to MCMV . env immunized mice . These results support the idea that MCMV . env immunized mice mount very potent antibody responses after FV challenge infection , which are able to mediate control over FV infection . To furthermore prove the importance of the antibody response in MCMV . env immunization , we depleted immunized mice of B cells using a CD20-specific B cell depleting antibody . In contrast to MCMV . env immunized mice , mice that were immunized with MCMV . env and depleted of B cells before FV challenge infection mostly failed to control FV infection . Only one B cell depleted mouse was able to control viremia ( Fig 11A ) , and all B cell depleted mice had higher viral loads in spleens than non-depleted mice going up to the same level as unvaccinated mice ( Fig 11B ) . The B cell depletion did not work equally well in all mice; interestingly , when we performed correlation analyses , we found strong inverse correlations between the B cell frequency compared to non-depleted mice and the viral loads in plasma as well as spleens ( viremia: r = -0 . 9411 , P = 0 . 017; spleen virus load: r = -0 . 9662 , P = 0 . 0074; Fig 11C and 11D ) . Taken together , these results strongly support the important role of antibodies and B cells in MCMV . env mediated protection . The induction of long-lasting immunity is a pre-requisite for any vaccine . Therefore , we analyzed if the MCMV . env immunization would also protect from an FV challenge infection applied three months later . CB6F1 mice were immunized once or twice as described before , and the development of FV-specific immune responses was monitored over time . The Env-specific CD4+ T cell response contracted quickly after the immunization to frequencies below 0 . 5% , and again we did not observe any impact of the second MCMV . env application ( Fig 12A ) . The titers of F-MuLV-binding antibodies did not change significantly and were rather low during the whole observation period ( Fig 12B ) . When mice were infected with FV three months after the initial immunization , they were able to mount strong neutralizing antibodies as observed in the previous experiments described above ( Fig 12C ) , and controlled the FV infection tightly with low viral loads in plasma ( Fig 12D ) , no splenomegaly ( Fig 12E ) and low spleen viral loads ( Fig 12F ) . These results show that protection conferred by MCMV . env immunization is long-lived . Another important consideration in the development of vector-based immunization strategies is the influence of pre-existing immunity against the vector . Therefore , we performed a vaccination experiment in mice that were infected with control MCMV five weeks before the MCMV . env immunization to induce MCMV immunity . When mice were infected with FV six weeks after the MCMV . env immunization , pre-immune mice were not able to control the FV infection and had high viral loads in plasma and spleens and developed splenomegaly ( Fig 13A–13C ) . An analysis of the immune responses revealed a trend to reduced induction of Env-specific CD4+ T cells and a severely impaired induction of neutralizing antibodies upon FV challenge infection in MCMV pre-immune mice compared to pre-naïve MCMV . env immunized mice ( Fig 13D and 13E ) . Taken together , our results show that MCMV . env immunization leads to strong protection of highly susceptible mice from FV infection , which relies on a strong anamnestic neutralizing antibody response that fully expands only after FV infection and is not sufficient to prevent infection at challenge , therefore , vaccine-induced CD4+ T cells as well as intrinsic CD8+ T cells are equally required for control of FV challenge . While the strong protection mediated by MCMV . env was long-lasting , it was limited to MCMV pre-naive mice .
CMV-based vectors have been one of the most attractive vector systems of the last years , mainly because a RhCMV-based SIV vaccine was able to confer potent protection to rhesus macaques from SIV infection . We show here that in the FV mouse model , an MCMV-based vector encoding F-MuLV Env conferred strong protection to highly FV-susceptible mice . Protection correlated strongly with the ability of the mice to mount a rapid and strong anamnestic antibody response upon challenge , but not directly with the strength of the Env-specific CD4+ T cell response , even though the cytokine profile of the Env-specific CD4+ T cells suggest that they may have some direct antiviral activity and thus contribute to control of the FV infection . It is a very interesting finding that MCMV . env immunization resulted in induction of strong CD4+ T cells as previous work with an MCMV vector encoding a tetanus toxin fragment had demonstrated potent antibody responses , but no appreciable cellular responses [8] , similar to a RhCMV-based vaccine against Ebola virus that was tested in rhesus macaques [9] . The data presented here show in fact the strongest CD4+ T cell response to any FV vaccine that we have tested so far: adenovirus-based vectors as well as attenuated F-MuLV induced Env-specific CD4+ T cell responses that were far lower than the response shown here for MCMV . env immunization [40] . The antibody response to MCMV . env immunization , on the other hand , was rather low and strengthening the antibody response might be a promising strategy for further improvement of MCMV vector-based immunization efficacy . We did not detect any Env protein incorporated in MCMV . env virions , and the targeted incorporation of Env into the MCMV particle could possibly lead to improved Env-specific antibody responses in the same way as we could demonstrate for a so-called expression-display adenovirus-based vector that presents the F-MuLV Env gp70 protein on the capsid surface [32] . There is only a limited number of epitopes known in FV , and whereas the number of described FV-derived CD4+ T cell epitopes has increased in recent years [44] , the Leader-Gag-derived GagL85-93 peptide is still the only confirmed CD8+ T cell epitope [43] . A CD8+ T cell epitope that was described for the Env protein [42] has never been confirmed in subsequent studies , and also our new data from MCMV . env immunized mice does not provide any indication for the presence of CD8+ T cells reactive to Env . Therefore , a strategy that might result in improved protection could be the incorporation of the Leader-Gag protein or the GagL85-93 epitope into an MCMV vector , as it has been demonstrated before that MCMV-based epitope vaccines were able to induce very potent immunogen-specific CD8+ T cell responses [6 , 7] . However , we have shown that the GagL85-93 epitope is rather weak and can be sub-dominant to vector-derived epitopes [45] , which might result in impaired GagL85-93-specific CD8+ T cell responses in the MCMV background . We have also shown before that Env suppresses CD8+ T cell responses to simultaneously or subsequently administered immunogens [40 , 46]; interestingly , we did not observe a suppressive effect of Env on MCMV-specific CD8+ T cell responses , so it would be intriguing to see if CD8+ T cell responses that are suppressed in DNA or adenovirus-based immunization are also unaffected by Env in the MCMV background . While the MCMV . env vaccine did not induce any Env-specific CD8+ T cells , we showed that the presence of CD8+ T cells at the time of challenge is required as mice mount an intrinsic , GagL85-93-specific CD8+ T cell response upon FV challenge that contributes to control . Interestingly , transfer of CD8+ T cells from MCMV . env immunized , FV-challenged mice did not confer protection to recipient mice , which may be attributable to the relatively low frequencies of GagL85-93-specific CD8+ T cells in the FV-challenged mice . In the past we have developed an adenovirus-based epitope vaccine which induced exclusively GagL85-93-specific CD8+ T cells and conferred strong protection from FV challenge [35] , but there the frequency of GagL85-93-specific CD8+ T cells was ~ 10% , i . e . approximately five times as high as the frequency observed here in the MCMV . env immunized , FV-challenged mice . Furthermore , we infected mice with the high dose of 5 000 SFFU , which is not easily controlled by the highly FV-susceptible CB6F1 mice , demonstrating the stringency of our infection model and the potency of the MCMV . env vaccine . One feature that makes CMV an interesting vaccine vector is its ability to establish a persistent infection , allowing for ongoing transgene expression and thereby acting as a self-boosting vaccine . To prevent strong activation of natural killer ( NK ) cells , we utilized an m157 deleted MCMV for vector construction [41]; this modification did not result in an exacerbated pathology in the infected mice , but did allow for slightly elevated replication in the early phase of MCMV infection . This reduced activation of NK cells may also explain why , contrary to a previous report [47] , we did not observe an effect of the control MCMV immunization on FV infection . Our challenge experiments showed that the outcome of an early FV challenge was far more variable than that of a late FV challenge , which we attribute to the higher neutralizing antibody levels that mice were able to mount after the late FV challenge . This delay in protection argues for the hypothesis that ongoing immunogen production by the MCMV . env vector promotes the development of protective immunity . On the other hand , antigen presenting cells of the B cell follicle are able to incorporate an antigen reservoir and to provide the antigen for B cell maturation for a prolonged time; therefore , experiments with a single-cycle MCMV would be necessary to confirm the supportive role of vaccine vector persistence . The development of a more attenuated or single-cycle MCMV might also be desirable for a translation into a CMV based vaccine for application in humans to increase the vaccine’s safety profile . Similarly , it would be interesting to analyse if different application routes result in comparable immune responses , as that would be necessary for a translation into vaccination of humans , and both questions shall be addressed in the future . It has been demonstrated in the past using attenuated retrovirus-based vaccines that complex immune responses are necessary to confer full protection from retrovirus infection [20]; however , in immunization experiments with other vaccines , we could show that very potent vaccines can be generated that induce only partial immune responses . When we immunized mice with calcium phosphate nanoparticles encapsulating the immunodominant Env123-141 CD4+ T cell epitope and the GagL85-93 CD8+ T cell epitope , highly susceptible CB6F1 mice that we also used in the study presented now were mostly able to control FV-induced disease and exhibited significantly reduced spleen viral loads [29] . Mice that we immunized with an advanced scheme of adenovirus-based immunization vectors mounted robust CD4+ and CD8+ T cell responses but low neutralizing antibody responses upon immunization , but were strongly protected from FV-induced disease and had very low spleen viral loads around the detection limit [40] . Interestingly , mice that we immunized with Fv-1b-restricted , N-tropic F-MuLV ( F-MuLV-N ) that is highly attenuated in CB6F1 mice ( Fv-1b/b ) showed the strongest control over FV , but here the protection seemed to rely mostly on neutralizing antibodies , as no FV-specific CD8+ T cells and low CD4+ T cells were detectable after immunization [40] . As the F-MuLV-N immunization was superior even to the advanced adenovirus-based immunization , these findings highlight the importance of antibodies for protection from retrovirus infection , and the results obtained in the MCMV immunization study presented here are well in line with these results . In our model , the MCMV immunization did not seem to induce any FV-specific CD8+ T cell response , and protection was established with a delay after immunization , arguing for an important role of an antibody response that requires longer time to mature . Importantly , the MCMV-based vaccine allowed even the highly FV-susceptible mice used in this study to tightly control both FV-induced disease and FV loads; many mice displayed spleen viral loads below the detection limit of the immunocytochemical assay , and this is a level of protection that is comparable to the protection we observed after F-MuLV-N immunization under identical conditions [40] , underlining the potency of our MCMV . env vaccine . Our data show that immunization efficacy was severely affected by a prior infection with MCMV . Most reports on CMV-based immunization have not addressed the influence of pre-existing immunity; the studies using RhCMV vectors for immunization against SIV have been performed in naturally RhCMV infected macaques though , and found the vaccine to be highly effective [10] . In contrast to the vector design employed in the macaque studies , which induces strong and broad CD8+ T cell responses , our vector design that leads to strong CD4+ T cell and anamnestic antibody responses is restricted to MCMV-naïve vaccinees . It would have to be explored if a different design of our MCMV vector would allow for induction of transgene-specific immunity in pre-existing MCMV immunity . The expression of the immunogen under an immediate early promotor might result in earlier accumulation of immunogen than with our current MCMV . env construct and provide meaningful amounts of immunogen in spite of early clearance of MCMV . env infected cells . Also the incorporation of Env into the MCMV particle , as discussed above , might be useful to allow for the induction of Env-specific immunity in spite of pre-existing MCMV immunity . Strategies such as these should be developed and carefully evaluated to obtain MCMV vectors that are also highly efficacious in pre-existing MCMV immunity . Overall , our results show that in our mouse model , MCMV-based vaccines prove highly effective tools for vaccination that lead to protection that is comparable in strength to the protection conferred by immunization with attenuated retrovirus , and highlight the importance of pursuing the development of CMV-based vectors further .
Mouse experiments were performed in accordance with the guidelines of the University Hospital Essen , Germany , the national animal protection law ( Tierschutzgesetz ( TierSchG ) ) and animal experiment regulations ( Tierschutz-Versuchstierverordnung ( TierSchVersV ) ) , and the recommendations of the Federation of European Laboratory Animal Science Association ( FELASA ) . The study was approved by the Northrhine-Westphalia State Office for Nature , Environment and Consumer Protection , Section 81 “Animal Research Affairs” ( LANUV NRW , Düsseldorf , Northrhine-Westphalia , Germany; permit numbers 84–02 . 04 . 2014 . A175 and 84–02 . 04 . 2017 . A091 ) . A murine fibroblast cell line from Mus dunni [48] and the murine hybridoma cell line 720 [49] ( both cell lines kindly provided by Dr . Kim J . Hasenkrug , NIAID , NIH , Hamilton , MT ) were maintained in RPMI medium ( Invitrogen/Gibco , Karlsruhe , Germany ) supplemented with 10% heat-inactivated fetal bovine serum ( Invitrogen/Gibco ) , 50 μg/ml gentamicin and 20 μg/ml ciprofloxacin . Primary mouse embryonic fibroblasts ( MEF ) were isolated according to described protocols [50] . Cells were maintained in a humidified 5% CO2 atmosphere at 37°C . The recombinant mouse cytomegalovirus ( MCMV; Murid herpesvirus 1 ) MCMV . env encoding Friend murine leukemia virus Env has been described before [51] . Briefly , MCMV . env was constructed by inserting an expression cassette containing the human CMV major immediate early promoter/enhancer and the F-MuLV Env coding sequence into the m157 open reading frame of an MCMV with MCK2-repaired background; incorporation as well as retention of the transgene was confirmed by PCR . An m157-deleted MCMV vector without transgene was used as control . For in vitro and in vivo characterization , a wildtype MCMV [52] was used as control . Uncloned , lactate dehydrogenase-elevating virus ( LDV ) -free FV stock was obtained from BALB/c mouse spleen cell homogenate ( 10% , wt/vol ) 14 days post infection with a B-cell-tropic , polycythemia-inducing FV complex [53] . For the analysis of transgene expression , MEF cells were infected with the MCMV vectors at an MOI of 5 . Cells were collected 4 , 24 and 48 hours after infection and cell lysates were subjected to SDS-PAGE and Western Blot analysis , probing with the F-MuLV gp70-specific hybridoma-derived antibody 720 [54] , an MCMV pIE1-specific antibody ( CROMA101; provided by Stipan Jonjić , University of Rijeka , Croatia ) and an actin-specific antibody ( Sigma-Aldrich , Munich , Germany ) . For the analysis of Env incorporation into vector particles , 2 × 105 PFU MCMV . env or MCMV , or different amounts of F-MuLV particles as positive control ( 1 × 105 FFU , 5 × 103 FFU , or 1 × 103 FFU ) were subjected to SDS-PAGE and Western Blot analysis , using a polyclonal goat-anti-gp70 antibody ( kindly provided by Dr . Christine Kozak , NIAID , NIH , Bethesda , MD ) . Detection of MCMV gB ( 15A12-H9; [50] ) served as positive control . Female CB6F1 hybrid mice ( BALB/c x C57BL/6 F1; H-2b/d Fv1b/b Fv2r/s Rfv3r/s ) and female BALB/c mice were purchased from Charles River Laboratories ( Sulzfeld , Germany ) . All mice were used when they were between 8 and 9 weeks of age . CB6F1 mice were immunized by intraperitoneal injection of 2 × 105 plaque forming units of the recombinant MCMV vector . If the vaccine was applied repeatedly , the second immunization was applied three weeks later . CB6F1 mice were challenged by the intravenous injection of 5 000 spleen focus-forming units . The development of FV-induced disease was monitored by palpation of the spleens of infected mice twice a week under general anaesthesia , and spleen sizes were rated on a scale ranging from 1 ( normal spleen size ) to 4 ( severe splenomegaly ) as described previously [55] . If mice showed overt signs of severe disease before the end of the experiment as rated by pre-determined termination criteria , they were euthanized and excluded from further analysis . Ten days post challenge ( p . c . ) , plasma samples from CB6F1 mice were obtained , and viremia was determined in a focal infectivity assay [56] . Serial dilutions of plasma were incubated with M . dunni cells for 3 days under standard tissue culture conditions . When cells reached ~100% confluence , they were fixed with ethanol , labeled with F-MuLV Env-specific MAb 720 [49] , and then with a horseradish peroxidase ( HRP ) -conjugated rabbit antimouse Ig antibody ( Dako , Hamburg , Germany ) . The assay was developed using aminoethylcarbazole ( Sigma-Aldrich , Deisenhofen , Germany ) as substrate to detect foci . Foci were counted , and focus-forming units ( FFU ) /ml plasma were calculated . 21 days p . c . , animals were sacrificed by cervical dislocation , the spleens were removed and weighed , and single-cell suspensions were prepared . Serial dilutions of isolated spleen cells were seeded onto M . dunni cells , and cells were incubated under standard tissue culture conditions for 3 days , fixed with ethanol , and stained as described for the viremia assay . Resulting foci were counted , and infectious centers ( IC ) /spleen were calculated . For the analysis of F-MuLV-binding antibodies , MaxiSorp ELISA plates ( Nunc , Roskilde , Denmark ) were coated with whole F-MuLV antigen ( 5μg/ml ) . After coating , plates were blocked with 10% fetal calf serum in PBS , and incubated with serum dilutions . Binding antibodies were detected using a polyclonal rabbit-anti-mouse HRP-coupled anti-IgG antibody and the substrate tetramethylbenzidine ( TMB+; both Dako Deutschland GmbH , Hamburg , Germany ) . Sera were considered positive if the optical density at 450 nm was 3-fold higher than that obtained with sera from naïve mice . To detect F-MuLV-neutralizing antibodies , serial dilutions of heat-inactivated plasma in PBS were mixed with purified F-MuLV and guinea pig complement ( Sigma Aldrich , Munich , Germany ) , incubated at 37°C for 60 min , and then added to M . dunni cells that had been plated at a density of 7 . 5 x 103 cells per well in 24-well plates the day before . Seventy-two hours later cells were stained as described for the viremia assay . Dilutions that resulted in a reduction of foci by 90% or more were considered neutralizing . F-MuLV-specific CD4+ T cells were analyzed in peripheral blood cells , lymph node or spleen cells two weeks after immunization or 21 days p . c . , ; erythrocytes were lysed before the staining when blood samples were used . Cells were stained with an allophycocyanin ( APC ) -coupled major histocompatibility complex ( MHC ) class II tetramer ( containing the I-Ab-restricted F-MuLV Env123-141 epitope EPLTSLTPRCNTAWNRLKL [57]; kindly provided by the MHC Tetramer Core Facility of the National Institutes of Health , National Institute of Allergy and Infectious Diseases , Atlanta , GA ) , fluorescein isothiocyanate ( FITC ) –anti-CD11b , peridinin chlorophyll protein ( PerCP ) –anti-CD43 , Brilliant Violet ( BV ) 510-anti-CD44 , BV605-anti-CD4 ( Becton Dickinson , Heidelberg , Germany ) and Fixable Viability Dye eFluor 780 ( eBioscience , Frankfurt , Germany ) . Data were acquired on an LSR II flow cytometer ( Becton Dickinson , Mountanview , CA ) and analyzed using FlowJo software ( Tree Star , Ashton , OR ) . F-MuLV-specific CD8+ T cells were analyzed in PBMC ten days p . c . ; cells were stained with a PE-coupled MHC I tetramer ( containing the H2-Db-restricted F-MuLV Gag-Leader derived GagL85-93 epitope AbuAbuLAbuLTVFL in which the cysteine residues of the original peptide sequence have been replaced by amino-butyric acid ( Abu ) to prevent disulfide bonding [43]; MBL , Woburn , MA ) , PerCP-anti-CD43 , BV450-anti-CD8 , BV510-anti-CD44 ( Becton Dickinson , Heidelberg , Germany ) and Fixable Viability Dye eFluor 780 ( eBioscience , Frankfurt , Germany ) . Data were acquired on an LSR II flow cytometer ( Becton Dickinson , Mountanview , CA ) and analyzed using FlowJo software ( Tree Star , Ashton , OR ) . For the analysis of vaccine-induced CD8+ T cells , spleens were isolated from mice 6 weeks after MCMV immunization or two weeks after FV infection , and depleted of CD4+ cells using the Miltenyi CD4+ T cell isolation kit ( Miltenyi Biotec , Bergisch Gladbach , Germany ) . Cells were stimulated in an IFNγ ELISpot plate ( 384-well ImmunoSpot , C . T . L . Europe , Bonn , Germany ) with pools of peptides derived from F-MuLV Env ( 6–8 peptides / pool; crude 18-mer peptides with 11 amino acid overlap covering the whole Env sequence were obtained from Peptides&Elephants , Henningsdorf , Germany ) in the presence of 10 units/ml of IL2 , using 10μg/ml of each peptide and 2 . 5 × 105 cells per stimulation . Alternatively , cells were stimulated with a pool of MCMV derived peptides ( M45985-993 ( HGIRNASFI ) , M57816-824 ( SCLEFWQRV ) , M102446-455 ( SIVDLRFAVL ) , m139419-426 ( TVYGFCLV ) and m14115-23 ( VIDAFSRL; [58] ) ) . Cells were stimulated for 48 hours and IFNγ foci were visualized according to the manufacturer’s instructions and counted using a BioReader-7000 Fz ( Bio-Sys , Karben , Germany ) . For the analysis of cytokine production by MCMV-specific CD8+ T cells , peripheral blood cells were restimulated in vitro with the MCMV-derived peptides as indicated above for 6 hours in the presence of 2 μg/ml brefeldin A . Cells were stained with BV421-anti-CD8 ( BioLegend ) , BV510-anti-CD44 ( Becton-Dickinson ) , Fixable Viability Dye eFluor 780 ( FVD-eF780; eBioscience , Frankfurt , Germany ) and FITC-anti-interferon γ ( IFNγ ) ( eBioscience ) . For the analysis of cytokine production by Env-specific CD4+ T cells , peripheral blood cells were stimulated for 40 hours in the presence of 10 units/ml IL2 and a pool of the Env-derived peptides Env123-141 ( EPLTSLTPRCNTAWNRLKL ) , Env57-71 ( ETVWAISGNHPLWTW ) , Env91-105 ( GLEYRAPYSSPPGPP ) , Env415-430 ( KGSYYLVAPAGTMWAC ) , Env267-281 ( PRVPIGPNPVLADQL ) and Env277-291 ( LADQLSFPLPNPLPK ) at a concentration of 10 μg/ml of each peptide , followed by an additional incubation for 6 hours in the presence of 2 μg/ml brefeldin A . Cells were stained with BV605-anti-CD4 ( BioLegend ) , BV510-anti-CD4 , FVD-eF780 , FITC-anti-IL10 ( Invitrogen ) , PE/Dazzle594-anti-IFNγ ( BioLegend ) , PE-Cy7-anti-IL17 ( Invitrogen ) , APC-anti-IL4 ( eBioscience ) , and eFluor450-anti-IL2 ( eBioscience ) . For the analysis of cytokine production by Env-specific CD8+ T cells or CD4+ T cells , spleen cells were stimulated in vitro with pools of peptides derived from F-MuLV Env ( 6–8 peptides / pool; crude 18-mer peptides with 11 amino acid overlap covering the whole Env sequence were obtained from Peptides&Elephants , Henningsdorf , Germany ) , combining three pools per stimulation ( Env Pool3 ) at a concentration of 10 μg/ml per peptide in the presence of 10 units/ml Il2 for 40 hours , followed by an additional incubation for 6 hours in the presence of 2 μg/ml brefeldin A . Cells were stained with BV605-anti-CD4 , BV650-anti-CD44 ( BioLegend ) , AlexaFluor700-anti-CD8 ( eBioscience ) , FVD-eF780 , FITC-anti-IL10 , PE/Dazzle594-anti-IFNγ , PE-Cy7-anti-TNFα ( BioLegend ) , APC-anti-IL4 , and eFluor450-anti-IL2 and BV510-anti-GzmB ( Becton-Dickinson ) . Data were acquired on an LSR II flow cytometer ( Becton-Dickinson , Mountain View , CA ) and analysed using FlowJo software ( Tree Star , Ashton , OR ) . For the transfer of plasma , blood was collected from MCMV . env or MCMV control vector immunized mice 14 days after FV challenge infection , mixed with 10 U/ml heparin and cleared of cells by centrifugation . Plasma collected from one immunized mouse ( ~300 μl ) was injected intravenously into one recipient mouse . For the transfer of CD8+ T cells , spleens were collected from MCMV . env immunized mice 6 weeks after immunization or 14 days after FV challenge infection , single cell suspensions were prepared and CD8+ T cells were isolated from 108 total spleen cells by magnetic cell sorting ( CD8+ T cell untouched isolation kit , Miltenyi , Bergisch-Gladbach , Germany ) . The CD8+ T cells isolated from one mouse were transferred into one recipient mouse by intravenous injection in 200 μl PBS with 50 U/ml heparin . For the depletion of CD4+ or CD8+ T cells , mice were injected on five consecutive days leading up to FV challenge infection with 250 μl of hybridoma-derived antibodies 191 . 1 or 169 . 4 [59] , respectively , followed by injections every other day until day 14 after FV challenge . To control the depletion efficacy , small volumes of blood were collected , stained with antibodies PE-anti-CD4 and BV421-anti-CD8 after erythrocyte lysis , and analyzed by flow cytometry . The depletion efficacy was higher than 98% . For the depletion of B cells , mice were injected once intravenously with 250 μg of the anti-CD20 antibody SA271G2 ( BioLegend ) . Statistical analyses were performed using the software GraphPad Prism 6 ( GraphPad Software , La Jolla , CA ) , testing with the Wilcoxon signed rank test for the comparison of two groups , the ordinary one-way analysis of variance ( ANOVA ) and Holm-Sidak post test or the Kruskal-Wallis one-way analysis of variance on ranks and Student-Newman-Keuls ( equally sized groups ) or Dunn’s ( unequally sized groups ) multiple comparison procedure for the comparison of three or more groups , or with Spearman ranked analysis for the determination of correlation . | CMV-based vectors have attracted a lot of attention in the vaccine development field , since they were shown to induce unconventionally restricted CD8+ T cell responses and strong protection in the SIV rhesus macaque model . In a mouse retrovirus model , we show now that immunization with a mouse CMV-based vector encoding retrovirus envelope conferred very strong protection , even though it was not designed to induce any CD8+ T cell responses . In this MCMV . env immunization , protection relied on the induction of CD4+ T cells and the ability to mount a strong anamnestic neutralizing antibody response upon retrovirus infection , but it was restricted to MCMV pre-naïve mice . In our model system , the MCMV based vector shows very high efficacy that is comparable to an attenuated retrovirus-based vaccine , and encourages the pursuit of this vaccination strategy . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"viral",
"transmission",
"and",
"infection",
"spleen",
"immunology",
"microbiology",
"animal",
"models",
"model",
"organisms",
"experimental",
"organism",
"systems",
"cytotoxic",
"t",
"cells",
"antibodies",
"viral",
"load",
"antibody",
"response",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"animal",
"studies",
"t",
"cells",
"mouse",
"models",
"immune",
"response",
"biochemistry",
"cell",
"biology",
"virology",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
]
| 2019 | Immunization with a murine cytomegalovirus based vector encoding retrovirus envelope confers strong protection from Friend retrovirus challenge infection |
The rhl quorum-sensing ( QS ) system plays critical roles in the pathogenesis of P . aeruginosa . However , the regulatory effects that occur directly upstream of the rhl QS system are poorly understood . Here , we show that deletion of gene encoding for the two-component sensor BfmS leads to the activation of its cognate response regulator BfmR , which in turn directly binds to the promoter and decreases the expression of the rhlR gene that encodes the QS regulator RhlR , causing the inhibition of the rhl QS system . In the absence of bfmS , the Acka-Pta pathway can modulate the regulatory activity of BfmR . In addition , BfmS tunes the expression of 202 genes that comprise 3 . 6% of the P . aeruginosa genome . We further demonstrate that deletion of bfmS causes substantially reduced virulence in lettuce leaf , reduced cytotoxicity , enhanced invasion , and reduced bacterial survival during acute mouse lung infection . Intriguingly , specific missense mutations , which occur naturally in the bfmS gene in P . aeruginosa cystic fibrosis ( CF ) isolates such as DK2 strains and RP73 strain , can produce BfmS variants ( BfmSL181P , BfmSL181P/E376Q , and BfmSR393H ) that no longer repress , but instead activate BfmR . As a result , BfmS variants , but not the wild-type BfmS , inhibit the rhl QS system . This study thus uncovers a previously unexplored signal transduction pathway , BfmS/BfmR/RhlR , for the regulation of rhl QS in P . aeruginosa . We propose that BfmRS TCS may have an important role in the regulation and evolution of P . aeruginosa virulence during chronic infection in CF lungs .
Pseudomonas aeruginosa is an important opportunistic pathogen that accounts for 10% of all hospital-acquired infections [1] , [2] . Most notably , P . aeruginosa is the leading cause of chronic pulmonary infections and mortality in cystic fibrosis ( CF ) patients [3] . The success of P . aeruginosa relies on the production and precise coordination of numerous virulence-associated factors such as lipopolysaccharide , flagella , type IV pili , exopolysaccharide alginate , toxins , proteases , lipases , pyocyanin , and rhamnolipids , which are primarily controlled by regulatory systems such as the quorum-sensing ( QS ) system and the two-component system ( TCS ) [4]–[12] . P . aeruginosa has two well-characterized acyl-homoserine lactone ( acyl-HSL ) - based QS systems , las ( LasR-LasI ) and rhl ( RhlR-RhlI ) [4]–[6] , [8]–[10] . In addition , a third Pseudomonas quinolone signal ( PQS ) acts as a link between the las and rhl QS systems , although PQS is not involved in sensing cell density [4]–[6] , [8]–[10] . The synthase of LasI catalyzes the synthesis of N- ( 3-oxododecanoyl ) homoserine lactone ( 3O-C12-HSL ) , whereas RhlI catalyzes the synthesis of N-butyryl-homoserine lactone ( C4-HSL ) , which induces their respective cognate transcriptional regulators LasR and RhlR , responsible for the activation of numerous QS-controlled genes [4]–[6] , [8]–[10] . The transcriptional regulator LasR is generally considered to sit at the top of the QS hierarchy in P . aeruginosa . LasR/3O-C12-HSL activates the transcription of rhlR , and RhlR/C4-HSL activates the transcription of rhlI and various virulence-associated genes [4]–[6] , [8]–[10] . However , RhlR is able to control the expression of LasR-specific factors independent of LasR [13] . 2- ( 2-hydroxylphenyl ) -thiazole-4-carbaldehyde ( IQS ) could activate the rhl system in a LasR-independent manner [14] . Thus , the regulation of the rhl QS system is much more complicated than previously thought . So far , LasR and Vfr are the two transcriptional regulators known to regulate the expression of rhlR directly , other than RhlR itself [4]–[6] , [8]–[10] , [15] . Pathogenic bacteria , including P . aeruginosa , probes its surrounding environment constantly and makes appropriate decisions during infection [7] , [11] , [12] , [16]–[18] . An important molecular device to achieve sampling of environmental signals is the two-component system ( TCS ) [12] , [16] . Classically , two-component systems are composed of an inner membrane-bound sensor , which is able to detect an environmental stimulus , and a response regulator , which is phosphorylated by the sensor kinase and which , in turn , modulates the expression of genes necessary for the appropriate physiological response [16] . Approximately 130 genes encoding for TCS components have been identified in the genome of P . aeruginosa [1] , [7] , [11] . This provides P . aeruginosa with a sophisticated capability to regulate diverse metabolic adaptations , virulence and antibiotic resistance processes [7] . In fact , a large number of TCSs or TCS components , such as GacSA , PhoPQ , SadARS , RetS , and LadS have been described as having a key role during the infection process [7] , [8] , [11] . However , the direct links between TCS and QS remain poorly understood [6]–[8] , [11] , [19]–[22] . The observation that the expression of BfmRS TCS was dramatically up-regulated in the lungs of cystic fibrosis patients compared to in vitro growth intrigued us [23] . We sought to determine the roles of BfmRS in virulence regulation in P . aeruginosa . In this study , we showed that BfmRS TCS directly controls rhl QS system and modulates the ability of P . aeruginosa to adapt to the host . We demonstrate that BfmRS TCS may play important roles in the regulation and evolution of bacterial virulence during long-term bacterial adaptation to lungs afflicted with cystic fibrosis .
In P . aeruginosa , BfmS ( PA4102 ) is a putative two-component sensor kinase with uncharacterized functions although its cognate response regulator BfmR ( PA4101 ) has been reported to play an important role in biofilm maturation [24] , [25] . To probe the biological roles of BfmS , we generated a bfmS null mutant strain ( ΔbfmS ) as described in the Materials and Methods section and in our previous studies [26] . Interestingly , ΔbfmS strain was defective in green pigment and rhamnolipids ( Figure 1A ) , which can be complemented by introducing the native copy of bfmS ( Table S1 in Text S1 ) into the ΔbfmS strain ( Figure 1A ) . Quantitative analysis of pyocyanin and rhamnolipids indicates that the deletion of bfmS results in a 3 . 5-fold decrease of pyocyanin production and a 5-fold decrease in rhamnolipid production respectively ( Figure S1A and S1B ) . Given that rhamnolipids promote the swarming motility of P . aeruginosa [27] , we next examined the swarming motility of a wild-type MPAO1 strain , a ΔbfmS strain and its complementary strain ( ΔbfmS/p-bfmS ) . As shown in Figure S1C , deletion of bfmS abolished swarming whereas both wild-type MPAO1 strain and the complementary strain swarmed on the surface of plates at 36 h . The rhlAB operon is required for rhamnolipid synthesis [4]–[6] , [9] , [10] . We therefore constructed an rhlA promoter-lux fusion ( rhlA-lux , Table S1 in Text S1 ) and measured its activity in a wild-type MPAO1 strain , a bfmS deletion strain ( ΔbfmS ) , and its complementary strain ( ΔbfmS/p-bfmS ) . The expression of rhlA-lux fusion in ΔbfmS was significantly lower than those of other strains when bacteria were grown in an M8-glutamate minimal medium supplemented with 0 . 2% glucose ( Figure S1D ) . This result suggests that the decreased expression of rhlAB in ΔbfmS strain is likely responsible for the reduction in rhamnolipid production . Since the expression of rhlAB is positively controlled by the rhl quorum-sensing system in P . aeruginosa [4]–[6] , [9] , [10] , we next sought to measure the RhlI-dependent autoinducer C4-HSL content in the wild-type MPAO1 strain , the ΔbfmS strain , and the complementary strain ( ΔbfmS/p-bfmS ) . We used the pDO100 ( pKD-rhlA ) system [28] ( Table S1 in Text S1 ) that carries a lux reporter fused with a rhlA promoter . As a result , supernatants prepared from either the wild-type MPAO1 strain or the complementary strain ( ΔbfmS/p-bfmS ) , but not the ΔbfmS strain , markedly promoted the luminescence values and thereby C4-HSL levels ( Figure 1B ) . We also observed that deletion of bfmS results in decreased rhlI promoter activity ( Figure S1E ) . Based on these results , we conclude that BfmS positively controls rhl QS system in P . aeruginosa . To further study the roles of bfmS , we used microarray analysis in order to compare the transcriptome of the ΔbfmS strain with that of the wild-type MPAO1 strain . As a result , we identified 131 genes with increased transcript levels ( ≥2-fold ) ( Table S2 in Text S1 ) and 71 genes with decreased transcript levels ( ≤2-fold ) ( Table S3 in Text S1 ) in the ΔbfmS strain versus wild-type MPAO1 strain . These 202 genes represent ∼3 . 6% of the total number of annotated genes in the P . aeruginosa PAO1 genome . Of those 202 genes , 42% encode hypothetical proteins of unknown functions ( Tables S2 and S3 in Text S1 ) . Grouping these genes according to their annotated function shows that they belong to several functional categories , primarily transport of small molecules , carbon compound catabolism , translation , and adaptation ( Tables S2 and S3 in Text S1 ) . Among the 131 genes whose expression is up-regulated in the ΔbfmS strain , 7 genes were up-regulated more than 10-fold . Interestingly , these 7 genes , including PA4100 , bfmR , PA4103 , PA4104 , PA4105 , PA4106 , and PA4107 , are located at or near the bfmRS ( PA4101-PA4102 ) loci ( Figure S2A , Table S2 in Text S1 ) . These microarray-based expression data are consistent with the operon predictions for P . aeruginosa , which suggested that PA4103 and PA4104 are organized into PA4103 operon ( PA4103-PA4104 ) while PA4105 , PA4106 and PA4107 are organized into PA4107 operon ( PA4107-PA4106-PA4105 ) ( www . pseudomonas . com ) . Among these genes , PA4100 encodes a dehydrogenase of unknown function , and bfmR encodes a two-component response regulator that acts as a biofilm maturation regulator , whereas PA4103 , PA4104 , PA4105 , PA4106 , and PA4107 encode hypothetical proteins . Although PA4103 contains a ferric reductase like transmembrane component ( pfam01794 ) and PA4107 contains a calcium binding motif ( cd00051 ) , their biological functions are unknown . Further characterization of the functions of these genes may provide insight into the roles of bfmS in P . aeruginosa . There are 11 genes whose expressions are down-regulated more than 10-fold in ΔbfmS strain ( Figure S2 , Table S3 in Text S1 ) . Among them , 5 genes ( rhlA , rhlB , antA , antB , and antC ) are already known to be controlled by the rhl QS system [29] , [30] . In addition , we observed that deletion of bfmS decreases the transcription of rhlI by approximately 66% ( Table S3 in Text S1 ) . We also found a moderate , 20% decrease in rhlR transcription in the ΔbfmS mutant compared with the parent , which is consistent with the results of an rhlR-lux reporter gene analysis ( Figure S3A ) . These results further suggest that bfmS positively controls the rhl QS . To further confirm the differentially expressed genes identified by the microarray analysis , 12 genes representing a range of microarray signal intensity and expression profiles were subjected to real-time ( RT ) PCR analyses . There was a high degree of consistency among data generated by these two methods ( Table S4 in Text S1 ) , which assures the reliability of microarray analysis in determining transcriptional changes . Since deletion of bfmS led to the over-expression of bfmR ( >90-fold ) ( Figure S2A , Table S2 in Text S1 ) , we next sought to determine if the elevated bfmR contributes to the phenotypes observed in the ΔbfmS strain . We generated a bfmRS double mutant strain ( ΔbfmRS ) ( Table S1 in Text S1 ) and performed phenotypic analysis . Interestingly , the ΔbfmRS strain and wild-type MPAO1 strain display similar phenotypes when bacteria are grown in Pyocyanin production broth ( PPB ) or on a cetyltrimethylammonium bromide ( CTAB plate ) ( Figure 1C ) . The introduction of a wild-type bfmR gene ( p-bfmR , Table S1 in Text S1 ) into ΔbfmRS strain restored its phenotypes similar to ΔbfmS strain ( Figure 1C ) . These results suggest that the effect of bfmS deletion on either the pigment production or rhamnolipids production in P . aeruginosa is likely mediated through the over-expression of bfmR . Moreover , qRT-PCR analysis indicates that the expressions of at least 12 genes were significantly affected by the deletion of bfmS , whereas their altered expression levels caused by bfmS deletion were suppressed by additional deletion of bfmR . Thus , bfmR may mediate most , if not all , the output of bfmS . We next tested if BfmS could affect the expression of BfmR . We constructed a bfmR promoter-lux fusion ( bfmR-lux , Table S1 in Text S1 ) and then measured its activity in a wild-type MPAO1 strain , a bfmS deletion strain ( ΔbfmS ) , its complementary strain ( ΔbfmS/p-bfmS ) , a bfmRS double deletion strain ( ΔbfmRS ) , and a ΔbfmRS strain harboring p-bfmR , p-bfmS or p-bfmRS ( Table S1 in Text S1 ) . As shown in Figure 1D , the activity of bfmR-lux in ΔbfmS strain was about 60-fold higher than that of the wild-type MPAO1 strain . Complementation with p-bfmS in the ΔbfmS strain restored the activity of bfmR-lux to levels similar to the wild-type strain ( Figure 1D ) . In addition , the activity of bfmR-lux in the ΔbfmRS strain was similar to that observed in the wild-type MPAO1 strain; however , the introduction of p-bfmR , but not p-bfmS or p-bfmRS , to the ΔbfmRS strain dramatically increased the activity of bfmR-lux ( >46-fold ) ( Figure 1D ) . Hence , BfmR can activate its own gene promoter in the absence of BfmS . We next evaluated if the absence of BfmS causes an accumulation of BfmR in P . aeruginosa . To this end , we constructed an integration vector mini-ctx-BfmR-Flag ( Table S1 in Text S1 ) and the resulting clone was mobilized into the wild-type MPAO1 and ΔbfmRS strain , yielding an MPAO1::BfmR-Flag strain and ΔbfmRS::BfmR-Flag strain , respectively . The ΔbfmRS::BfmR-Flag strain displayed a pigment-deficient phenotype as observed for either the ΔbfmS strain or the ΔbfmRS/p-bfmR strain ( Figure S4A ) . The cell lysates of MPAO1::BfmR-Flag strain , ΔbfmRS::BfmR-Flag strain and its complementary strain ( ΔbfmRS::BfmR-Flag/p-bfmS ) were subjected to Western blot analysis using anti-FLAG antibodies . As shown in Figure S4B , a large amount of BfmR-Flag in the ΔbfmRS::BfmR-Flag strain was detected . In contrast , no detectable signal was obtained for the BfmR-Flag generated from the MPAO1::BfmR-Flag strain or the complementary strain ( ΔbfmRS::BfmR-Flag/p-bfmS ) ( Figure S4B ) . Therefore , the absence of BfmS leads to an accumulation of its cognate response regulator BfmR . Since BfmR activates its own gene promoter , we next aimed to test if BfmR could bind its own promoter . We performed electrophoretic mobility shift assay ( EMSA ) using 6His-BfmR protein and DNA fragments containing bfmR , rhlA , and rhlC promoter regions , respectively . We found that 6His-BfmR could shift the bfmR promoter DNA , although it failed to shift the rhlA or rhlC promoter DNA ( Figure 2A ) . We further determined the specific DNA sequence that BfmR can recognize in the bfmR promoter region by using a dye-primer-based DNase I footprint assay . We uncovered three BfmR-protected regions in the bfmR promoter DNA ( Figure 2B ) . Interestingly , all three BfmR-protected regions contain a consensus sequence GATACAnnGC ( where n is any nucleotides , Figure 2C ) . Using RAST ( http://rsat . ulb . ac . be/rsat/ ) , we found that 41 promoters ( −1 bp to −400 bp of the coding region ) of P . aeruginosa PAO1 , including the PA4017 promoter , contain a putative BfmR-binding motif ( GATACAnnGC ) ( Table S5 in Text S1 ) . As expected , BfmR could shift the PA4107 promoter DNA ( Figure S5A ) in our EMSA analysis , although it failed to shift the rhlI promoter DNA ( Figure S5A ) . Interestingly , we also observed that BfmR is able to bind to PA4103 promoter DNA ( Figure S5B ) that lacks a canonical BfmR-binding motif ( GATACAnnGC ) ( Table S5 in Text S1 ) . Using a dye-primer-based DNase I footprint assay , we uncovered that the BfmR-protected region of PA4103 promoter DNA contains a non-canonical BfmR-binding motif ( GATACAnnAC , the mismatch is underlined ) ( Figure S5C ) , which is subsequently determined to be required for the BfmS-mediated regulation of PA4103 promoter activity ( Figure S5D ) . Thus , it is likely that BfmR directly control the expression of the PA4107 and PA4103 operon . As aforementioned , BfmR negatively controls the rhl QS system of P . aeruginosa , which does not need to bind promoter of rhlI ( Figure S5A and S5B ) . These observations prompted us to determine if BfmR binds to the promoter of rhlR . EMSA analysis indicated that 6His-BfmR bound to the rhlR promoter DNA but not to the promoter region of rhlC that serves as a negative control ( Figure 3A ) . Dye-primer-based DNase I footprint assay indicated that there were three BfmR-protected regions in the promoter of the rhlR ( Figure 3B ) . Interestingly , protected region I ( −220 to −193 from the start codon of rhlR ) harbored a putative BfmR-binding motif ( GATACTnnGC ) with one mismatch ( underlined ) ( Figure 3B ) , oriented in the opposite direction of the transcription of rhlR . Protected region II extends from nucleotide −151 to nucleotide −171 while protected region III extends from nucleotide −69 to nucleotide −85 , relative to the start codon of rhlR , respectively ( Figure 3B ) . There are 44 bfmS-regulated genes harbor a consensus sequence ( GATACAnnGC with or without one mismatch ) in their promoter region ( −1 bp to −400 bp of the coding region ) ( Tables S2 and S3 in Text S1 ) . Additionally , there are 984 promoters ( −1 bp to −400 bp of the coding region ) harbor the consensus sequence ( GATACAnnGC without or with one mismatch ) in the PAO1 genome . These observations suggest that BfmR may serves as a global regulator affecting expression of a large number of genes . We next elucidated if BfmR regulates the expression of rhlR . To do this , we measured rhlR promoter-lux fusion activity in the wild-type MPAO1 strain , the ΔbfmS strain , the complementary strain ( ΔbfmS/p-bfmS ) , the ΔbfmRS strain , and the ΔbfmRS strain harboring p-bfmR ( ΔbfmRS/p-bfmR ) . The low-phosphate ( 0 . 32 mM Pi ) M8-glutamate minimal medium supplemented with 0 . 2% glucose , used as phosphate limitation , served to stimulate the expression of rhlR [21] . As shown in Figure 3C and S3B , the activity of rhlR-lux in the ΔbfmS strain was more than 8-fold lower than that observed in the wild-type MPAO1 strain . Complementation with p-bfmS in the ΔbfmS strain restored the activity of rhlR-lux similar to that of the wild-type strain ( Figure 3C ) . Moreover , ΔbfmRS strain exhibited rhlR-lux activity similar to that observed in the wild-type MPAO1 strain , while the introduction of p-bfmR into the ΔbfmRS strain caused a 3 . 8-fold decrease in rhlR-lux activity ( Figure 3C ) . Thus , it is likely that bfmS activates the expression of rhlR by repressing BfmR , which acts to negatively regulate rhlR expression and rhl QS . This notion was further substantiated by the observations that under low-phosphate growth conditions ΔbfmS also exhibits decreased rhlI ( Figure S3C ) and rhlA ( Figure S3D ) promoter activity , and lowered C4-HSL content ( Figure S3E ) as compared to wild-type MPAO1 . To determine if the putative BfmR-binding motif ( GATACTnnGC ) ( Figure 3B ) is involved in the BfmR-mediated inhibition of rhlR-lux activity , we deleted the first five residues ( GATACT ) in the consensus sequence ( yielding rhlR-D-lux , Table S1 in Text S1 ) , and examined the ability of the mutant sequence to permit the inhibition of the reporter gene in ΔbfmS strain . As shown in Figure 3D , the rhlR-lux activity in ΔbfmS strain was approximately 8-fold lower than that observed in the wild-type MPAO1 strain or in the ΔbfmRS strain . However , the rhlR-D-lux activity was about 2-fold lower in the ΔbfmS strain compared to the wild-type MPAO1 strain or the ΔbfmRS strain ( Figure 3D ) . Thus , the five residues ( GATACT ) are required for the full inhibition of rhlR-lux activity in ΔbfmS strain , demonstrating the importance of these conserved binding-site residues . However , besides GATACT sequence elements , additional regulatory sequence elements within the promoter region of rhlR are most likely involved in BfmR-mediated inhibition of rhlR-lux activity , given that the rhlR-D-lux activity in ΔbfmS strain is still decreased , although to a much lesser extent than that of rhlR-lux ( Figure 3D ) . Like many other response regulators [31] , BfmR can be phosphorylated by acetyl phosphate ( Figure S6A ) and hence activated in vitro ( Figure S6B ) . We further observed that the predicted phosphorylation site , aspartate residue D55 , is required for the activation of BfmR in vitro and in vivo ( Figure S7 ) . As acetyl∼P is an intermediate in the acetate kinase ( AckA ) -phosphate acetyltransferase ( Pta ) pathway [31] , we hypothesized that the AckA-Pta pathway may be involved in the activation of BfmR . Thus , we constructed a mutant strain ( ΔbfmSΔackA-pta , Table S1 in Text S1 ) with deletion of both the bfmS gene and ackA-pta operon and measured the activity of bfmR-lux as well as the C4-HSL content in this strain and the ΔbfmS strain , respectively . As shown in Figure 4 , the expression of bfmR-lux was lower ( Figure 4A ) while the C4-HSL content was higher ( Figure 4B ) in the ΔbfmSΔackA-pta strain than that of the ΔbfmS strain , respectively . The introduction of wild-type ackA-pta operon ( p-ackA-pta , Table S1 in Text S1 ) into the ΔbfmSΔackA-pta strain was able to restore either the activity of bfmR-lux ( Figure 4A ) or the C4-HSL content to the level of the ΔbfmS strain ( Figure 4B ) . Therefore , in the ΔbfmS strain , acetyl phosphate or the component that is dependent on the Acka-Pta pathway is required for the full activity of BfmR . As BfmS modulates the rhl QS system that contributes significantly to the virulence of P . aeruginosa [4]–[6] , [9] , [10] , we infected romaine lettuce leaves with P . aeruginosa to see if BfmS controls bacterial virulence . The pathogenicity assay revealed a significant difference in the manifestation of infection symptoms caused by the ΔbfmS strain compared to wild-type MPAO1 strain . Relative to wild-type MPAO1 , the ΔbfmS strain failed to cause severe necrotic lesions of the leaves , which can be complemented by introducing the wild-type bfmS gene into the ΔbfmS strain ( Figure 5A ) . In addition , the ΔbfmRS strain exhibited a virulence phenotype similar to that of a wild-type MPAO1 strain , while the introduction of p-bfmR into the ΔbfmRS strain led to a low virulence phenotype ( Figure S8 ) . Moreover , the constitutive expression of rhlR in ΔbfmS strain could restore either the virulence of ΔbfmS strain to the level of the wild-type MPAO1 strain ( Figure 5B ) , suggesting that the decreased expression of rhlR is likely responsible for the attenuated virulence of ΔbfmS strain in the lettuce leaf model of P . aeruginosa infection . Since cytotoxicity and invasion of P . aeruginosa are useful traits for this pathogen [32] , we further characterized BfmS to check if it regulates the cytotoxicity or the invasion of P . aeruginosa in a murine lung epithelial cell line ( MLE-12 ) , a widely used in vitro model for studying host-pathogen interactions [33]–[35] . Using an MTT assay , we found that about 50% MLE-12 cells were killed when challenged with wild-type MPAO1 strain; however , only 5% of MLE-12 cells were killed after inoculation with the ΔbfmS strain ( Figure 5C ) . Using a colony forming unit ( CFU ) assay , we showed that deletion of bfmS significantly increases ( p<0 . 01 ) the internalization of P . aeruginosa by approximately 50% ( Figure 5D ) . Further , the invasive and cytotoxic phenotypes of ΔbfmS strain could be completely restored to the wild-type levels by the introduction of p-bfmS ( Figure 5C and 5D ) . Thus , deletion of bfmS causes a loss of cytotoxic capacity while it enhances the invasion of P . aeruginosa MPAO1 to MLE-12 cells . To further determine the virulence regulated by BfmS , a mouse model of acute pneumonia was used as described in our previous study [26] . C57BL/6J mice were intranasally infected with approximately 5×106 CFU of wild-type MPAO1 , bfmS null mutant ΔbfmS , and its complementary strain ΔbfmS/p-bfmS . Figure 5E shows the CFU of bacteria recovered from the lungs compared to the initial inoculum at 18 h post infection , with a geometric mean indicated for each group . In this assay , wild-type MPAO1 was recovered in numbers approximately at 3 . 13% of the inoculum dose from lungs with a result 5 . 6-fold higher than that ( 0 . 56% ) of the ΔbfmS strain . Further , bacteria of complementary strain ( ΔbfmS/p-bfmS ) were recovered from lungs with 3 . 43% , similar to that of the wild-type MPAO1 strain ( Figure 5E ) . These results indicate that deletion of bfmS decreases P . aeruginosa survival in the mouse lungs in this model and thus reduced virulence . The P . aeruginosa DK2 lineage is highly successful and has been isolated from ∼40 cystic fibrosis ( CF ) patients since the start of the sampling program in 1973 [36] . We noted that the DK2 lineage-specific mutations in BfmS are point mutations that cause two amino acid substitutions , proline replaces leucine 181 ( L181P ) , and glutamine replaces glutamic acid ( E376Q ) . Among them , L181P was fixed in the DK2 lineage before the year 1979 , while E376Q was subsequently fixed in the DK2 lineage after 1991 [36] . We next investigated the regulatory effect associated with these amino acid substitutions observed in the BfmS . We created p-bfmSL181P , p-bfmSE376Q , and p-bfmSL181P/E376Q plasmids ( Table S1 in Text S1 ) and introduced them to the ΔbfmS strain , and tested their effects on bfmR-lux activity . Interestingly , L181P and E376Q substitutions in BfmS caused a 16-fold and a 2-fold increase in the activity of bfmR-lux , respectively ( Figure 6A ) . More significantly , the combined substitution ( L181P/E376Q ) led to a 27-fold increase in the activity of bfmR-lux ( Figure 6A ) . Accordingly , L181P or L181P/E376Q substitutions in BfmS decreased the RhlI-dependent autoinducer C4-HSL content ( Figure 6B ) . These data strongly suggest that mutations in specific residues of BfmS result in activation of BfmR . The DK2 lineage-specific mutations ( L181P , L181P/E376Q ) may abolish the negative regulatory effects of BfmS on BfmR or alternatively , it may transform BfmS into a positive regulator of BfmR and therefore , activate BfmR ( Figure 6A and 6B ) . To discriminate between these two possibilities , we tested the effect of these amino acid substitutions in BfmS on the activity of bfmR-lux when bacteria were grown in M8-glutamate minimal medium supplemented with sodium acetate . BfmS did not function as a negative regulator of BfmR when bacteria were grown in this media , given that the ΔbfmS strain exhibited similar bfmR-lux activity as the MPAO1 strain ( Figure 6C ) . However , the ΔbfmS/p-bfmSL181P strain and ΔbfmS/p-bfmSL181P/E376Q strain displayed bfmR-lux activity 2 . 5-fold and 6-fold higher than that the activity observed in the reference strain ( ΔbfmS/p-bfmS ) , respectively ( Figure 6C ) . These results suggest that the L181P or L181P/E376Q amino acid substitution render BfmS as a positive regulator of BfmR . Further , the introduction of p-bfmSL181P or p-bfmSL181P/E376Q to the bfmRS double deletion mutant strain ( ΔbfmRS ) failed to increase the activity of bfmR-lux ( Figure S9A ) , suggesting that the effect of L181P or L181P/E376Q substitutions in BfmS on the bfmR-lux activity is mediated through the activation of BfmR . BfmS is a member of the HisKA subfamily of bacterial histidine kinases and it is predicted that the conserved H238 residue is required for its kinase activity [37] . We next investigated the regulatory role associated with this residue by changing His to Ala . We created p-bfmSH238A and p-bfmSL181P/E376Q/H238A plasmids ( Table S1 in Text S1 ) and introduced them to the ΔbfmS strain , and examined their effects on the bfmR-lux activity . As shown in Figure 6D , H238A substitution in BfmS has no significant effect on the activity of bfmR-lux when bacteria were grown in M8-glutamate minimal medium supplemented with 0 . 2% glucose . However , H238A substitutions in the BfmSL181P/E376Q abolished its ability to induce the bfmR-lux activity ( Figure 6D ) , which suggests that the H238 residue or the kinase activity is required for BfmSL181P/E376Q to activate BfmR . This hypothesis is further supported by the fact that amino acid substitution L181P/E376Q , but not L181P/E376Q/H238A , causes overproduction of phosphorylated and unphosphorylated BfmR ( Figure 6E ) . Furthermore , we noted that the RP73-specific mutation in bfmS causes the substitution of arginine by histidine at the codon 393 ( R393H ) [38] . The P . aeruginosa strain RP73 was isolated after 16 . 9 years of chronic lung infection in a CF patient [38] . Like the L181P substitution or the L181P/E376Q substitution , the R393H substitution in BfmS resulted in increased bfmR-lux activity ( Figure S9B and S9C ) and decreased level of C4-HSL ( Figure S9D ) , suggesting that BfmSR393H activates BfmR . As the regulatory effects of these BfmS variants ( BfmSL181P , BfmSL181P/E376Q , and BfmSR393H ) on BfmR have been changed from negative to positive , we therefore term them “reverse function” mutants . Although we failed to obtain either the soluble full-length BfmS protein or the cytoplasmic region of BfmS that prevented us from the assays of the phosphatase/kinase activities of BfmS in vitro , our genetic analyses clearly indicate that specific missense mutations ( L181P , L181P/E376Q , or R393H ) can convert BfmS from a repressor to an activator of BfmR .
In this study , we uncovered a novel signal transduction pathway , BfmS/BfmR/RhlR , for the regulation of the rhl QS system in P . aeruginosa . We demonstrated that BfmS has profound effects on the expression of virulence-associated traits and the ability of P . aeruginosa to adapt to the host . In addition , we found that deletion of bfmS leads to a dramatic increase in biofilm formation and that BfmR mediates this effect ( Figure S10 ) . This appearance is consistent with previous observations that BfmR is a biofilm maturation regulator [24] , [25] . Intriguingly , BfmS is able to switch its function from a repressor to an activator of BfmR in cystic fibrosis ( CF ) isolates such as DK2 strains and RP73 strain . A proposed model for signal transduction by BfmRS TCS is shown in Figure 7 . The prototypical two-component regulatory system is composed of a sensor kinase and a response regulator . In general , the sensor kinase senses an environment change and communicates it via phosphorylation to its cognate response regulator , and hence activates the response regulator's function . We demonstrated that BfmS functions as a negative regulator of BfmR ( Figure 1D and Figure S4 ) , whose activation requires phosphorylation on D55 ( Figure S7 ) . Therefore , BfmS might act as a phosphatase as opposed to a kinase for BfmR under our experimental conditions . In fact , many two-component sensors are bifunctional , catalyzing both the phosphorylation and dephosphorylation of their cognate response regulator [37] , [39]–[41] . For some sensor kinases , the phosphatase activity may be the critical function in vivo [42]–[45] . Interestingly , the BfmS homologue in Pseudomonas syringae , RhpS , has also been shown to be a negative regulator of its cognate regulator RhpR in our previous studies [44] , [45] . In the absence of BfmS , the Acka-Pta pathway can modulate the activity of BfmR ( Figure 4 ) . These observations suggest that the activation of BfmR is shaped by BfmS as well as by the nutritional status of P . aeruginosa . However , the expression of bfmR-lux in the ΔbfmSΔackA-pta strain was still about 12-fold higher than that in the wild-type MPAO1 strain ( Figure 4A ) , indicating that acetyl phosphate or the component that is dependent on the Acka-Pta pathway is likely not the sole trigger of BfmR activation . This notion was further substantiated by the observation that the deletion of ackA-pta operon only partially alleviates the inhibition of QS signal C4-HSL production caused by the absence of bfmS ( Figure 4B ) . The rhlR gene encodes the transcriptional regulator RhlR , which has a central role in the quorum-sensing response , and is therefore very important for P . aeruginosa to co-ordinate its virulence in order to establish a successful infection [4] , [6] , [8]–[10] , [46] , [47] . We found that BfmR binds to and represses the rhlR promoter ( Figure 3 ) . A DNase I footprint analysis demonstrated that the BfmR-protected region ( binding site I ) of the rhlR promoter has a putative BfmR-binding motif ( GATACTnnGC ) that is crucial to the BfmR-mediated inhibition of rhlR-lux activity ( Figure 3D ) , thereby reinforcing the likelihood that BfmR directly regulates rhlR . However , the detailed effect of BfmR on the rhlR gene expression awaits further study , since the rhlR promoter harbors multiple transcription start sites [15] , regulatory sequences [15] , and at least three BfmR-protected regions ( Figure 3B ) . To our knowledge , this is the first evidence of a two-component regulator regulating rhlR in a direct manner . However , this finding is in contrast to a previous report suggesting that BfmR functions independently of QS signaling [24] . The exact cause of this discrepancy remains unknown . In the previous report [24] , Petrova et al drew the conclusion based on the fact that deletion of bfmR has no significant effect on the transcript abundance of rhlA and lasB . Consistent with this , we observed that deletion of bfmRS has no significant effect on rhl-dependent phenotypes ( Figure 1A and 1B ) and the expression of rhlA ( Table S4 in Text S1 ) . However , deletion of bfmS alone leads to the activation of BfmR ( Figure 1D and S4B ) , which in turn directly binds to the promoter and decreases the expression of the rhlR ( Figure 3 ) , causing the inhibition of the rhl QS system ( Figure 1 and S3 ) . Additionally , the decreased expression of rhlR in ΔbfmS strain ( Figure 3C and 3D ) may contribute to the attenuated virulence in lettuce leaves ( Figure 5A ) and the reduced production of QS signal C4-HSL , pyocyanin and rhamnolipids ( Figure 1 and S1 ) , since the constitutive expression of rhlR in ΔbfmS strain could restore these phenotypes to wild-type levels or higher ( Figure 5B and S11 ) . Moreover , the expression of rhlI in the ΔbfmS strain was significantly lower ( >2-fold ) than that of the wild-type strain ( Tables S3 and S4 in Text S1 ) . Thus , the bfmS deletion , which results in activation of BfmR , affects all aspects of the rhl QS system . Deletion of bfmS impacts the expression of 202 genes that comprise 3 . 6% of the P . aeruginosa genome ( Tables S2 and S3 in Text S1 ) . These observations suggest that BfmS acts as a global regulator in P . aeruginosa . Besides regulating rhl quorum sensing , BfmS also regulates the expression of a large number of genes such as PA4103 , PA4104 , PA4105 , PA4106 and PA4107 , whose transcripts are likely independent of the quorum-sensing regulated [48] . Thus , it is not surprising that BfmS has a profound effect on the expression of virulence-associated traits and the ability of P . aeruginosa to adapt to the host ( Figure 5 ) . Interestingly , BfmS has a positive impact on acute virulence phenotypes ( Figure 5A and 5E ) , but a negative effect on biofilm formation ( Figure S10 ) that acts as a major virulence-associated trait contributing to chronic infections [49] . This formation suggests that BfmS may play an important role in mediating the switch between the acute and chronic infection lifestyles of P . aeruginosa . P . aeruginosa can cause serious acute and chronic infections in humans and it underwent numerous genetic adaptations during evolution in the CF airways , resulting in remodeling of the regulatory networks to match the fluctuations in the environment of CF lung [36] , [50]–[52] . bfmS in P . aeruginosa CF isolates are often found to undergo missense mutations . For instance , L181P ( point mutations causing the substitution of leucine by proline at codon 181 ) or L181P/E376Q in at least 10 DK2 strains [36] , R393H in RP73 strain [38] , A21P/T120K/L164F in either LESB58 or LES431strain , L164F in c7447m strain , A4T in CIG1 strain , T120K/L163V/L164F in C3719 strain , D295N in CF5 strain , and P6S/L164F in CF614 strain ( http://www . ncbi . nlm . nih . gov/ ) . We found that specific missense mutations in bfmS gene ( L181P , L181P/E376Q , and R393H ) result in elevated BfmR activity ( Figure 6 and Figure S9 ) , which contributes to biofilm formation ( Figure S10 ) [24] , [25] as well as to the inhibition of the rhl QS system ( Figures 1 , 3 , and S3 ) . It is well known that biofilm formation enables P . aeruginosa to cause persistent infections [49] while the loss of quorum sensing is one of the dominating changes that occur during the adaptive process of the P . aeruginosa in the CF lung [36] , [50] . Thus , the naturally occurring missense mutations in BfmS may provide a selective advantage to either DK2 strains or RP73 strain during the course of chronic infection in CF lungs . However , it should be noted that the P . aeruginosa community in the CF lung is very dynamic [50] , [51] , [53] and only a fraction of the isolates will most probably possess these mutations . Therefore , the role of these missense mutations in the chronic lung infection awaits further investigation . Intriguingly , although BfmS functions as a negative regulator of BfmR ( Figure 1D and S4B ) , the naturally occurring missense mutations in bfmS gene ( L181P , L181P/E376Q , and R393H ) can produce BfmS variants that no longer repress , but instead activate BfmR ( Figure 6 and Figure S9 ) . These “reverse function” mutants of BfmS may exhibit an elevated ratio of kinase to phosphatase activity on BfmR , given that the activation of BfmR requires the phosphorylation on D55 ( Figure S7 ) . In agreement with this notion , we found that H238 , the conserved histidine residues predicted to be involved in the autophosphorylation of BfmS , is required for BfmSL181P/E376Q to activate BfmR ( Figure 6D and 6E ) . The occurrence of BfmS “reverse function” mutants is not strain dependent , as evidenced by the fact that either DK2 lineage-specific mutations or RP73-specific mutation in bfmS could reverse its function against BfmR . The L181P mutation was located in the HAMP domain of BfmS , while the E376Q and the R393H mutation were located in ATP-binding domain . Currently , it is not clear how these missense mutations change the function of BfmS . Although further studies are needed to elucidate this elegant mechanism , our genetic analyses clearly indicated that naturally occurring missense mutations in P . aeruginosa gene could result in reverse of function , rather than simply loss ( weakened ) or gain ( strengthened ) of function . Noticeably , bfmRS operon and BfmR-activated transcripts such as PA4103-4104 and PA4107-4106-4105 ( Figure S2 , Tables S2 and S4 in Text S1 ) , were dramatically up-regulated in the lungs of cystic fibrosis patients compared to in vitro planktonic bacteria , indicating that BfmRS system is likely activated during chronic infection in CF lungs [23] . These genes also exhibit much higher gene expression levels in some P . aeruginosa CF isolates such as DK2-lineage strains ( late stage infection isolates ) [54] and E601 strain [55] compared to wild-type laboratory strain PAO1 . These observations and results from the current study suggest that BfmRS TCS may sense and respond to environmental stress in CF lungs . We envision that further studies aimed at the characterization of the stimuli that BfmRS and/or its variants detect within the host could be of great importance to a full understanding of the mechanisms that make P . aeruginosa a successful pathogen and to the development of novel strategies to limit its infections .
Animal experiments were performed in strict accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals approved by the State Council of People's Republic of China ( 11-14-1988 ) . All animal procedures were approved by the Institutional Animal Care and Use Committee ( IACUC ) of Shanghai Public Health Clinical Center ( Permit Number: 2013P201 ) . The laboratory animal usage license number is SYXK-HU-2010-0098 , certificated by Shanghai Committee of Science and Technology . The bacterial strains and plasmids used in this study are listed in Table S1 in Text S1 . Unless noted otherwise , P . aeruginosa MPAO1 [56] and its derivatives were grown in Luria-Bertani ( LB ) broth , Pyocyanin production broth [57] ( PPB: 20 g peptone , 1 . 4 g MgCl2 , 10 g K2SO4 , 20 ml glycerol per liter; pH 7 . 0 ) , or M8-glutamate minimal medium [27] ( 6 g Na2HPO4 , 3 g KH2PO4 , 0 . 5 g NaCl , 0 . 24 g MgSO4 , 0 . 5 g glutamate per liter; pH 7 . 4 ) supplemented with 0 . 2% glucose , as indicated . E . coli cultures were grown in Luria-Bertani ( LB ) broth . All cultures were incubated at 37°C with shaking ( 250 rpm ) . For plasmid maintenance in P . aeruginosa , the medium was supplemented with 100 µg/ml carbenicillin or 100 µg/ml kanamycin when required . For plasmid maintenance in E . coli , the medium was supplemented with 100 µg/ml carbenicillin , 50 µg/ml kanamycin , 300 µg/ml trimethoprim , or 10 µg/ml gentamicin , as appropriate . For marker selection in P . aeruginosa , either 30 µg/ml gentamicin or 10 µg/ml tetracycline were used when required . For gene replacement , a SacB-based strategy [58] was employed as described in our previous study [26] . To construct the bfmS null mutant ( ΔbfmS ) , polymerase chain reactions ( PCRs ) were performed in order to amplify sequences upstream ( 1 , 574 bp ) and downstream ( 1 , 562 bp ) of the intended deletion . The upstream fragment was amplified from MPAO1 genomic DNA using primers BfmSupF ( with EcoRI site ) and BfmSupR ( with XbaI site ) , while the downstream fragment was amplified with primers , BfmSdownF ( with XbaI site ) and BfmSdownR ( with HindIII site ) . The two PCR products were digested and then cloned into the EcoRI/HindIII-digested gene replacement vector pEX18Ap , yielding pEX18Ap::bfmSUD . A 1 . 8 kb gentamicin resistance cassette was cut from pPS858 with XbaI and then cloned into pEX18Ap::bfmSUD , yielding pEX18Ap::bfmSUGD . The resultant plasmid , pEX18Ap::bfmSUGD , was electroporated into MPAO1 with selection for gentamicin resistance . Colonies were screened for gentamicin sensitivity and loss of sucrose ( 5% ) sensitivity , which typically indicates a double-cross-over event and thus marks the occurrence of gene replacement . The ΔbfmS strain was further confirmed by PCR . A similar strategy was used to construct the ΔbfmRS strain as described above . Briefly , the upstream fragment ( 1 , 832 bp ) of the intended deletion was amplified with primers BfmRupF ( with EcoRI site ) and BfmRupR ( with XbaI site ) . The downstream fragment ( 1 , 562 bp ) was amplified with primers , BfmSdownF ( with XbaI site ) and BfmSdownR ( with HindIII site ) . A 1 . 8 kb gentamicin resistance cassette was cut from pPS858 with XbaI and then cloned into pEX18Ap::bfmRSUD , yielding pEX18Ap::bfmRSUGD . Again , a similar strategy was used to construct the ΔbfmSΔackA-pta strain . Primers Acka-up-F ( with KpnI site ) and Acka-up-R ( with BamHI site ) amplified the upstream fragment ( 2 , 245 bp ) of the intended deletion of ackA-pta operon in ΔbfmS . Primers Pta-domn-F ( with BamHI site ) and Pta-domn-R ( with HindIII site ) amplified the downstream fragment ( 1 , 709 bp ) . A 2 . 3 kb tetracycline resistance cassette was amplified from the integration vector mini-CTX-lacZ with primers , Mini-TC-F ( with BamHI site ) and Mini-TC-F ( with BamHI site ) . The resultant plasmid , pEX18Ap::acka-ptaUTD , was electroporated into ΔbfmS strain with selection for tetracycline resistance . Colonies were screened for tetracycline sensitivity and loss of sucrose ( 5% ) sensitivity , which typically indicate a double-cross-over event and thus mark the occurrence of gene replacement . PCR further confirmed the deletion of pta-acka loci . Primers bfmRflag-F ( with HindIII site ) and bfmRflag-R ( with BamHI site ) ( Table S6 in Text S1 ) were used to perform PCR of the BfmR gene that was meant to fuse with the Flag-tag . A 1 , 586-bp PCR product covering the region from 848 bp upstream and the BfmR gene ( not including the stop codon ) was generated . The HindIII- and BamHI-digested PCR product was cloned into the HindIII and BamHI sites of the mini-CTX-lacZ [59] to generate mini-ctx-BfmR-Flag . The resulting plasmid was conjugated into P . aeruginosa MPAO1 and ΔbfmRS strains and the construct was integrated into the attB site as described previously though a diparental mating using E . coli S17 λ-pir as the donor , yielding a MPAO1::BfmR-Flag strain and a ΔbfmRS::BfmR-Flag strain , respectively ( Table S1 in Text S1 ) . In these mutant strains , parts of the mini-CTX-lacZ vector containing the tetracycline resistance cassette were deleted using a flippase ( FLP ) recombinase encoded on the pFLP2 plasmid . To construct the plasmid for constitutive expression of bfmR , a 806 bp PCR product covering 15 bp of the bfmR upstream region , the bfmR gene , and 50 bp downstream of bfmR was amplified using primers BfmR ( comp ) Fwr ( with HindIII site ) and BfmR ( comp ) Rev ( with BamHI site ) . The product was digested with HindIII and BamHI and ligated into PAK1900 [60] in the same orientation as plac to generate p-bfmR . To construct the plasmid for the constitutive expression of bfmS , a 1 , 385 bp PCR product covering 30 bp of the bfmS upstream region , the bfmS gene , and 50 bp downstream of bfmS was amplified using primers BfmS ( comp ) Fwr ( with HindIII site ) and BfmS ( comp ) Rev ( with BamHI site ) , and then cloned into PAK1900 , yielding p-bfmS . To construct the plasmid for the constitutive expression of bfmRS , a 2 , 107 bp PCR product covering 15 bp of the bfmR upstream region , the bfmRS operon , and 50 bp downstream of bfmS was amplified using primers BfmR ( comp ) Fwr ( with HindIII site ) and BfmS ( comp ) Rev ( with BamHI site ) and then cloned into PAK1900 , yielding p-bfmRS . To construct the plasmid for the constitutive expression of rhlR , a 770 bp DNA fragment covering 44 bp of the rhlR upstream region and the rhlR was amplified using primers RhlR-OE-F ( with HindIII site ) and RhlR-OE-R ( with HindIII site ) and then cloned into PAK1900 . The construct with rhlR in the same orientation as plac was confirmed by DNA sequencing , yielding p-rhlR . To construct the plasmid for constitutive expression of acka-pta operon , a 3 , 563 bp DNA fragment covering 136 bp of acka upstream region , the acka-pta operon , and a 65 bp downstream of pta was amplified using primers Acka-comp-F ( with HindIII site ) and Pta-comp-R ( with BamHI site ) and then cloned into PAK1900 , yielding plasmid p-ackA-pta . The five mutations , p-bfmRD55A , p-bfmSL181P , p-bfmSE376Q , and p-bfmSL181P/E376Q , and p-bfmSR393H were obtained using the QuikChange II site-directed mutagenesis kit ( Stratagene ) . For generating p-bfmRD55A , the primer pair BfmR ( D55A ) -F/BfmR ( D55A ) -F was used . For generating p-bfmSL181P , primer pair PA4102L181P-F/PA4102L181P-R was used . For generating p-bfmSE376Q , the primer pair PA4102E376Q-F/PA4102E376Q-R was used . For generating p-bfmSL181P/E376Q , primer pairs PA4102L181P-F/PA4102L181P-R and PA4102E376Q-F/PA4102E376Q-R were used . For generating p-bfmSR393H , the primer pairs R393H-F/R393H-R was used . All constructs were sequenced to ensure that no unwanted mutations resulted . Full-length of bfmR was cloned into pET28a with a thrombin-cleavable N-terminal His-tag . Primers bfmR-F ( with NdeI site ) and bfmR-R ( XhoI ) were used to amplify the bfmR gene from P . aeruginosa MPAO1 chromosomal DNA . The amplified fragments were ligated into similarly cut pET28a ( Novagen ) in order to produce the plasmids pET28a-6His-BfmR . pET28a-6His-BfmRD55A was obtained by using the primer pair BfmR ( D55A ) -F/BfmR ( D55A ) -R and a QuikChange II site-directed mutagenesis kit ( Stratagene ) . The protein was expressed in the E . coli strain BL21 star ( DE3 ) and purifications were performed as described in our previous studies [26] , [61] , [62] . Briefly , bacteria were grown at 37°C overnight in 10 ml of LB medium ( containing 50 µg/ml kanamycin ) with shaking ( 250 rpm ) . The next day , the cultures were transferred into 1 L of LB medium ( containing 50 µg/ml kanamycin ) incubated at 37°C with shaking ( 250 rpm ) until the OD600 reached 0 . 6 , and then IPTG ( isopropyl-1-thio-β-d-galactopyranoside ) was added to a final concentration of 1 . 0 mM . After 4 h incubation at 30°C with shaking ( 250 rpm ) , the cells were harvested by centrifugation and stored at −80°C . The cells were lysed at 4°C by sonication in lysis buffer [10 mM Tris ( pH 7 . 4 ) , 300 mM NaCl , 1 mM PMSF , and 2 mM DTT] . Clarified cell lysate was loaded onto a HisTrap HP column ( Amersham Biosciences ) , washed with Ni-NTA washing buffer , and eluted with Ni-NTA elution buffer . The fractions containing 6His-BfmR or 6His-BfmRD55A were concentrated and loaded onto a Superdex-200 gel filtration column with a running condition of 10 mM Tris ( pH 7 . 4 ) , 300 mM NaCl , and 2 mM DTT . The purified protein was >90% pure as estimated by a 12% ( wt/vol ) SDS/PAGE gel . The DNA sequence of the extracellular sensory domain of BfmS consisting of 121 residues ( Gln34-Trp154 ) was amplified from MPAO1 genomic DNA with the primers PA4102-EX-F ( with NcoI ) and PA4102-EX-R ( with BamHI ) by PCR and was subsequently cloned into pET28b using NcoI and BamHI as the restriction enzymes . Following confirmation by DNA sequencing , the recombinant plasmid ( pET28b-bfmS34–154 ) was transformed into E . coli strain BL21 star ( DE3 ) . The extracellular sensory domain of BfmS ( designated BfmS34–154 ) was expressed and purified as described above with some modifications . Briefly , bacteria were grown at 37°C overnight in 10 ml of LB medium ( containing 50 µg/ml kanamycin ) with shaking ( 250 rpm ) . The next day , the cultures were transferred into 1 L of LB medium ( containing 50 µg/ml kanamycin ) incubated at 37°C with shaking ( 200 rpm ) until the OD600 reached 0 . 6 , and then IPTG was added to a final concentration of 1 . 0 mM . After 16 h incubation at 16°C with shaking ( 200 rpm ) , the cells were harvested by centrifugation and stored at −80°C . The cells were lysed at 4°C by sonication in lysis buffer [50 mM Tris-HCl , pH 8 . 0 , 100 mM NaCl , 10% glycerol 1 mM PMSF , and 2 mM DTT] . Clarified cell lysate was loaded onto a HisTrap HP column ( Amersham Biosciences ) , and eluted with Ni-NTA elution buffer ( 50 mM Tris-HCl , pH 8 . 0 , 100 mM NaCl , 20 mM imidazole 10% glycerol , and 1 mM DTT , ) . The fractions containing BfmS34–154 were concentrated and loaded onto a Superdex-75 gel filtration column with a running condition of 20 mM Tris-HCl , pH 8 . 0 , 100 mM NaCl , 10% glycerol , 1 mM DTT . The purified protein was >90% pure as estimated by a 12% ( wt/vol ) SDS/PAGE gel . The plasmid pMS402 [63] carrying a promoterless luxCDABE reporter gene cluster was used to construct promoter-luxCDABE reporter fusions of the bfmR as described previously [28] , [64] . For bfmR-lux , the bfmR promoter region ( −463 to +18 of the start codon ) was amplified by PCR using the primers PMS402-bfmR-F ( with XhoI site ) and PMS402-bfmR-R ( with BamHI site ) . For rhlA-lux , the rhlA promoter region ( −526 to −20 of the start codon ) was amplified by PCR using the primers pms402-rhlA-F ( with XhoI site ) and pms402-rhlA-R ( with BamHI site ) . For rhlR-lux , the rhlR promoter region ( −450 to +19 of the start codon ) was amplified by PCR using the primers pms402-rhlR-1F ( with BamHI site ) and pms402-rhlR-R ( with BamHI site ) . The promoter oriented in the same direction as luxCDABE was selected for further analysis . To generate rhlR promoter mutant rhlR-D ( deletion of GATACT , which is the central part of the putative BfmR-binding site on the reverse DNA strand ) , the DNA fragment was amplified using primers pms402-rhlR-1F/pms402-rhlR-R and subsequently cloned into pGEM-T vector . rhlR-D ( rhlR promoter lacking putative BfmR-binding site ) was obtained using a QuikChange II site-directed mutagenesis kit ( Stratagene ) and primer pair pms402-rhlR ( D1 ) F/pms402-rhlR ( D1 ) R . For 4103-lux , the PA4103 promoter region ( −659 to +19 of the start codon ) was amplified by PCR using primers pms402-p4103-F ( with XhoI site ) and pms402-p4103-R ( with BamHI site ) . To generate PA4103 promoter mutant 4103-M ( GATACA was mutated to ATATAT ) , primer pair p4103-mutation-F/p4103-mutation-R was used as described above . The promoter regions were cloned into the XhoI-BamHI site or BamHI site ( for rhlR-lux ) upstream of the lux genes on pMS402 and the cloned promoter sequences were confirmed by DNA sequencing . The constructs were transformed into MPAO1 or its derivatives by electroporation . Use of these lux-based reporters , gene expression under different conditions was measured as counts per second ( cps ) of light production with a 2104 EnVision Multilabel Plate Readers or Synergy 2 ( Biotek ) . Relative light units were calculated by normalizing CPS to OD600 . The bacterial growth and the extraction of total RNAs were performed as described above . The total DNase-treated RNA ( 5 µg ) was reversely transcribed to synthesize cDNA using the PrimeScript RT reagent Kit ( Takara ) with random primers according to the manufacturer's protocol . The resulting cDNA were diluted by 1∶2 , 1∶4 , and 1∶8 respectively . Triplicate quantitative assays were performed on 1 µl of each cDNA dilution with the THUNDERBIRD SYBR qPCR Mix and 300 nM primers using an Applied Biosystems 7500 Fast Real-Time PCR System . Dissociation curve analysis was performed in order to verify product homogeneity . The gene-specific primers used for Quantitative real-time PCR for PA4100 , PA4103 , PA4107 , PA4108 , ntrB , oprH , phoB , hmgA , rhlA , antA , nasA , and rhlI are listed in Table S6 in Text S1 . The amplicon of 16S rRNA was used as an internal control in order to normalize all data . Relative expression levels of interest genes were calculated by the relative quantification method ( ΔΔCT ) as previously described [65] , [66] . P . aeruginosa was grown at 37°C for 24 h on M8-glutamate minimal agar plate ( M8-glutamate minimal medium supplemented with 0 . 2% glucose , and solidified with 2% agar ) . To prepare cell lysates for the Phos-tag gel assay , bacteria cells were scraped from the plate and immediately resuspended in 60 µl of lysis buffer [50 mM Tris-Cl ( pH 7 . 5 ) , 150 mM NaCl , 1 mM MgCl2 , 0 . 1% Triton X-100 , 15 µg/ml DNaseI , 0 . 5 mM PMSF , 1 mM DTT ) with 0 . 1% ( vol/vol ) Lysonase . Sufficient lysis was achieved by repeated pipetting up and down for 10 s followed by addition of 20 µl of 4×SDS loading buffer . Resulting cell lysates ( 10 µl ) were immediately loaded onto a Phos-tag gel for electrophoresis . BfmR-flag and BfmR-flag∼P were separated on 10% acrylamide gels containing 25 µM acrylamide-Phos-tag ligand ( Wako Pure Chemical ) and 50 µM MnCl2 as previously described [67] . Electrophoresis was performed at 30 mA at 4°C for 80 min in Tris-Glycine-SDS running buffer ( 25 mM Tris , 192 mM glycine , 0 . 1% SDS , pH 8 . 4 ) . After electrophoresis , the Phos-tag gel was washed 10 min at RT with Transfer Buffer [20% ( v/v ) methanol , 50 mM Tris , 40 mM glycine] supplied with 1 mM EDTA to remove Zn2+ from the gel , then the gel was incubated at room temperature with gentle shaking for another 10 min in Transfer Buffer twice to remove EDTA . Samples resolved on gels were transferred to PVDF ( Bio-Rad ) membranes through semi-dry transfer assembly ( Bio-Rad ) for 30 min at room temperature . BfmR-Flag proteins were detected by Western blot analysis using a mouse anti-Flag monoclonal antibody ( Cat#: AGM12165 , Aogma ) followed by a secondary , sheep anti-mouse IgG antibody conjugated to horseradish peroxidase ( HRP ) ( Code#: NA931 , GE Healthcare ) . For detection of ClpP protein , anti-ClpP polyclonal antibody and anti-rabbit IgG antibody conjugated to horseradish peroxidase ( HRP ) ( Code#: NA934 , GE Healthcare ) were used . Anti-ClpP polyclonal antibody , which cross-reacts with the ClpP of Pseudomonas aeruginosa , was prepared by immunizing a rabbit with a Staphylococcus aureus full-length ClpP protein ( NWMN_0736 ) . For detection of BfmS protein , anti-BfmS polyclonal antibody ( prepared by immunizing a rabbit with a BfmS34–154 protein , Shanghai Immune Biotech CO . , Ltd ) and anti-rabbit IgG antibody conjugated to horseradish peroxidase ( HRP ) ( Code#: NA934 , GE Healthcare ) were used . For detection of RNAP protein , anti-RNAP ( Neoclone , #WP003 ) antibody and anti-mouse IgG antibody conjugated to horseradish peroxidase ( HRP ) ( Code#: NA931 , GE Healthcare ) . Immunoblots for ClpP and RNAP served as loading control . The membrane is exposed to X-ray film ( Kodak ) or the chemiluminescent is detected by a Imaging Quant LAS-4000 ( GE ) , according to the manufacturer's recommendation . Pyocyanin was extracted from culture supernatants and measured using previously reported methods [68] . Briefly , P . aeruginosa was grown in Pyocyanin production broth [57] ( PPB: 20 g peptone , 1 . 4 g MgCl2 , 10 g K2SO4 , 20 ml glycerol per liter; pH 7 . 0 ) for 36 h at 37°C with shaking ( 250 rpm ) . The culture was subsequently centrifuged and filtered ( pore size , 0 . 22 µm ) . 1 . 5 ml of chloroform was added to 2 . 5 ml of culture supernatant . After extraction , the chloroform layer was transferred to a fresh tube and mixed with 1 ml of 0 . 2 N HCl . After centrifugation , the top layer ( 0 . 2 N HCl ) was removed and its absorption measured at 520 nm . Concentrations , expressed as micrograms of pyocyanin produced per ml of culture supernatant , were determined by multiplying the optical density at 520 nm ( OD520 ) by 17 . 072 . Rhamnolipids production was estimated by inoculating strains on M8-based agar plates supplemented with 0 . 2% glucose ( m/v ) , 2 mM MgSO4 , 0 . 0005% ( m/v ) methylene blue , and 0 . 02% ( m/v ) cetyltrimethylammonium bromide , as described previously [68] , [69] . The orcinol assay was used to directly assess the amount of rhamnolipids in the sample as previously described [28] . After a culture of 48 h in LB medium at 37°C with shaking ( 250 rpm ) , 1 ml of the culture supernatant was extracted twice with 2 ml of diethyl ether . The pooled ether fractions were evaporated to dryness and the remainder was dissolved in 100 µl of distilled water and mixed with 100 µl of 1 . 6% orcinol , and 800 µl of 60% sulfuric acid . After heating for 30 min at 80°C in the dark , the samples were cooled for 3 h at room temperature in the dark . Absorbance at 421 nm ( A421 ) was measured . Rhamnolipid concentrations were calculated by comparing A421 values with those obtained for rhamnose standards between 0 and 1000 µg/ml , assuming that 1 µg of rhamnose corresponds to 2 . 5 µg of rhamnolipids . The motility assay was carried out as described previously [27] , [68] . Swarming medium was based on M8-glutamate minimal medium [27] ( 6 g Na2HPO4 , 3 g KH2PO4 , 0 . 5 g NaCl , 0 . 24 g MgSO4 , 0 . 5 g glutamate per liter; pH 7 . 4 ) , supplemented with MgSO4 ( 2 mM ) , glucose ( 0 . 2% ) , and Casamino acid ( 0 . 5% ) , and solidified with 0 . 5% agar . Bacteria were inoculated with a toothpick onto swarm agar plates . Swarm agar plates were incubated for 24 hours at 37°C and then incubated for more time at room temperature . Phosphorylation of 6His-BfmR was detected by utilizing the Pro-Q Diamond phosphorylation gel stain as described by the manufacturer ( Invitrogen ) . Purified 6His-BfmR and 6His-BfmRD55A were incubated with buffer ( 10 mM Tris pH 8 . 0; 1 mM DTT; 5 mM MgCl2; 10 mM KCl; 50 mM acetyl phosphate ) at 37°C for 30 min . The acetyl phosphate-treated samples of 6His-BfmR and 6His-BfmRD55A were resolved on a 12% SDS polyacrylamide gel , and then the gel was immersed in fixing solution ( 10% acetic acid , 50% methanol ) for 30 min and subsequently washed three times with deionized water each for 10 min . The gel was stained with Pro-Q Diamond phosphoprotein gel stain for 60 min , followed by washing with deionized water for 30 min . The entire procedure was conducted at room temperature . Fluorescent output was recorded using an Tanon-5200 multi . The electrophoretic mobility shift experiments were performed as described in our previous studies with some modifications [26] , [61] , [62] . Briefly , 20 µl of the DNA probe mixture ( 30 to 50 ng ) and various amounts of purified proteins in binding buffer ( 10 mM Tris-Cl , pH 8 . 0; 1 mM DTT; 10% glycerol; 5 mM MgCl2; 10 mM KCl ) were incubated for 30 min at 37°C . When indicated , 50 mM acetyl phosphate was added to the solution . Native polyacrylamide gel ( 6% ) was run in 0 . 5× TBE buffer at 85 V at 4°C . The gel was stained with GelRed nucleic acid staining solution ( Biotium ) for 10 min , and then the DNA bands were visualized by gel exposure to 260-nm UV light . DNA probes were PCR-amplified from P . aeruginosa MPAO1 genomic DNA using the primers listed in Table S6 in Text S1 . The probes for bfmR promoter , a 481 bp DNA fragment covering the promoter region of bfmR ( from −463 to +18 of the start codon ) was amplified using primers bfmR-F ( EMSA ) and bfmR-R ( EMSA ) . For rhlR promoter , a 470 bp DNA fragment covering the promoter region of rhlR ( from −450 to +20 of the start codon ) was amplified using primers rhlR-F ( EMSA ) and rhlR-F ( EMSA ) . For rhlI promoter , a 446 bp DNA fragment covering the promoter region of rhlI ( from −444 to +2 of the start codon ) was amplified using primers rhlI-F ( EMSA ) and rhlI-R ( EMSA ) . For rhlA promoter , a 572 bp DNA fragment covering the promoter region of rhlA ( from −591 to −19 of the start codon ) was amplified using primers rhlA-F ( EMSA ) and rhlA-R ( EMSA ) . For rhlC promoter , a 540 bp DNA fragment covering the promoter region of rhlC ( from −549 to −9 of the start codon ) was amplified using primers rhlC-F ( EMSA ) and rhlC-F ( EMSA ) . For PA4103 promoter , a ca . 0 . 7 kb DNA fragment ( 4103-P ) containing the promoter region of PA4103 ( from −659 to +19 of the start codon ) was amplified from plasmid 4103-lux DNA using primers pZE . 05 and pZE . 06 . For PA4107 promoter , a 360 bp DNA fragment ( 4107-P ) covering the promoter region of PA4107 ( from −490 to −131 of the start codon ) was amplified from P . aeruginosa MPAO1 genomic DNA using primers PA4108-F and PA4108-R . All PCR products were purified by using a QIAquick gel purification kit ( QIAGEN ) . The published DNase I footprint protocol was modified [70] in this study in the same way as described in our previous study [61] . Briefly , PCR was used to generate DNA fragments using the primer sets as detailed in Table S6 in Text S1 . For amplification of bfmR promoter , primers bfmR-F ( EMSA ) and 6FAM-bfmR-R were used . For amplification of the rhlR promoter , primers rhlR-F ( EMSA ) and 6FAM-rhlR-R were used . For amplification of the PA4103 promoter , p4103-F ( EMSA ) and p4103-R-FAM were used . All PCR products were purified by with QIAquick gel purification kit ( QIAGEN ) . 50 nM 6-carboxyfluorescein ( 6-FAM ) -labeled promoter DNA and various amounts of 6His-BfmR ( as indicated ) in 50 µl of binding buffer ( 10 mM Tris-Cl , pH 8 . 0; 1 mM DTT; 10% glycerol; 5 mM MgCl2; 10 mM KCl; 50 mM acetyl phosphate ) were incubated at room temperature for 10 min . 0 . 01 unit of DNase I was added to the reaction mixture and incubated for 5 more min . The digestion was terminated by adding 90 µl of quenching solution ( 200 mM NaCl , 30 mM EDTA , 1% SDS ) , and then the mixture was extracted with 200 µl of phenol-chloroform-isoamyl alcohol ( 25∶24∶1 ) . The digested DNA fragments were isolated by ethanol precipitation . 5 . 0 µl of digested DNA was mixed with 4 . 9 µl of HiDi formamide and 0 . 1 µl of GeneScan-500 LIZ size standards ( Applied Biosystems ) . A 3730XL DNA analyzer detected the sample , and the result was analyzed with GeneMapper software . Overnight P . aeruginosa cultures were washed and diluted 100-fold in M8-glutamate minimal medium ( 6 g Na2HPO4 , 3 g KH2PO4 , 0 . 5 g NaCl , 0 . 24 g MgSO4 , 0 . 5 g Glutamate per liter; pH 7 . 4 ) supplemented with glucose ( 2 g/L ) . The bacteria were subsequently grown at 37°C for 48 h ( OD600≈1 . 0 ) with shaking ( 250 rpm ) . Total RNA was immediately stabilized with RNAprotect Bacteria Reagent ( Qiagen , Valencia , CA ) and then extracted using a Qiagen RNeasy kit following the manufacturer's instructions . The total DNase-treated RNA samples were then analyzed by CapitalBio Corp for Chip ( Affymetrix ) assay . Briefly , samples were labeled according to the manufacturer ( Affymetrix , Santa Clara , CA ) and then hybridized to the Affymetrix GeneChip P . aeruginosa genome array ( catalog number AFF-900339 ) for 16 h at 50°C though the use of the GeneChip hybridization oven at 60 rpm . Washing , staining , and scanning were performed using the Affymetrix GeneChip system . The data were normalized using Robust Multi-array Average ( RMA ) [71] . Gene expression analysis was performed using three independent mRNA samples for each strain . Microarray data were analyzed using SAM ( Significance Analysis of Microarrays ) software [72] . Criterion such as cutoff limitation for fold change ≥2 or ≤0 . 5 and q-value ≤5% was used to select differential expression genes . All data were submitted to the ArrayExpress database ( http://www . ebi . ac . uk/arrayexpress ) under accession number E-MTAB-1983 . The autoinducer of the rhl system , C4-HSL , was measured using an rhlA promoter-based P . aeruginosa strain , pDO100 ( pKD-rhlA ) [28] . This detection system was developed by fusing the C4-HSL-responsive rhlA promoter upstream of luxCDABE and introducing the construct into pDO100 , a rhlI mutant strain [28] . Procedures were modified from the protocol described previously [28] . Briefly , the reporter strain pDO100 ( pKD-rhlA ) was grown in LB medium plus 100 µg/ml kanamycin overnight at 37°C with shaking ( 250 rpm ) and diluted to an OD600 of 0 . 05 in fresh LB plus kanamycin . 90 µl was subsequently added to the wells of a 96-well microtitre plate . A 10 ml portion of the samples or medium control was added to the wells . The luminescence value was measured in a 2104 EnVision Multilabel Plate Readers or Synergy 2 ( Biotek ) , and calculated from the luminescence value minus that of the medium control . The data are presented as CPS and are not normalized to OD600 of pDO100 ( pKD-rhlA ) . In this assay , the growth curves of pDO100 ( pKD-rhlA ) are identical . Different strains of bacteria were grown overnight in Luria-Bertani ( LB ) broth at 37°C with shaking . Then , the bacteria were subject to pelleting by centrifugation at 5 , 000 g and resuspended in 10 ml of fresh LB broth and allowed to grow until the mid-logarithmic phase . OD600 nm was measured , and the density was adjusted to 0 . 25 OD ( 0 . 1 OD = 1×108 cells/ml ) . Mammalian cells were washed once with PBS after overnight culture in full medium , and changed to a serum-free and antibiotic-free medium immediately before infection . Cells were infected by various strains at a multiplicity of infection ( m . o . i ) of 10∶1 bacteria-to-cell ratio at indicated time points for 2 h . The cells were washed three times with PBS to remove surface bacteria and incubated with 100 µg/ml polymyxin B for 1 h . The cells were lyzed to evaluate the internalized bacteria using CFU assay on agar dishes as described in our previous studies [73] , [74] . The killing of MLE-12 cells by bacterial infection was performed by continuing incubation for 24 h and cell survival measured using MTT assay [75] . MLE-12 cells were cultured in 96-well plates as above . After incubation for 24 h , MTT dye was added to the cells in each well with at a final concentration of 1 µg/ml . Then , the cells were incubated at 37°C until the color developed . The yellow color may change to brown upon reduction by enzymes . The reaction was terminated by adding 100 µl of stop solution ( 10% DMSO , 10% SDS in 50 mM HEPES buffer ) . The plate was left at room temperature overnight . The next day , the 560-nm absorbance was read using a plate reader in order to quantify the dye conversion [76] . Background correction was done with controls containing only the dye . A lettuce leaf virulence assay was performed as described previously [77]–[79] . Briefly , P . aeruginosa strains were grown aerobically overnight at 37°C with shaking ( 250 rpm ) in PPB broth or PPB broth containing carbenicillin ( 100 µg/ml ) when appropriate , washed , resuspended , and diluted in sterile MgSO4 to a bacterial density of 1×109 CFU/ml . Lettuce leaves were prepared by washing with sterile distilled H2O and 0 . 1% bleach . Samples ( 10 µl ) were then inoculated into the midribs of Romaine lettuce leaves . Containers containing Whatman paper moistened with 10 mM MgSO4 and inoculated leaves were kept in a growth chamber at 37°C for five days . Symptoms were monitored daily . As a control , lettuce leaves were inoculated with 10 mM MgSO4 . All P . aeruginosa strains were grown at 37°C overnight in PPB medium with shaking ( 250 rpm ) , diluted 100-fold in fresh PPB medium , and incubated at 37°C for 2 . 5–3 . 0 h until the cultures reached OD600 0 . 8 . Bacteria were collected by centrifugation , washed , and suspended in PBS buffer . Viable P . aeruginosa were enumerated by colony formation on Pseudomonas isolation agar ( PIA ) ( Difco ) plates in order to quantify the infectious dose . Mouse infections were carried out as described previously [26] , using 8-week-old female C57BL/6J mice obtained from Shanghai SLAC Laboratory Animal Co . Ltd . and housed under specified pathogen-free conditions . Mice were anaesthetized with pentobarbital sodium ( intraperitoneal injection , 80 mg/kg ) and intranasally infected with c . 5×106 cfu of each bacterial isolate . After that , animals were sacrificed 18 h post infection . Lungs were aseptically removed and homogenized in PBS plus 0 . 1% Triton X-100 in order to obtain single-cell suspensions . Serial dilutions of each organ were plated on Pseudomonas isolation agar ( PIA ) ( Difco ) plates . Bacterial burden per organ was calculated and is expressed as a ratio of the inoculum delivered per animal . Statistical analysis was performed using Prism software ( GraphPad ) . Two-tailed Student's t tests or Mann–Whitney test was used to calculate p-values ( two-tailed ) using Prism software ( GraphPad ) , as indicated . * p<0 . 05 , ** p<0 . 01 , *** p<0 . 001 . | The rhl quorum-sensing ( QS ) system allows P . aeruginosa to regulate diverse metabolic adaptations and virulence . However , how rhl QS system is regulated remains largely unknown . Here , we report that two-component sensor BfmS controls rhl QS system by repressing its cognate response regulator BfmR , which directly suppresses the expression of rhl QS regulator RhlR gene and reduces the production of QS signal molecule N-butanoyl-L-homoserine lactone ( C4-HSL ) . We find that BfmS is critical to the ability of P . aeruginosa to modulate the expression of virulence-associated traits and adapt to the host . Intriguingly , although wild-type BfmS is a repressor of BfmR , naturally occurring missense mutation ( L181P , L181P/E376Q , or R393H ) can convert its function from a repressor to an activator of BfmR , leading to BfmR activation , which in turn reduces the level of rhl QS signal C4-HSL . These results , therefore , provide important and novel insight into the regulation and evolution of P . aeruginosa virulence . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"bacteriology",
"biology",
"and",
"life",
"sciences",
"microbiology"
]
| 2014 | A Novel Signal Transduction Pathway that Modulates rhl Quorum Sensing and Bacterial Virulence in Pseudomonas aeruginosa |
With the approach of winter , many insects switch to an alternative protective developmental program called diapause . Drosophila melanogaster females overwinter as adults by inducing a reproductive arrest that is characterized by inhibition of ovarian development at previtellogenic stages . The insulin producing cells ( IPCs ) are key regulators of this process , since they produce and release insulin-like peptides that act as diapause-antagonizing hormones . Here we show that in D . melanogaster two neuropeptides , Pigment Dispersing Factor ( PDF ) and short Neuropeptide F ( sNPF ) inhibit reproductive arrest , likely through modulation of the IPCs . In particular , genetic manipulations of the PDF-expressing neurons , which include the sNPF-producing small ventral Lateral Neurons ( s-LNvs ) , modulated the levels of reproductive dormancy , suggesting the involvement of both neuropeptides . We expressed a genetically encoded cAMP sensor in the IPCs and challenged brain explants with synthetic PDF and sNPF . Bath applications of both neuropeptides increased cAMP levels in the IPCs , even more so when they were applied together , suggesting a synergistic effect . Bath application of sNPF additionally increased Ca2+ levels in the IPCs . Our results indicate that PDF and sNPF inhibit reproductive dormancy by maintaining the IPCs in an active state .
To synchronize with the Earth’s rhythmic environment , most higher organisms have evolved endogenous time-keeping systems [1 , 2] . While the highly conserved circadian clock is well-characterized [3 , 4] , our knowledge of the seasonal clock that governs the overwintering response ( diapause ) in insects is still fragmentary [5 , 6] . Diapause refers to an alternative developmental program , typically induced by the shortening days and falling temperatures of the approaching winter [1 , 7] . Diapausing animals are characterized by low metabolic rate , drastically decreased food intake , extended lifespan , and increased stress resistance [8–15] . The fruit fly Drosophila melanogaster exhibits an adult reproductive ‘diapause’ manifested by arrested ovarian development which is stimulated by low temperatures and can be enhanced by short photoperiods [15 , 16] . While the Drosophila literature refers to this phenomenon as ‘diapause’ rather than ‘dormancy’ or ‘overwintering’ , we recognize that it is not a classic photoperiodically-induced state because it requires cold-temperature to induce the reproductive quiescence . Nevertheless , it shows features that are commonly associated with responses that are resistant to environmental stresses [7 , 15 , 16] . An increasing body of evidence suggests that insulin-like signaling is a key regulator of diapause in numerous species [13 , 17–20] . In Drosophila , 4 of the 8 identified insulin-like proteins ( DILP1 , -2 , -3 , -5 ) are produced in 14 median neurosecretory cells ( designated as Insulin Producing Cells , IPCs ) of the Pars intercerebralis ( PI ) , which are anatomically connected to the key neuroendocrine system that governs hormonal regulation of gonadal arrest [21–25] . Even though the center for dormancy control is believed to be in the PI [26–28] , the neurosecretory cells in this brain area , including the IPCs , do not have a circadian clock , and would therefore have to receive any timing information ( if any ) from other cells [29–33] . A challenging question is how the environmental signals that trigger dormancy ( i . e . decreasing photoperiod and temperature ) including putative timing information , are perceived , interpreted , and converted into hormonal signals in the brain , leading to the overwintering phenotype [34] . Even though several neuropeptides , neurotransmitters , and peptide hormones have already been identified as modulators of function of the IPCs ( reviewed in [35 , 36] ) , there are still gaps in our understanding of how the activity of these cells is controlled . Recent research revealed a synaptic connection between the IPCs and one group of dorsal clock neurons ( DN1 ) , raising the possibility of a direct modulatory effect exerted by the circadian system [33] . Indeed , natural variants of the timeless clock gene in D . melanogaster , namely the s-tim and ls-tim allelic variants , are known to have a dramatic effect on the inducibility of reproductive dormancy [37 , 38] . In our study , we primarily focused on neuronal clusters that , based on their axonal projections to the dorsal brain ( dorsal protocerebrum ) , could play an intermediary role in conveying dormancy-inducing signals towards the IPCs . Neurites of the small ventral lateral neurons ( s-LNvs ) project to the dorsal protocerebrum , where they rhythmically release the circadian neuromodulator PIGMENT DISPERSING FACTOR ( PDF ) [39] . PDF is also expressed in the large ventral lateral neurons ( l-LNvs ) , and in two groups of non-clock cells , a developmentally transient neuronal cluster in the Tritocerebrum ( designated as PDF-Tri ) , and a small number of neurons in the eighth abdominal neuromere of the ventral ganglion ( designated as PDFAb neurons ) [40] . PDF is a key coordinator of pacemaker interactions and behavioral rhythms [41–43] , of sleep and arousal [44 , 45] and of sexual behavior [46] . PDF may also be involved in diapause regulation in the blow fly Protophormia terraenovae [47] , the mosquito Culex pipiens [48] , and the bean bug Riptortus pedestris [49] . However , its effects appear to be contradictory and the mechanisms through which PDF acts on diapause are unclear . Short Neuropeptide F ( sNPF ) has been implicated in the modulation of diapause in the Colorado potato beetle [50] and has been shown to stimulate ovarian development in the locust [51 , 52] . In Drosophila sNPF increases food intake and body size [53] and enhances growth [54] . This peptide is broadly produced in the Drosophila nervous system [53 , 55] , including the s-LNvs [56] . A small set of bilaterally symmetric neurons in the Pars lateralis ( PL ) , defined as dorsal-lateral peptidergic neurons ( DLPs ) , also express sNPF . DLPs have axon terminations in the proximity of the IPCs , and co-express the multi-functional neuropeptide Corazonin ( Crz ) [57] , which has also been proposed as a diapause regulating peptide in the hawkmoth Manduca sexta [58] . The G protein-coupled receptors for PDF , sNPF , and Crz ( PDFR , sNPFR1 , and CrzR , respectively ) have already been characterized and extensively studied in Drosophila [57 , 59–67] . Interestingly , sNPFR1 and CrzR have been found to influence the activity of the IPCs [53 , 54 , 57 , 68] . However , to date no studies have reported PDF signaling to the IPCs . In the present study , we demonstrate that PDF and sNPF produced by PDF-positive ( PDF+ ) neurons , reduce the induction of dormancy in Drosophila , mainly through a direct effect on the IPCs . Conversely , the Corazonin-expressing DLP neurons do not seem to be involved . Using live imaging , we show that the IPCs respond to both PDF and sNPF peptides with increasing levels of cAMP and to sNPF additionally with increasing Ca2+ levels suggesting that PDF and sNPF positively modulate IPCs activity and thereby inhibit gonadal arrest .
To examine whether the PDF-expressing neurons have a role in the regulation of reproductive arrest , we used a Pdf-Gal4 driver to target gene expression specifically in PDF+ neurons including both s- and l-LNvs [39 , 41] . First , a bacterial depolarization-activated sodium channel ( Na+ChBac ) was expressed in these neurons ( Pdf>Na+ChBac ) . Such manipulation will enhance the release of neurotransmitters and neuropeptides , including PDF and sNPF [69] . Pdf>Na+ChBac flies showed significantly lower levels of dormancy compared to controls ( p<0 . 001; Fig 1A ) . Importantly , Na+ChBac expressing and control flies shared the same timeless ( tim ) background , as tim alleles ( s- and ls-tim ) affect the overall level of reproductive arrest ( S1 Table ) [37] . We also overexpressed PDF in the same subset of cells ( Pdf>Pdf ) , which again resulted in a significant decrease in the incidence of dormancy compared to controls ( p<0 . 001; Fig 1A ) . In addition to PDF , the s-LNvs co-express the neuropeptide sNPF . Since sNPF is widely present in the nervous system [53 , 55] , we started by overexpressing this neuropeptide with a pan-neuronal driver . This manipulation ( elav>2xsNPF ) produced a very significant reduction in dormancy in the experimental flies compared to the controls ( p<0 . 001; Fig 1B ) . Considering that both the elav-Gal4 and the UAS-2xsNPF lines carry the ls-tim allele ( S1 Table ) , which is known to promote ovarian quiescence [37] , the antagonistic effect of sNPF is quite dramatic . We then narrowed the overexpression of the neuropeptide specifically to the PDF+ neurons ( Pdf>2xsNPF ) and detected a similar and highly significant reduction of ovarian arrest in the experimental flies compared to the controls ( p<0 . 001; Fig 1C ) . These results suggest that the Pdf-expressing tissues ( the s-LNvs , l-LNvs , the PDF-Tri and the PDF-Ab ) are making the major contribution to the inhibition of dormancy . To test the importance of the s-LNvs we used the R6-Gal4 driver [70] which is active in the s-LNvs and in some other neurons but not in the remaining PDF+ cells [71] . Again , R6>2xsNPF flies showed significantly reduced dormancy when compared to controls ( p<0 . 001; Fig 1C ) . Thus , we conclude that sNPF , likely released from the s-LNvs , is involved in the negative regulation of gonadal arrest . We then considered the opposite manipulation , namely reduced neuronal excitability , which ultimately results in reduced release of neuropeptides . Neuronal overexpression of the potassium channel Ork increases potassium efflux and causes membrane hyperpolarization , thereby preventing the firing of action potentials [72] . Pdf>Ork flies showed higher levels of ovarian arrest compared to controls ( p<0 . 001; Fig 2A ) . Furthermore , genetically ablating the PDF+ neurons by overexpressing the pro-apoptotic protein hid ( head involiution defective ) , Pdf>hid ) , also caused a larger proportion of females to undergo dormancy compared to controls ( p<0 . 001; Fig 2B ) . As the PDF neurons co-express more than just PDF we also examined whether the Pdf01 null mutation would alter dormancy levels at two photoperiods , LD8:16 and LD16:8 . The experiment was performed on the ls-tim background . Fig 2C reveals that there is a significant genotype effect ( p = 1 . 3 x 10−4 ) with Pdf01 mutants showing significantly elevated levels of ovarian quiescence plus a significant photoperiodic effect ( p = 6 x 10−6 ) , with no significant interaction , so the enhancement in dormancy occurs at both photoperiods equally . We also examined the effects of overexpressing the PDF receptor ( PDFR ) in the IPC cells using dilp2 ( p ) -Gal4 [21 , 73] in a homozygous receptor mutant han background . This was compared to parental controls ( dilp2 ( p ) -Gal4 and UAS-PDFR ) that were both homozygous and heterozygous for han , as well as the dilp2 ( p ) >+ wild-type control . All six genotypes were placed on the s-tim background . Fig 2D reveals that overexpression of PDFR causes a highly significant reduction of dormancy ( p<0 . 001 ) compared to all the corresponding han/han and han/+ controls , but is not significantly different from the dilp2 ( p ) >+ wild-type . Consequently , we appeared to have rescued the mutant phenotype in this genetic background . The heterozygous han/+ background controls also show a highly significant reduction of dormancy compared to their corresponding homozygous mutant controls ( both p<0 . 001 ) and the dilp2 ( p ) >+ wild type ( p<0 . 01 ) , further underscoring the dosage effect of PDFR on the phenotype . All of these experiments are consistent with the view that the s-LNvs modulate dormancy levels via the neuropeptides PDF and sNPF . However , since Pdf-Gal4 is also expressed in the non-circadian PDF-Tri and PDF-Ab neurons [39 , 41] , an influence of the latter cannot be excluded . The IPCs express sNPFR1 [54 , 57 , 68]; its activation by sNPF stimulates organismal growth by promoting the transcription of insulin-like peptides genes [54] . To investigate whether sNPFR1 signaling in the IPCs modulates ovarian arrest we expressed a dominant negative form of the receptor ( UAS-sNPFR1-DN ) under the control of two IPCs-specific drivers: dilp2 ( p ) -Gal4 and InsP3-Gal4 [21 , 73] . The former drives gene expression from the 2nd larval instar , the latter becomes active mainly after larval development [73] . Inhibition of sNFR1 from early larval stages ( dilp2 ( p ) >sNPFR1-DN ) increased only marginally the proportion of dormancy ( Fig 2E ) . However , when the receptor was repressed later in development ( InsP3>sNPFR1-DN ) , a significantly higher proportion of flies showed gonadal arrest compared to controls ( p<0 . 001; Fig 2E ) . Both drivers are specific for the 14 IPCs in the brain [21 , 73] , so we speculate that different degrees of dormancy might reflect differences in the strength of the drivers or compensatory phenomena that occur early in development . We also used the InsP3 driver to knockdown sNPFR1 . We observed a significant increase in gonadal arrest in InsP3> SNPFR1 RNAi compared to the Gal4 ( p = 0 . 017 ) and UAS controls ( p = 0 . 006 ) ( Fig 2F ) . These results are consistent with our previous observations regarding the antagonist nature of sNPF on this phenotype ( Fig 1B and 1C ) , and also suggest a role for sNPF signaling in the IPCs in the regulation of reproductive arrest . The dorsal-lateral peptidergic neurons ( DLPs ) are 6–7 bilaterally symmetric neurons in the Pars Lateralis ( PL ) whose axons ends in the proximity of the IPCs [57] . The DLPs produce the neuropeptides Corazonin ( Crz ) and sNPF through which they modulate the activity of the IPCs , as the latter express the relevant receptors , CrzR ( Corazonin Receptor ) and sNPFR [57] . The DLPs affect survival , stress resistance and levels of circulating carbohydrates and lipids [57 , 74] . Since diapause is associated with marked changes of these parameters , we questioned whether the DLPs are involved in the regulation of this seasonal response . We used two DLP-specific Crz-Gal4 driver lines , Crz1-Gal4 and Crz2-Gal4 , to overexpress Na+ChBac and sNPF , respectively . Both Crz1>Na+ChBac and Crz2>2xsNPF flies showed a reduction in the proportion of females undergoing gonadal dormancy ( p<0 . 001 ) compared to one of the parental controls but not the other ( S1 Fig ) . Therefore , although we cannot totally exclude an effect of the DLPs on ovarian arrest we can conclude that their involvement , if any , is not as robust as that observed for the PDF+ neurons . Next , we asked whether the dorsal terminals of the s-LNvs in the dorsal vicinity might be close enough to the IPCs to enable such sNPF and PDF signaling . By performing ICC with anti-DILP2 and anti-PDF we could not see direct contacts overlapping between the s-LNvs and the IPCs ( Fig 3A ) . However , when we expressed GFP in the IPCs ( dilp2 ( p ) >GFP ) , we observed that the dorsal projections of the s-LNvs are in close proximity to fine processes originating from the IPCs ( Fig 3B ) . To test whether these fine processes are dendrites , we expressed the dendritic marker DenMark in the IPCs ( dilp2 ( p ) >DenMark ) [75] . We found prominent labeling in the IPC processes , indicating that they are of dendritic origin ( Fig 3C and 3D ) . This explains why we could not see them by anti-DILP2 labeling . The IPCs are neurosecretory cells that are crucial for initiating seasonal responses [13 , 17 , 19 , 20] . Since we have shown that PDF and sNPF have a modulatory effect on diapause ( Fig 1 ) , and that PDF+ and sNPF+ neurons appear to contact the IPCs ( Fig 3B ) [57] , we asked whether the IPCs can respond directly to these neuropeptides . The PDF receptor signals primarily via cAMP [59 , 61 , 76] , whereas the signaling cascade following activation of the receptor for sNPF uses cAMP , at least in part [77–79] . Thus , we expressed a genetically encoded fluorescence resonance energy transfer ( FRET ) based cAMP sensor in the IPCs ( dilp2 ( p ) >Epac1camps ) to monitor real-time cAMP levels [62] . A similar experimental design had previously been adopted with success to investigate the presence of the PDF receptor in clock neurons [62] . We bath-applied 10 μM of synthetic PDF to acutely dissected fly brains . This resulted in a slow rise in the intracellular amount of cAMP , measured as the average FRET signal between 100 and 1000 s after the application of the neuropeptide ( light-blue curve and bar; Fig 4A and 4B ) . cAMP FRET signals were 10% ca . higher ( 100–1000 s; p<0 . 001 ) compared to the negative ( modified minimal hemolymph-like solution , HL3 , light-grey curve and bar; Fig 4A and 4B ) control . A similar increment was also observed after presenting PDF together with the sodium channel blocker tetrodotoxin ( TTX , dark-blue curve and bar , 100–1000 s; p<0 . 001; Fig 4A and 4B ) . The latter prevents neuronal communication , suggesting that PDF activates the IPCs directly . However , the short-term ( 100-200s ) response to PDF was unchanged compared to the negative control , either with or without TTX ( Fig 4C and 4D ) . Similar observations were made upon the application of 10 μM synthetic sNPF . This resulted in a slow but significant cAMP increase in the IPCs either in the absence ( yellow curve and bar , 100–1000 s; p<0 . 001 ) or in the presence ( orange curve and bar , 100–1000 s; p<0 . 001 ) of TTX , reflecting a direct activation of the IPCs by sNPF ( Fig 4E and 4F ) . Moreover , as we saw for PDF , the short-term responses to sNPF did not differ from the negative control ( Fig 4G and 4H ) . However , when we applied sNPF and PDF together ( sNPF+PDF ) , at a concentration of 10 μM for each peptide , we recorded an increase in cAMP FRET signal , reaching a level ~15% higher than that for the negative control ( red , 100–1000 s; p<0 . 001; Fig 5A and 5B ) . Moreover , the short-term response was particularly distinctive compared to the applications of single peptides , revealing a ~8% increase in cAMP FRET signals ( red curve and bar; 100–200 s; p<0 . 01; Fig 5C and 5D ) . We repeated the experiment but halving the concentration of each peptide to 5 μM ( sNPF1/2+PDF1/2 ) . This also resulted in a significant increase in cAMP ( pink curve and bar , 100–1000 s; p<0 . 05; Fig 5A and 5B ) . However , we noticed that following an initial increase , after 400 s the concentration of cAMP slowly declined ( pink curve and bar; Fig 5A ) . Focusing on the short-term response , the co-application of ( sNPF1/2+PDF1/2 ) resulted in higher cAMP FRET signals compared to the bath treatment with single peptides; however , the difference was statistically significant only with regard to PDF ( 100-200s; p<0 . 01; Fig 5C and 5D ) . We also tested ( sNPF+PDF+TTX ) and found an increase of ~13% in the amount of cAMP FRET signal compared to the negative control ( magenta curve and bar , 100–1000 s; p<0 . 001; Fig 5E and 5F ) . This response was not statistically different from single peptide applications in the presence of TTX although in the interval 300–600 s the levels of cAMP for ( sNPF+PDF+TTX ) were higher ( Fig 5E and 5F ) . Interestingly , although there were no differences in cAMP levels when comparing the application of sNPF+PDF with or without TTX in the interval 100–1000 s , the distinctive short term response disappeared in the presence of TTX ( Fig 5G and 5H ) . This suggests that the short-term response of the IPCs to the combined application of ( sNPF+PDF ) is indirectly mediated by other cells . We assessed whether the larger increase in cAMP levels induced in the IPCs by the co-application of sNPF+PDF is a specific response . We co-applied sNPF with the following neuropeptides: SDNFMRFamide ( SDNFMRFa ) , adipokinetic hormone ( AKH ) , Drosophila tachykinin ( DTK ) and allatostatin-C ( Ast-C ) . AKH and DTK were chosen because they are known regulators of IPC activity [21 , 73 , 80 , 81] . When sNPF was co-applied with SDNFMRFa , AKH or DTK , the levels of cAMP were significantly reduced compared to the sNPF+PDF co-application ( p<0 . 05 , p<0 . 01 , p<0 . 05 , respectively ) and were indistinguishable from the negative control ( Fig 6 ) . Conversely , the co-application of sNPF+Ast-C resulted in a significant increase in the amount of cAMP but the addition of Ast-C alone was sufficient to evoke similar cAMP responses in the IPCs ( Fig 6 ) . To our knowledge , no data are available on the distribution of the Ast-C receptor in Drosophila , and no link between IPCs and Ast-C has been reported previously . In summary , our data show that the strong and rapid responses of the IPCs we observed are unique to the co-application of PDF and sNPF . In particular , they are not caused either by the combination of other peptides or by receptor cross-activation due to high peptide doses applied . To verify whether the responses to PDF were mediated by its receptor , we carried out the treatment in the PDFR-null ( han ) [59] background ( han; dilp2 ( p ) >Epac1camps ) . The application of PDF in the mutant no longer evoked an increase of cAMP ( light-blue; 100–1000 s; Fig 7A–7D ) , and surprisingly , neither did sNPF ( yellow curve and bar; 100–1000 s; Fig 7A–7D ) . The rapid increase in cAMP levels induced by the co-application of sNPF+PDF was also obliterated in the mutant ( red curve and bar; 100–200 s; Fig 7C and 7D ) . To test whether sNPF acts via PDFR , we rescued PDFR expression in the IPCs in the han mutant background ( han; dilp2 ( p ) >PDFR; Epac1camps ) . In these flies , IPCs strongly responded to PDF , but not to sNPF . It is therefore unlikely that sNPF signals via PDFR ( Fig 7E–7H ) . We observed a slight increase in cAMP ~400s after sNPF application , but this was not significantly different from the negative controls ( yellow curve and bar; 100–1000 s; Fig 7E and 7F ) . The fast response to sNPF was completely absent ( yellow curve and bar; 100–200 s; Fig 7G and 7H ) . This shows that the absence of the response to sNPF in the han mutant cannot be directly attributed to the loss of PDFR . In several cell types PDF signals primarily through cAMP rather than calcium [59 , 61 , 76] . Here we tested whether PDF or sNPF affect the level of intracellular Ca2+ in the IPCs by expressing the genetically encoded Ca2+ sensor GCaMP3 . 0 ( dilp2 ( p ) >GCaMP3 . 0 ) [82] Incubating PDF with freshly dissected brains did not produce change in Ca2+ levels , which were indistinguishable from the negative control ( Fig 8 ) . However , the bath application of sNPF induced a small but significant increase in the Ca2+ signal ( 100-200s , 3 . 4 ± 2 . 8% , p<0 . 05 ) . This was still detectable in the presence of TTX ( 100-200s , 5 . 4 ± 3 . 5% , p<0 . 05 ) , suggesting that this response is not mediated by interneurons but is due to the direct activation of the IPCs by sNPF ( Fig 8 ) .
We have constitutively activated the PDF+ neurons with Na+ChBac [69] and observed a significant reduction in levels of gonadal arrest ( Fig 1A ) . Since this treatment increases membrane excitability , it likely increases release of both PDF and sNPF , whose overexpression also led to a significant reduction in dormancy ( Fig 1B and 1C ) . Manipulations with opposite effect , namely reduction of membrane excitability of the PDF+ neurons through overexpression of the K+ channel Ork [72] or induction of cell death by overexpression of the pro-apoptotic gene hid , resulted in an enhanced dormancy response ( Fig 2A and 2B ) . Furthermore , the Pdf0 mutant showed a significantly elevated response compared to controls and overexpressing PDFR in the IPCs significantly reduced gonadal arrest on both heterozygous and homozygous mutant han backgrounds to wild-type levels . These results consistently support a model where PDF and SNF act to antagonise dormancy , possibly by enhancing dILP expression . When we expressed a dominant negative form of the sNPF receptor in the IPCs , or downregulated the receptor with RNAi , higher dormancy induction was observed especially when using a driver that is active later in development ( Fig 2E and 2F ) . Numerous sNPF-expressing neurons can potentially target the IPCs , including the PDF+ sLNvs and the DLPs . However , the manipulations of the DLPs did not affect reproductive quiescence ( S1 Fig ) . One puzzling aspect of the results is that they appear , at least superficially , to contradict a recent study in which Pdf01 females showed relatively low levels of dormancy and did not reveal a photoperiodic effect [83] . The LD cycles used in these experiment were LD16:8 versus LD10:14 so not as extreme as ours and this might have had a damping effect on any photoperiodicity . Furthermore , the low dormancy levels of the mutant females suggested that the mutation might have been on the s-tim background . A congenic wild-type was not compared to the mutant , so it is difficult to predict whether such a control would have had lower levels of reproductive arrest which would be consistent with our findings . In addition , the further stress of starvation was incorporated into the experimental paradigm . It would therefore be of interest to examine whether our results , which suggest that the neuropeptides released by clock cells antagonise reproductive quiescence , can be generalized under a variety of different unfavourable environmental conditions . Using a whole brain ex-vivo preparation , we observed that the IPCs responded to bath-applications of sNPF and PDF with increasing cAMP levels . The responses persisted when synaptic connections with the rest of the brain were inhibited by TTX , suggesting a direct effect on the IPCs of these neuropeptides ( Fig 4A–4H ) . Expression of sNPFR1 in the IPCs had already been described [53 , 54 , 57 , 68] . Interestingly , sNPFR1 couple to more than one Gα-protein subtype , since both excitatory ( through Gsα [77 , 78] ) and inhibitory effects ( through Goα [79] ) on cAMP levels have been documented . In addition , sNPF can signal by suppressing Ca2+ in some circadian clock clusters [42] and peptidergic PTTH neurons [84] . Here , we found that sNPF increased Ca2+ levels in the IPCs showing that it can also have activating Ca2+ effects . Similar multiple G-protein coupling has also been reported for the neurokinin-1 and -2 receptors [85 , 86] , as well as for the glucagon receptor in human atrial membranes [87] . The expression of the PDFR in the IPCs is less well-characterized . A previous study using fluorescent in situ hybridization reported prominent PDFR expression in the PI . However , it did not identify the PDFR positive cells [60] . Moreover , another study suggested that the cAMP responses evoked in the IPCs by PDF are not as robust as those registered in clock neurons [62] . Possibly , lower ligand efficacy might simply reflect less PDFR in these cells . The strong cAMP responses we observed after driving PDFR in han mutants under control of dilp2 ( p ) -Gal4 supports this conclusion ( Fig 6E–6H ) . On the other hand , there are 14 IPCs in the PI which show heterogeneous protein and neuropeptide composition [29 , 88] , and rhythmic electrophysiological parameters [33] . Thus , it is possible that , like the s-LNvs [89] , the responsiveness of the IPCs to PDF is also influenced by time of day . In any case , here we found significant cAMP increasing effects of PDF on the IPCs . When we applied sNPF and PDF together , the short-term ( 100–200 s ) response of the IPCs to the combined peptides was greater than the sum of their separate activities , pointing to a synergistic effect between the two molecules ( Fig 5A–5D ) . Similar interactions were not observed when sNPF was co-applied with other Drosophila neuropeptides , suggesting that the interaction of sNPF and PDF is specific . Nevertheless , the addition of TTX dampened the short-term response , suggesting that additional cells participate in the synergism between sNPF and PDF ( Fig 5E–5H ) . Interestingly , han mutants lacking PDFR did not respond to PDF , sNPF or sNPF+PDF co-application ( Fig 7A–7D ) . For sNPF , these results are puzzling , but since the rescue of PDFR in the IPCs only restored the response to PDF but not to sNPF , we conclude that sNPF does not act via PDFR . Still , we cannot completely exclude a cross-talk between PDFR and sNPFR1 . For instance , there is evidence that GPCRs can engage in homo- or hetero-oligomeric complexes , resulting in cooperativity ( reviewed in [90] ) . Alternatively , an interaction downstream of the receptors may occur , for example at the level of the signalosomes . PDFR and sNPFR associate with specific and different signalosomes that may have enhancing or opposing effects on cAMP levels [76–79 , 91] . However , further studies are required to investigate these possibilities using experimental settings that are closer to physiological conditions than bath applications of peptides . Interestingly , sNPF ( produced in the s-LNvs and in a subset of dorsal lateral neurons ) and PDF interact on clock neurons and set the phase of their cytosolic Ca2+ rhythms according to neuron cluster [42] . PDF primarily regulates Ca2+ rhythms in the LNd and DN3 clusters , while sNPF orchestrates that of the DN1 [42] . Thus , the two signaling pathways act dynamically to facilitate the right timing of circadian neuronal activities . Similar complex regulation could also influence the seasonal clock system . While the Pdf mRNA and protein do not cycle , it has been reported that PDF cycles at the nerve terminals with a circadian rhythm [39] . Thus , we presume that the daily rhythmic stimulation of the IPCs under normal summer conditions of warm days and long photoperiods maintains the expression of dILPs and suppresses dormancy . A simple model would have low temperatures and short photoperiods reducing the expression of PDF and sNPF from the sLNvs terminals , which might be expected to reduce IPC activation and enhance reproductive arrest levels . This could mean that the transcription/translation of these neuropeptides could be reduced under colder conditions , or alternatively , that their release was reduced from the sLNv terminal . This scenario might imply higher levels of PDF within the sLNv soma if it is sequestered there . Alternatively , the receptors for PDF and SNPF on the IPCs might be less sensitive to their ligands at lower temperatures , which would have the same effect on reproductive arrest . Future studies are required to examine the dynamic nature of PDF/SNPF and receptor expression under simulated winter conditions . PDF and sNPF–most likely from the s-LNvs can maintain D . melanogaster , originally a tropical species , in the reproductive state . However , it is intriguing that high-latitude Drosophila species such as D . montana , D . littoralis , D . ezoana and D . virilis lack PDF in the s-LNvs [92–95] . These species have a high incidence of reproductive arrest even under long-daylengths , an adaptation to the low temperatures even under summer photoperiods at these clines . For example , D . ezoana enters diapause when day-length falls below 16 hours [96] . Although we do not know whether the s-LNvs of the high-latitude species still express sNPF , we speculate that the lack of PDF-signaling to the IPCs of these species might facilitate the termination of the reproductive state under short-day condition , and induce ovarian arrest [97] . The role of the s-LNvs as a source of PDF and sNPF may thus provide the entry point into the neuronal mechanism that allows D . melanogaster to detect the environmental conditions that predispose them to reproductive dormancy .
Flies were reared at 23°C , 70% relative humidity , under 12-hour light/12-hour dark cycles ( LD 12:12 ) on cornmeal standard food . The following , previously described fly strains were used: Hu-S Dutch natural population [37] , han5304 [59] , UAS-PDFR [59] , gal1118 Gal4 enhancer trap [98] , R6-Gal4 [70] , UAS-Epac1camps [62] , UAS-hid [99] , UAS-Pdf [41] , and UAS-DenMark [75] . InsP3-Gal4 was a gift from Michael J . Pankratz [73] , dilp2 ( p ) -Gal4 ( p , precocious ) was a gift from Eric J . Rulifson [21] , UAS-2xsNPF [53] and UAS-sNPFR1-DN [54] were kindly provided Kweon Yu . Out of these lines , we newly combined han5304; dilp2 ( p ) -Gal4 , han5304;;UAS-Epac1camps , and han5304;UAS-Epac1camps/CyO;UAS-PDFR . The transgenic line Pdf-LexA , LexAop-CD4::GFP11/CyO;UAS-CD4::GFP1-10/TM6b for GRASP was a gift from François Rouyer . UAS-OrkΔ-C ( designated as UAS-Ork [72] ) and UAS-Na+ChBac [69] ) were gifts from Michael B . O’Connor . Pdf01 mutants were backcrossed for 8 generations to a natural wild-type ls-tim line from The Netherlands [37] . The following lines were ordered from Bloomington Drosophila Stock Center: Canton-S wild-type strain ( #1 ) , Crz1-Gal4 ( #51976 ) , Crz2-Gal4 ( #51977 ) , ElavC155-Gal4 ( designated as elav-Gal4 , #458 ) , Pdf-Gal4 ( #6900 ) , UAS-CD8-GFP ( BDSC #5137 ) , and Oregon-R ( #2376; used as control for experiments with han mutant ) . In order to use a control in which the specific transgene ( Gal4 or UAS ) was in the same condition of heterozygosis as in the experimental line , we have crossed all the parental lines to the w1118 strain ( generic genotype w1118;P-element/+ ) . In Drosophila , two allelic variants of the timeless circadian clock gene ( s-tim/ls-tim ) were found to significantly influence reproductive arrest: ls-tim allele promotes reproductive quiescence at every photoperiod [37 , 38] . Considering this modulatory effect , controls were generated according to the tim allele of the UAS and Gal4 strains used in the experiments ( see below for PCR timeless genotyping ) , with the help of w1118; s-tim ( gift from Matthias Schlichting ) and w1118; ls-tim ( from our lab ) lines that express either the s- or ls-tim isoform . All fly stocks and crosses used for dormancy assays were maintained at 23°C in LD 12:12 . Newly eclosed flies ( within 5 h post-eclosion ) were placed in plastic vials , and immediately subjected to 12°C in LD 8:16 . After 11 days , flies were killed in abs EtOH , and ovaries of females were dissected in PBS . Dormancy levels were scored by considering the absence of yolk deposition in the ovarian follicles [7] . These data were presented as the proportion of females with ovarian arrest among all the dissected individuals . On average , 5 replicates of n>60 flies were dissected for each genotypes . Flies were reared at 23°C in LD 12:12 , and newly eclosed individuals were placed in short days ( LD 8:16 ) at different temperatures ( 12 , 18 , 23°C ) for 11 days . Females were collected at ZT1 ( Zeitgeber Time 1 , 1 h after light-on ) , and immediately fixed in 4% paraformaldehyde ( PFA ) in PBS for 100 min at room temperature ( RT ) . After 3 washes in PBS , brains were dissected in ice-cold PBS , fixed in PFA 4% for 40 min at RT , and subsequently washed 6 times in PBS containing 0 . 3% Triton X-100 ( PBS-T ) . Next , samples were permeabilized in 1% PBS-T , followed by an overnight blocking step in 1% bovine serum albumin ( BSA ) in 0 . 3% PBS-T at 4°C . Afterwards , brains were incubated in primary antibody solution ( diluted in 0 . 1% BSA , 0 . 3% PBS-T ) for 3 days at 4°C . After 6 washes in 1% BSA in 0 . 3% PBS-T at RT , another blocking step was performed in 1% BSA in 0 . 3% PBS-T at 4°C , followed by hybridization with the secondary antibody ( diluted in 0 . 1% BSA , 0 . 3% PBS-T ) overnight at 4°C . The primary antibodies used in this study: mouse anti-GFP ( 1:500; Thermo Fisher Scientific ) , rabbit-anti-GFP ( 1:1000; A6455 , Invitrogen ) and anti-PDF ( 1:5000; mAb C7 , Developmental Study Hybridoma Bank , donated by Justin Blau ) , rabbit anti-DILP2 , anti-sNPF recognizing the sNPF propeptide ( for both 1:2000; Jan A . Veenstra [100 , 101] ) . The secondary antibodies used: Alexa Fluor 488 ( goat anti-mouse , 1:250 ) Cy3 ( goat anti-rabbit , 1:500 ) ( Thermo Fisher Scientific ) , DyLight488 ( goat-anti-rabbit 1:250 ) and DyLight649 ( goat anti-mouse , 1:250 ) ( Jackson ImmunoResearch , Dianova ) . Staining was visualized adopting either a semi-confocal ( Nikon Eclipse 80i equipped with a QiCAM Fast Camera using the Image ProPlus software ) or confocal microscope ( ZEISS LSM700 running ZEN Lite software or Leica SP8 ) . To real-time monitor cAMP or Ca2+ concentration changes in the IPCs , live optical imaging was performed using the genetically encoded cAMP sensor , Epac1-camps [62] and the genetically encoded Ca2+ sensor GCaMP3 . 0 [82] . Relying on the UAS-GAL4 binary system [102] , the sensors were expressed specifically in the IPCs ( dilp2 ( p ) >Epac1-camps or dilp2 ( p ) >GCaMP3 . 0 ) . Female flies , maintained at 25°C in LD 12:12 , were anesthetized on ice before brain dissections in cold hemolymph-like saline ( HL3 [103] ) and mounted at the bottom of a cap of a plastic Petri dish ( 35x10 mm , Becton Dickenson Labware , New Jersey ) in HL3 with the dorsal surface up . Brains were allowed to recover from dissection 15 min prior to imaging . Live imaging was conducted by using an epifluorescent imaging setup ( VisiChrome High Speed Polychromator System , ZEISS Axioskop2 FS plus , Photometrics CoolSNAP HQ CCD camera , Visitron Systems GmbH ) Visitron Systems GmbH ) using a 40x dipping objective ( ZEISS 40x/1 . 0 DIC VIS-IR ) . IPCs were brought into focus and regions of interest were defined on single cell bodies using the Visiview Software ( version 2 . 1 . 1 , Visitron Systems , Puchheim , Germany ) . Time-lapse frames were imaged with 0 . 2 Hz and 4x binning by exciting the CFP fluorophore of the cAMP sensor with 434/17 nm light or the GFP fluorescence of GCaMP3 with 488/10 nm light . For cAMP imaging , CFP and YFP emissions were separately detected using a Photometrics DualView2 beam splitter . After measuring baseline FRETs for ~100 s , substances were bath-applied drop-wise using a pipette , and imaging was performed for 1000 s . The neuropeptides used in this study were applied in a final concentration of 10 μM in 0 . 1% DMSO in HL3 . The water-soluble forskolin derivate NKH477 ( 10 μM , Sigma Aldrich ) or the cholinergic agonist carbamylcholine ( 1mM , Sigma Aldrich ) served as positive controls in cAMP and Ca2+ imaging , respectively , while HL3 alone with 0 . 1% DMSO was used as negative control . In the case of tetrodotoxin ( TTX ) treatments , brains were incubated for 15 min in 2 μM TTX in HL3 prior to imaging and substances were co-applied together with 2 μM TTX . Inverse Fluorescence Resonance Energy Transfer ( iFRET ) was calculated over time according to the following equation: iFRET = CFP/ ( YFP-CFP*0 . 357 ) [62] . Thereby , raw CFP and YFP emission data were background corrected; in addition , YFP data were further corrected by subtracting the CFP spillover into the YFP signal , which was determined as 0 . 357 ( 35 . 7% of the CFP signal ) . Next , iFRET traces of individual neurons were normalized to baseline and were averaged for each treatment . Finally , maximum iFRET changes were calculated for individual neurons to quantify and contrast response amplitudes of the different treatments . The following synthetic neuropeptides were used in this study: pigment dispersing factor ( PDF: NSELINSLLSLPKNMNDAa; Iris Biotech GmbH ) , short neuropeptide F-1 ( sNPF-1: AQRSPSLRLRFa; Iris Biotech GmbH ) , adipokinetic hormone ( AKH: pQLTFSPDWa , NovoPro Bioscience ) , allatostatin-C ( Ast-C: pEVRYRQCYFNPISCF , gift from Paul H . Taghert ) , Tachykinin 4 ( DTK-4: APVNSFVGMRa , gift from Paul H . Taghert ) , dFMRFamide 4 ( SDNFMRFa , gift from Paul H . Taghert ) . For Ca2+ imaging , brains expressed the GCamp3 . 0 sensor in the insulin producing cells ( dilp2 ( p ) >GCaMP3 . 0 ) [104] . The preparation of the brain samples was the same as in the case of cAMP imaging , and the same microscope was used with a modified setup , measuring GFP fluorescence without a beam splitter . The cholinergic agonist carbamylcholine ( 1 mM CCh ) was used to generate rapid Ca2+ increases ( Nakai et al . 2001 ) . After subtraction of background fluorescence , changes in fluorescence intensity were calculated for each ROI as Δ ( F/F0 ) = [ ( Fn—F0 ) /F0] x 100 with Fn as fluorescence intensity at time point n and F0 as the baseline fluorescence calculated prior to the application of the different substances to the brain . To ensure genetic homogeneity for the tim locus , between the experimental flies and their corresponding controls , all the strains used in this study were genotyped in order to identify the tim allele present in their genome ( summarized in S1 Table ) . The genomic DNA was extracted from individual adult females ( 10 flies per genotype ) by homogenizing them in 50 μl of extraction buffer ( Tris HCl pH = 8 . 2 10 mM , EDTA 2 mM , NaCl 25 mM ) ; after addition of 1 μl of Proteinase K ( 10 mg/ml ) samples were incubated at 37°C for 45 min , followed by 3 min at 100°C . The tim region containing the polymorphic site was amplified using a reverse primer ( 5’-AGATTCCACAAGATCGTGTT-3’ ) and two different forward primers ( ls-tim: 5’-TGGAATAATCAGAACTTTGA-3’; s-tim: 5’-TGGAATAATCAGAACTTTAT-3’ ) that allow selective amplification of the different tim alleles [37] . Larvae of the following genotypes: Act>sNPFR1-RNAi , Act>+ and +>sNPFR1-RNAi , were reared under standard conditions at 23°C and LD12:12 until eclosion . Newly eclosed female flies were collected and subsequently exposed to low ( 12°C ) or high ( 23°C ) temperature and short photoperiod ( 8h:16hL:D ) for 11 days . mRNA was isolated from whole bodies of 10 females . mRNA was reverse-transcribed with SuperScript II First-Strand Synthesis SuperMix ( Invitrogen ) . PCRs were performed on a CFX96 Touch Real Time PCR Detector System ( Bio Rad ) with GoTaq qPCR Master Mix ( Promega ) , using the following primers: sNPFR- F: 5′- CGACCATCAGATGCACCA -3′ , R: 5′-CGTCCGTCTCGTCTGTCC -3′; rp49 F: 5′- ATCGGTTACGGATCGAACAA-3′ , R: 5′- GACAATCTCCTTGCGCTTCT-3′ . The results are shown as relative expression ratios obtained with the 2-ΔΔCt method ± SEM . RP49 was used as reference . Results are shown in S2 Fig . Data were analysed with R statistical software ( version 3 . 0 . 1 , www . r-project . org ) and plotted using GraphPad Prism 6 software . In the case of normally distributed data ( Shapiro-Wilk normality test , p>0 . 05 ) , statistical significance was tested by one- or two-way ANOVA with post-hoc Tukey's HSD tests , while data that were not normally distributed were analyzed by Wilcoxon or Mann-Whitney test . In the case of multiple comparisons , raw p-values were further adjusted using Bonferroni correction , and these corrected p-values served as significance levels . When analyzing dormancy assays , all data were transformed to arcsine . For simplicity , figures in the Results section show untransformed data ( dormancy , % ) . | Diapause is a hormonally mediated process that allows insects to predict and respond to unfavourable conditions by altering their metabolism and behavior to resist the oncoming environmental challenges . In Drosophila melanogaster females a protective state of reproductive dormancy is induced by lower temperatures and shorter photoperiods that mimic the approach of winter . By genetically manipulating the circadian pacemaker s-LNvs cells , which express two neuropeptides , Pigment dispersing factor ( PDF ) and short Neuropeptide F ( sNPF ) , we were able to modulate levels of gonadal arrest . PDF and sNPF appear to act as antagonists to dormancy , as do the Drosophila insulin-like peptides ( dILPs ) that are expressed in the insulin producing cells ( IPCs ) . Indeed , we observe that the axonal projections from the s-LNvs appear to overlap with those from the IPCs implying that the clock cells signal to the IPCs . We confirm this possible communication by applying the two synthetic peptides to the IPCs and detecting a response in the IPC signal transduction pathway . We conclude that the clock neurons activate the IPCs via PDF and sNPF , which in turn release the dILPs , antagonise dormancy and lead to reproductive growth , thereby uncovering a neurogenetic circadian-overwintering axis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"diabetic",
"endocrinology",
"neuroscience",
"animals",
"hormones",
"animal",
"models",
"fluorophotometry",
"physiological",
"processes",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"diapause",
"molecular",
"biology",
"techniques",
"neuropeptides",
"drosophila",
"research",
"and",
"analysis",
"methods",
"spectrum",
"analysis",
"techniques",
"fluorescence",
"resonance",
"energy",
"transfer",
"insulin",
"animal",
"cells",
"animal",
"studies",
"neurochemicals",
"proteins",
"hyperexpression",
"techniques",
"endocrinology",
"molecular",
"biology",
"insects",
"peptide",
"hormones",
"spectrophotometry",
"arthropoda",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"cellular",
"neuroscience",
"eukaryota",
"cell",
"biology",
"post-translational",
"modification",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"signal",
"peptides",
"organisms"
]
| 2019 | Peptidergic signaling from clock neurons regulates reproductive dormancy in Drosophila melanogaster |
In Caenorhabditis elegans the Toll-interleukin receptor domain adaptor protein TIR-1 via a conserved mitogen-activated protein kinase ( MAPK ) signaling cascade induces innate immunity and upregulates serotonin ( 5-HT ) biosynthesis gene tph-1 in a pair of ADF chemosensory neurons in response to infection . Here , we identify transcription factors downstream of the TIR-1 signaling pathway . We show that common transcription factors control the innate immunity and 5-HT biosynthesis . We demonstrate that a cysteine to tyrosine substitution in an ARM motif of the HEAT/Arm repeat region of the TIR-1 protein confers TIR-1 hyperactivation , leading to constitutive tph-1 upregulation in the ADF neurons , increased expression of intestinal antimicrobial genes , and enhanced resistance to killing by the human opportunistic pathogen Pseudomonas aeruginosa PA14 . A forward genetic screen for suppressors of the hyperactive TIR-1 led to the identification of DAF-19 , an ortholog of regulatory factor X ( RFX ) transcription factors that are required for human adaptive immunity . We show that DAF-19 concerts with ATF-7 , a member of the activating transcription factor ( ATF ) /cAMP response element-binding B ( CREB ) family of transcription factors , to regulate tph-1 and antimicrobial genes , reminiscent of RFX-CREB interaction in human immune cells . daf-19 mutants display heightened susceptibility to killing by PA14 . Remarkably , whereas the TIR-1-MAPK-DAF-19/ATF-7 pathway in the intestinal immunity is regulated by DKF-2/protein kinase D , we found that the regulation of tph-1 expression is independent of DKF-2 but requires UNC-43/Ca2+/calmodulin-dependent protein kinase ( CaMK ) II . Our results suggest that pathogenic cues trigger a common core-signaling pathway via tissue-specific mechanisms and demonstrate a novel role for RFX factors in neuronal and innate immune responses to infection .
Innate immunity is an integral part of the stress response program in which the host activates a range of defense genes to enhance the chance of survival against internal and environmental threats . In mammals , signals associated with pathogenic microbes trigger Toll-like receptors to recruit Toll-interleukin receptor ( TIR ) domain adaptor proteins , thereby forming scaffolds with downstream signaling cascades leading to transcriptional upregulation of defense genes . A growing body of evidence indicates that classical immune proteins including Toll-like receptors and TIR domain adaptor proteins are expressed in the developing and mature brain in mammals [1] , [2] . It has been proposed that certain common molecular mechanisms may function in neurons and non-neuronal tissues to induce physiologically distinct responses to aversive cues [1] , [2] , [3] . Except a few cases , the gene targets of immune factors in neurons are not known and it is unclear whether those immune signaling cascades are differentially regulated in neurons and non-neuronal tissues . Consequently , identification of upstream regulators and downstream effectors of conserved core immune signaling pathways may provide insights into the regulation of the immunity as well as the regulation of neural plasticity . Our laboratory has focused on genetic dissection of environment-dependent transcriptional regulation of the tph-1 gene , encoding the rate-limiting serotonin ( 5-HT ) biosynthesis enzyme tryptophan hydroxylase , in the nematode Caenorhabditis elegans . Previously , we showed that tph-1 expression in a pair of ADF chemosensory neurons in the head sensory organ Amphid is modulated by two layers of transcriptional regulation according to growth conditions: signaling through the OCR-2/OSM-9 TRPV channel turns on the basal tph-1 expression under optimal growth conditions , and aversive growth conditions further upregulate tph-1 expression independently of OCR-2/OSM-9 [4] , [5] . Work from several laboratories suggests that tph-1 expression in the ADF neurons responds to pathogenic food . C . elegans feeds on bacteria and is killed by a large number of pathogenic microbes in its natural environment [6] . In an elegant study , it showed that feeding worms with the human opportunistic pathogen Pseudomonas aeruginosa PA14 triggers upregulation of tph-1 in the ADF neurons leading to aversive learning and avoidance behavior [7] . A subsequent study indicated that the TIR-domain adaptor protein TIR-1 , which was initially identified as an upstream regulator of a conserved mitogen-activated protein kinase ( MAPK ) signaling pathway in the innate immunity [8] , is required for PA14-induced tph-1 upregulation and PA14 avoidance behavior [9] . However , the C . elegans genome lacks a homolog of nuclear factor-kappaB ( NF-κB ) , the major transcriptional activator of the mammalian innate immunity [10] . In addition , deletion of the sole C . elegans Toll receptor gene tol-1 did not affect the intestinal immunity [11] or tph-1 expression [7] . These observations suggest that the TIR-1 signaling cascade may involve evolutionarily more ancient upstream players and downstream transcription factors . Activating transcription factor ( RFX ) transcription factors were first identified in human subjects of bare lymphocyte syndrome , a hereditary immunodeficiency disease , and are required for the expression of the major histocompatibility complex class II ( MHC II ) genes [12] . RFX proteins bind to the X-box motif on the MHC II promoters and interact with cAMP response element-binding ( CREB ) protein and other cofactors to form a higher order “enhanceosome” , which then recruits the non-DNA-binding transcriptional activator CIITA to turn on MHC II expression [13] . RFX factors have since been identified in diverse eukaryotic species [14] , [15] , [16] and are expressed broadly in neuronal and non-neuronal cells in animals , suggesting additional roles for RFX factors in biological processes of multiple tissues . Studies of the sole C . elegans RFX factor daf-19 have uncovered its role in the development of dendritic cilia of sensory neurons [17] , [18] . Subsequent studies found RFX factors regulating ciliogenesis in Drosophila [19] and mouse [20] , demonstrating one aspect of RFX function conserved across phyla . In this paper , we identified DAF-19 as a key transcriptional regulator of tph-1 in the ADF neurons and intestinal antimicrobial genes in C . elegans . We found that , analogous to the RFX-CREB interaction for MHC II expression in human immune cells , DAF-19 concerts with ATF-7 , a member of activating transcription factor ( ATF ) /CREB superfamily of transcription factors , acting downstream of the TIR-1 signaling cascade to control transcriptional responses to pathogenic bacterial food in C . elegans . We show that the TIR-1-DAF-19/ATF-7 pathway is differentially regulated to induce tph-1 upregulation and intestinal immunity in response to P . aeruginosa PA14 . Thus , our data suggest that pathogenic signals may trigger a common core signaling pathway via cell-specific mechanisms and a RFX transcription factor acts in an ancient host to regulate 5-HT biosynthesis and the innate immunity .
We carried out a forward genetic screen to identify components underling aversive environment-induced tph-1 upregulation in C . elegans . The levels of tph-1 expression in identified neurons in living C . elegans can be estimated by quantifying fluorescence intensity of a green fluorescence protein ( GFP ) driven by the tph-1 promoter ( tph-1::gfp ) [4] . A pair of ADF neurons is the only chemosensory neurons producing 5-HT in a hermaphrodite C . elegans [21] . Each ADF neuron projects a single dendrite to the tip of the nose where the ciliated sensory endings are exposed to the external environment and its axon extends to the nerve ring , the brain of C . elegans [22] . We started with a strain expressing a stably chromosomally integrated tph-1::gfp transgene in ocr-2 ( yz5 ) mutant background , in which tph-1::gfp expression in the ADF neurons is visible under aversive growth conditions but not under optimal growth conditions , providing a visual assay for environment-dependent changes in tph-1 expression [5] . We isolated mutagenized worms with enhanced ADF tph-1::gfp under optimal growth conditions , and analyzed the mutants in the ocr-2 background as well as after the ocr-2 mutation being outcrossed . yz68 is one of the mutants retrieved from the screen . Through single nucleotide polymorphism-based ( SNP ) mapping , RNA-interference ( RNAi ) -mediated inactivation of candidate genes in the mapped contig , and sequencing the yz68 mutant genome , we identified a nucleic acid change predicting a substitution of cysteine426 by tyrosine ( C426Y ) in the fourth ARM motif of the HEAT/Arm repeat region of the TIR-1 protein ( Figure 1A ) . Several experimental data suggest that the C426Y substitution causes TIR-1 constitutive activation . First , whole mount anti-5-HT-antibody staining showed that ADF 5-HT immunoreactivity in tir-1 ( yz68 ) ;ocr-2 double mutants was elevated compared to the ocr-2 single mutant ( Figure 1D ) . As 5-HT is being secreted , 5-HT immunostaining does not fully reflect the rate of 5-HT biosynthesis . With this caveat in mind , the results suggest that tir-1 ( yz68 ) enhanced 5-HT in the ADF neurons . Second , transgenic expression of tir-1 ( yz68 ) cDNA under the gpa-13 promoter ( Pgpa-13::tir-1 ( yz68c ) ) , which is expressed in ADF , AWC and ASH sensory neurons in the head region , recapitulated tph-1::gfp upregulation ( Figure 1B ) . Third , RNAi of tir-1 in tir-1 ( yz68 ) ;ocr-2 mutants blocked the tph-1 upregulation ( Figure 1C ) . The tir-1 ( ok1052 ) mutation , which causes mixed gain- and loss-of-function tir-1 phenotypes in the AWC neuron development [23] but does not affect the innate immunity [24] , caused only a modest increase in ADF tph-1::gfp ( Figure 1C ) . Collectively , these data suggest that the C426Y substitution alters a site critical for TIR-1 activation in the ADF neurons . Unlike the ocr-2 mutation , tir-1 ( tm3036 ) loss-of-function ( lf ) mutants and RNAi of tir-1 do not lead to a dramatic reduction in ADF tph-1::gfp ( Figure 1C ) . Likewise , inactivation of TIR-1 downstream MAPKKK nsy-1 or MAPKK sek-1 by RNAi and deletion , respectively , did not downregulate tph-1::gfp , although RNAi of nsy-1 did abolish tph-1::gfp upregulation by tir-1 ( yz68gf ) ( Figure S1 ) . These observations are consistent with a published work [9] . These data suggest that TIR-1 signaling is not a major regulator of basal tph-1 expression in the ADF neurons under optimal growth conditions . Previously , we identified that a number of aversive conditions may induce tph-1 upregulation in the ADF neurons [5] . We asked whether tir-1 signaling specifically mediates the response to pathogenic bacteria , or it is responsible for tph-1 upregulation under all aversive conditions . We first analyzed the intensity of tph-1::gfp in the ADF neurons in tir-1 ( lf ) and tir-1 ( gf ) mutants fed with the pathogenic P . aeruginosa strain PA14 . Following feeding on PA14 for 6 hr , ADF fluorescence in WT animals was ∼1 . 4-fold higher than their sibling on OP50 , but tir-1 ( lf ) mutants failed to upregulate tph-1::gfp under the same conditions ( Figure 2A ) , consistent with prior reports [7] , [9] . The tir-1 ( yz68gf ) mutant also did not exhibit a significant increase in tph-1::gfp following PA14 feeding , suggesting that the pathogen signals cannot further enhance tir-1 ( yz68gf ) protein activity ( Figure 2A ) . We next tested if TIR-1 function is required for tph-1 upregulation during dauer formation . Under the conditions of starvation , high growth temperature and high levels of pheromones , C . elegans develops into a stress-resistant dauer larva through a series of cellular and physiological remodeling and turns on a battery of stress genes [25] . We previously showed that WT and ocr-2 mutant worms upregulated ADF tph-1::gfp when they entered the dauer stage [5] . We therefore induced tir-1 ( lf ) and tir-1 ( gf ) mutants to form dauers by treating the worms with dauer pheromones . We observed ADF tph-1::gfp upregulation in both tir-1 ( tm3036lf ) and tir-1 ( tm3036lf ) ;ocr-2 double mutant dauers as compared with corresponding L4-stage animals ( Figure 2B ) . ADF tph-1::gfp was increased in tir-1 ( yz68gf ) mutants during dauer formation and this increase was more evident in the tir-1 ( yz68gf ) ;ocr-2 double mutant dauers ( Figure 2B ) . Thus neither TIR-1 deficiency nor TIR-1 hyperactivation can block tph-1::gfp upregulation induced by dauer formation . We previously showed that mutations that alter the morphology of dendritic cilia of ADF neurons cause tph-1 upregulation [5] . We therefore tested whether aberrant cilia trigger TIR-1 leading to tph-1::gfp upregulation . If this were the case , then we can expect that inactivation of tir-1 blocks tph-1::gfp upregulation in cilial mutants . Mutations of daf-6/Patched , which is expressed in the glia ensheathing the dendritic cilia of the chemosensory neurons [26] , and the Intraflagellar Transport ( IFT ) gene dyf-1 that is essential for cilia formation [27] conferred ADF tph-1::gfp upregulation [5]; however , inactivation of tir-1 did not block tph-1::gfp upregulation in daf-6 ( Figure 2C ) or dyf-1 mutants ( Figure S2 ) . Furthermore , a triple mutant of daf-6 , tir-1 ( yz68gf ) ;ocr-2 showed higher ADF tph-1::gfp than daf-6;ocr-2 and tir-1 ( yz68gf ) ;ocr-2 double mutants ( Figure 2D ) . Together these data suggest that the ADF neurons can detect and discriminate multiple aversive cues , and indicate that TIR-1 is selectively involved in the pathogen signaling transduction pathway . To identify the effectors of TIR-1 signaling , we carried out a suppressor screen for mutants that abrogate tph-1 expression in tir-1 ( yz68gf ) mutants . Using a combination of SNP mapping , non-complementation tests and sequencing the mutant genomes , we identified that two mutations , yz69 and yz70 , are allelic to the daf-19 gene , encoding the sole C . elegans ortholog of the RFX transcription factors ( Figure 3A ) . Subsequent experiments with our alleles and the previously existing daf-19 ( m86 ) -null mutation revealed that DAF-19 function is required not only for tir-1 ( yz68gf ) to upregulate tph-1::gfp , but also for ADF tph-1::gfp expression under optimal growth conditions ( Figure 3B ) , during dauer formation ( data not shown ) and in aberrant cilia backgrounds ( Figure 3E; Figure S2 ) . The reduced tph-1::gfp in daf-19 mutants was fully rescued by transgenic expression of WT daf-19 genomic sequence ( Figure 3B ) or daf-19 cDNA driven by the gpa-13 promoter ( Figure 3D ) . An implicit concern was that daf-19 deficiency alters ADF cell fates . To directly analyze the effect of daf-19 deficiency on tph-1 expression , we inactivated daf-19 by RNAi after 5-HT phenotypes established . Figure 3C shows that 100% of larval stage 1 ( L1 ) worms expressed tph-1::gfp prior to RNAi treatment , 33% of the animals lost ADF GFP after 24 hr RNAi treatment , and by 48 hr , 100% of the animals showed no GFP in the ADF neurons . DAF-19 is required for the expression of cilia IFT components [17] . Although we showed that aberrant cilia induced ADF tph-1::gfp upregulation [5] ( Figure S2 ) , complete lacking cilia could inhibit tph-1 expression . To rule out this possibility , we used lipophilic dye DiI staining to examine the dendritic cilia morphology of the chemosensory neurons in worms treated with daf-19 RNAi . We observed that RNAi of daf-19 eliminated ADF tph-1::gfp prior to a detectable change in the cilia ( Figure 3C ) . We investigated whether the reduced tph-1 expression is a secondary consequence of reduced ocr-2 and tir-1 expression in daf-19 mutants . daf-19 ( yz69 ) ( Figure 3Hii ) and daf-19 ( m86 ) -null ( not shown ) mutants expressed a GFP reporter for ocr-2 ( ocr-2::gfp ) in ADF and other chemosensory neurons indistinguishable from WT animals . Diminished tph-1::gfp expression also cannot be ascribed to reduced tir-1 expression , as transgenic expressing tir-1 ( yz68 ) cDNA by the gpa-13 promoter failed to increase ADF tph-1::gfp in daf-19 mutants ( Figure 3G ) , although the Pgpa-13::daf-19c transgene did ( Figure 3D ) , indicating that the gpa-13 promoter is expressed in daf-19 mutants but that the tir-1 ( yz68gf ) protein cannot stimulate tph-1 expression in the absence of DAF-19 function . To assess whether daf-19 deficiency abolishes all 5-HT phenotype genes , we analyzed the expression of cat-1 , encoding the vesicular monoamine transporter required for 5-HT synaptic release [28] . GFP-tagged CAT-1 was expressed and localized properly in the ADF neurons of daf-19 mutants ( Figure 3Iii ) . Thus , diminished tph-1::gfp in daf-19 mutants cannot be attributed to altered ADF cell fate . Because daf-19 appeared to be required for tph-1 expression controlled by multiple mechanisms , we hypothesized that DAF-19 may concerts with other transcriptional regulators to confer specificity to particular signaling pathway activation . In light of the partnership of RFX and CREB proteins for MHC II expression , we further hypothesized that an analogous mechanism might underscore transcriptional responses to pathogenic signals in C . elegans . ATF-7 , an ortholog of the mammalian ATF2/ATF7/CREB5 family of transcription factors [29] , has been implicated as a transcriptional repressor of antimicrobial genes in the intestine; activation of PMK-1 p38 MAPK de-represses ATF-7 , hence upregulating the antimicrobial genes [30] . We therefore analyzed tph-1::gfp expression in atf-7 ( gf ) and atf-7 ( lf ) mutants . atf-7 ( gf ) mutation presumably constitutively represses TIR-1 signaling targets . atf-7 ( gf ) did not downregulate ADF tph-1::gfp under optimal growth conditions ( Figure 4A ) , as seen in mutants with reduced tir-1 , TIR-1 downstream nsy-1/MAPKKK and sek-1/MAPKK ( Figure 1C , Figure S1 ) [9] . Like tir-1 ( lf ) mutants , atf-7 ( gf ) , as well as daf-19 mutants , failed to upregulate ADF tph-1::gfp following 6 hr feeding on PA14 ( Figure 2A , Figure 4B ) . By contrast , both atf-7 ( gf ) and atf-7 ( lf ) mutants upregulated ADF tph-1::gfp during dauer formation ( Figure 4C ) . These data suggest that ATF-7 confers the specificity to upregulate tph-1 in response to pathogenic bacteria . We tested further whether constitutive repression function of atf-7 ( gf ) could suppress tph-1::gfp upregulation by tir-1 ( yz68gf ) . We crossed the Pgpa-13::tir-1 ( yz68c ) transgene into atf-7 ( gf ) mutants . The Pgpa-13::tir-1 ( yz68c ) transgene conferred a significantly increase of ADF tph-1::gfp as tested in ocr-2 mutant background but not in ocr-2;atf-7 ( gf ) double mutants ( Figure 4A ) . Thus , tph-1 upregulation induced by PA14 and tir-1 ( yz68gf ) requires both DAF-19 and ATF-7 . daf-19 is expressed broadly in neurons but also in the intestine ( Figure S3 ) [31] , [32] . We used epistasis analysis to investigate the role of DAF-19 in TIR-1-mediated innate immunity . When incubated on a lawn of PA14 , tir-1 ( yz68gf ) mutants exhibited enhanced resistance to killing by PA14 , in contrast to the heightened susceptibility of tir-1 ( lf ) mutants ( Figure 5A ) . daf-19 ( m86 ) and daf-19 ( yz70 ) mutants exhibited heightened susceptibility to PA14 compared to WT ( Figure 5B ) . Similarly , RNAi of daf-19 enhanced the susceptibility to PA14 compared to mock RNAi ( Figure 5C ) . Transgenic expression of daf-19 cDNA in the intestine partially rescued PA14 resistance in daf-19 mutants ( Figure 5D ) . daf-19 ( m86 ) -null mutation suppressed enhanced immune resistance of the tir-1 ( yz68gf ) mutants , showing that the enhanced pathogen resistance of tir-1 ( yz68gf ) mutants also requires DAF-19 function ( Figure 5E ) . However , the daf-19 ( m86 ) ;tir-1 ( yz68gf ) double mutants did not display heightened pathogen susceptibility as seen in daf-19 ( m86 ) single mutants , suggesting additional transcriptional regulators involved in the immunity induced by TIR-1 activation . We tested the functional relationship between DAF-19 and ATF-7 in the innate immunity . Both atf-7 ( gf ) and atf-7 ( lf ) mutants exhibited heightened susceptibility to killing by PA14 , although the immunodeficiency phenotype of atf-7 ( lf ) mutants is weaker [30] . If ATF-7 and DAF-19 function in parallel , we could expect a stronger immunodeficiency phenotype in a double mutant of atf-7 ( lf ) and daf-19 relative to the single mutants . Contrary to the prediction , atf-7 ( lf ) ;daf-19 double mutants displayed a survival rate on PA14 comparable to the atf-7 ( lf ) single mutant ( Figure 5F ) . This result is more consistent with the model in which DAF-19 regulates ATF-7 targets in the immune system . The exact detoxification mechanisms of the C . elegans immunity are not known . To validate the role of DAF-19 in the innate immunity , we made use of the fact that bacterial infections cause intestine to induce the transcription of a battery of secretory proteins that are thought to produce antimicrobial effects [33] , [34] . Transcriptional regulation of those candidate antimicrobial genes has been used as an assay for genetic delineation of C . elegans immune pathways [35] , [36] . For example , the atf-7 ( gf ) allele , as well as a number of tir-1 ( lf ) alleles , were identified based on the diminished intestinal expression levels of a GFP reporter for the ShK-like toxin peptide gene T24B8 . 5 ( T24B8 . 5::gfp ) , and atf-7 ( lf ) alleles were identified as suppressors of the diminished T24B8 . 5::gfp of atf-7 ( gf ) [30] . We therefore analyzed the same integrated T24B8 . 5::gfp reporter in tir-1 and daf-19 mutants . On a lawn of standard bacterial food E . coli OP50 , T24B8 . 5::gfp intensity in the intestine of tir-1 ( yz68 ) was substantially enhanced relative to WT animals ( Figure 6B , 6I ) , further validating the constitutive activity of tir-1 ( yz68gf ) . By contrast , intestinal T24B8 . 5::gfp was markedly reduced in daf-19 mutants as in atf-7 ( gf ) and tir-1 ( lf ) mutants ( Figure 6C , 6E , 6I ) . Thus , DAF-19 deficiency results in downregulation of an immune gene marker regulated by TIR-1 and ATF-7 . Similar to the effect of the daf-19 mutation on tir-1 ( yz68gf ) immunity ( Figure 5E ) , the daf-19 ( m86 ) and daf-19 ( yz69 ) mutations diminished the increased T24B8 . 5::gfp expression in tir-1 ( yz68gf ) mutants , although the GFP level in the daf-19; tir-1 ( yz68gf ) double mutants was higher compared to the daf-19 single mutants ( Figure 6D , 6I ) . We also detected an increase in T24B8 . 5::gfp in the atf-7 ( lf ) intetine relative to WT animals; two tested daf-19 alleles both reversed the increased T24B8 . 5::gfp in atf-7 ( lf ) mutants ( Figure 6F , 6G , 6H , 6I ) . To confirm the role for DAF-19 in the regulation of candidate antimicrobial genes , we used quantitative real-time RT-PCR ( qPCR ) to measure the expression of endogenous T24B8 . 5 as well as three other ATF-7-regulated candidate antimicrobial genes . Every tested antimicrobial gene was reduced in two tested daf-19 alleles compared to WT animals ( Figure 7B ) . In contrast , the message levels of these antimicrobial genes were increased in tir-1 ( yz68gf ) mutants ( Figure 7B ) . daf-19 mutation diminished the increases of those antimicrobial genes in tir-1 ( yz68gf ) mutants ( Figure 7B ) , similar to that seen with T24B8 . 5::gfp in the intestine . As the controls , we analyzed atf-7 ( lf ) and atf-7 ( gf ) mutants . We observed that the message levels of K08D8 . 5 and F35E12 . 5 were significantly reduced , and C17H12 . 8 elevated in the atf-7 ( lf ) mutant ( Figure 7A ) , as previously reported [30] . Consistent with our observation of enhanced T24B8 . 5::gfp in atf-7 ( lf ) intestine , we found T24B8 . 5 message level increased in atf-7 ( lf ) mutants ( Figure 7A ) . We did observe dramatically reduced message levels of all tested genes in atf-7 ( gf ) and tir-1 ( lf ) mutants ( Figure 7A , 7B ) , as previously reported [30] , [36] . Our data thus far indicated that DAF-19 is required for TIR-1 signaling to upregulate tph-1 in the ADF neurons and candidate antimicrobial genes in the intestine . We wished to determine whether DAF-19 regulates every TIR-1 target , or it selectively mediates TIR-1 regulation of pathogen inducible genes . It has been well established that , TIR-1 specifies asymmetrical expression of the olfactory receptor STR-2 in one of two AWC olfactory neurons during the development [23] . We observed that neither AWC neuron expressed str-2::gfp in the tir-1 ( yz68gf ) mutant , as seen in mutants with excessive TIR-1 activity [23] . However , daf-19 mutants did not exhibit the AWC phenotype seen in tir-1 ( lf ) mutants ( Figure S4 ) . Thus , DAF-19 is critical for TIR-1 signaling to induce transcriptional responses to pathogenic bacteria in the 5-HT neurons and the intestine , but is not required for TIR-1 to regulate neural development . The finding of the shared transcriptional effectors of TIR-1 signaling in the ADF neurons and intestine raised an intriguing question as to whether pathogenic bacterial signals trigger the neurons and immune cells in the same manner in C . elegans . Although little is known about how the worm senses the presence of pathogens and relays the signals to TIR-1 , the protein kinase D DKF-2 is thought to promote transcriptional responses to PA14 by activating the TIR-1 signaling pathway [37] . However , we found that the dkf-2 ( pr3 ) -null mutation did not suppress tph-1::gfp upregulation by tir-1 ( yz68gf ) or PA14 ( Figure 8A ) . Previously , we showed that a gain-of-function mutation in the calcium/calmodulin-dependent protein kinase II ( CaMKII ) UNC-43 upregulates ADF tph-1::gfp [4] . We hypothesized that UNC-43 could be a component of the TIR-1 signaling pathway in the ADF neurons , similar to its involvement in TIR-1-mediated AWC development [23] . To test this hypothesis , we analyzed tph-1::gfp in unc-43 ( lf ) ;tir-1 ( yz68 ) double mutant . We observed that unc-43 ( lf ) abrogated ADF tph-1::gfp upregulation in tir-1 ( yz68 ) mutants ( Figure 8B ) . Furthermore , PA14 failed to induce ADF tph-1::gfp upregulation in both the unc-43 ( lf ) single and unc-43 ( lf ) ;tir-1 ( yz68gf ) double mutants ( Figure 8A ) . We previously showed that unc-43 ( lf ) does not block ADF tph-1::gfp upregulation during dauer formation [5] . These observations together suggest that UNC-43 selectively regulates TIR-1-mediated tph-1 upregulation and that tir-1 ( yz68gf ) cannot bypass UNC-43 function . In cultured mammalian neurons , CaMKII activation alters the subcellular localization of signaling components to initiate cellular responses [38] . UNC-43 has been shown enriched in postsynaptic sites of a number of neuronal types and was co-immunopecipitated with TIR-1 when co-expressed in cultured mammalian cells [23] , [39] . If UNC-43 regulates TIR-1 subcellular distribution , then it is possible that the subcellular distribution of the TIR-1 ( yz68gf ) protein also depends on UNC-43 . We tested this idea by comparing the subcellular distribution of GFP-tagged TIR-1 and TIR-1 ( yz68gf ) proteins expressed in chemosensory neurons by the gpa-13 promoter in WT and unc-43 ( lf ) backgrounds . In WT animals , TIR-1 ( WT ) ::GFP was observed in punctate structures in the axons around nerve ring , with more diffused fluorescence seen in the cell bodies ( Figure 8C ) . By contrast , TIR-1 ( yz68gf ) ::GFP displayed increased punctate structures in the axons as well as in the cell bodies ( Figure 8D ) . Importantly , TIR-1 ( WT ) ::GFP and TIR-1 ( yz68gf ) ::GFP punctate structures were reduced in unc-43 ( lf ) mutants ( Figure 8C , 8D ) . Based on these observations , we speculate that the C426Y substitution of TIR-1 ( yz68gf ) alters the protein conformation , thereby enhancing its interaction with UNC-43 in particular cellular compartments where TIR-1 interacts with MAPK signaling components , but that TIR-1 ( yz68gf ) cannot efficiently interact with its downstream components in the absence of UNC-43 . Collectively , these results suggest that PA14 triggers distinct mechanisms to activate TIR-1 signaling to induce DAF-19/ATF-7 targets in the ADF neurons and the intestine . Our data indicate that UNC-43 is required for TIR-1 signaling in ADF , and DKF-2 is not .
Parallel to the C . elegans NSY-1/MAPKKK-SEK-1/MAPKK-PMK-1/p38 MAPK pathway , the ASK-1/MAPKKK to p38 MAPK pathway regulates mouse innate immunity [40] . While members of the NF-κB family of transcription factors are the major effectors in mammalian innate immune responses , the ASK-1-p38 innate immune pathway is independent of NF-κB [40] . Since C . elegans lacks NF-κB , this MAPK pathway was proposed to act via effectors evolutionarily more ancient than NF-κB in the host defense systems [40] . In this study , we identify that DAF-19 RFX is a transcription factor downstream of this core innate immune pathway in C . elegans . Our genetic analyses suggest that DAF-19 concerts with ATF-7 , a member of the ATF/CREB superfamily of transcription factors , to upregulate tph-1 in the ADF neurons and antimicrobial factors in the intestine in response to the pathogenic P . aeruginosa strain PA14 , reminiscent of the RFX-CREB partnership for human MHC II gene expression . Several lines of evidence point to RFX factors as an ancient mechanism for enhancing survival under aversive conditions , one predating the divergence of stress responses and immunity . First , RFX is a regulator of cell cycle in a nutrient sensing pathway in Schizosaccharomyces pombe [41] and an effector of the DNA damage and replication checkpoint pathway essential for Saccharomyces cerevisiae survival under replicative stress [42] . Second , biochemical experiments have identified a role for RFX factors in RAS signaling-regulated transcription in mammalian epithelial cells [43] . Third , in addition to mediate transcriptional responses to PA14 , daf-19 controls the decision to enter the stress-resistant dauer stage that is specialized for enduring aversive environmental conditions [17] . Fourth , 5-HT is perhaps one of the most ancient mechanisms of stress responses conserved across phyla [44] . Pharmacological and biophysical experiments have long demonstrated that 5-HT biosynthesis in mammals is highly sensitive to environmental conditions and can be upregulated by a variety of metabolic , psychological and physical stressors in a region-specific manner [45] , [46] , although systematic dissection of 5-HT biosynthesis in the mammalian brain has not been feasible . Our results raise the possibility that the mechanisms that regulate 5-HT biosynthesis and certain aspects of innate immunity may be interrelated . What could be the mechanism by which DAF-19 regulates tph-1 and those antimicrobial genes ? Earlier studies of RFX regulation of human MHC II genes provided evidence that RFX factors regulate gene expression indirectly through recruiting additional transcriptional regulators to the promoters [13] . In a recent study , RFX factors were shown to protect promoters against epigenetic silencing by DNA methylation through recruiting chromatin-remodeling factors [47] . Thus one plausible mechanism could be that DAF-19 and ATF-7 bind to a promoter element shared in common among the pathogen-inducible genes , and infection induces the TIR-1 signaling cascade leading to phosphorylation of ATF-7 [30] and transcriptional activation of the targets . However , genomewide X-box motif search identified hundreds of candidate genes , only a few antimicrobial genes are among them [48] and no X-box can be recognized in the promoter of the tph-1 , T24B8 . 5 , C17H12 . 8 , K08D8 . 5 or F35E12 . 5 genes . Thus , DAF-19 is likely to bind non-consensus X-boxes on the promoters via co-regulators . Alternatively , DAF-19 may modulate chromatin structure , facilitating the binding of other transcriptional regulators to the promoters . We favor the binding-via-cofactor model because our data thus far suggest that DAF-19 is involved in multiple environment-dependent regulations of tph-1 expression , whereas ATF-7 selectively regulates the response to PA14 . We speculate that DAF-19 interacts with distinct cofactors that are regulated by distinct environmental cues ( Figure 9A ) . Further experiments are required to determine whether DAF-19 directly interacts with ATF-7 . In addition , the possibility that DAF-19 and ATF-7 regulate yet unidentified transcription factor ( s ) , which in turn regulate pathogen-inducible genes , cannot be excluded . It is perhaps interesting to note that our two daf-19 alleles both are located in the predicted dimerization ( DIM ) domain , suggestive for the importance of protein-protein interaction in DAF-19 function . While the exact mechanism of DAF-19 on the target gene promoters remains to be elucidated , our data showed that DAF-19 and ATF-7 regulate common immune gene markers . Remarkably , by comparing the list of TIR-1 signaling targets identified by microarray analysis [36] with the database of several independent expression profiling of daf-19 mutants [49] , [50] , [51] , we found that 102 out of 215 TIR-1 gene targets are among those differentially expressed in daf-19 mutants; one of them is C17H12 . 8 , which was confirmed by our qPCR . We showed that daf-19 mutations can suppress increased expression of immune gene markers in tir-1 ( yz68gf ) mutants . However , daf-19-null;tir-1 ( yz68gf ) double mutants did not display reduced expression of those immune genes as seen in the daf-19 single mutants , judging by T24B8 . 5::gfp in living worms and qPCR . Consistent with the gene expression analyses , the daf-19-null mutation blocked the enhanced resistance to killing by PA14 in tir-1 ( yz68gf ) mutant but the daf-19;tir-1 ( yz68gf ) double mutant did not exhibit heightened susceptibility as daf-19 single mutants did . These observations may be consistent with the model in which DAF-19 and ATF-7 interact with additional transcription factor ( s ) that also contribute to the regulation of those antimicrobial genes . There is converging evidence from neurobiology and immunology suggesting that the brain immune privilege is not absolute [52] . Internal and external pathogenic products can infiltrate into the CNS and there is extensive bi-directional communications between the CNS and immune systems [53] , [54] . The finding of a large number of classical immune proteins in the CNS has led to the idea that common molecular mechanisms may be involved in neuronal and immune responses to pathogenic signals [1] , [2] , [53] . Our identification of the TIR-1-DAF-19/ATF-7 pathway in regulating tph-1 and antimicrobial genes supports this idea . Moreover , our genetic analysis suggests that activation mechanisms of this core signaling pathway in the neurons and intestine during infection differ . A prior work showed that DKF-2 is required for TIR-1 signaling cascade to induce intestinal immunity [37] . We found that dkf-2-null did not prevent PA14-induced tph-1::gfp upregulation . Instead , we identified a requirement of UNC-43 CaMKII for tph-1::gfp upregulation induced by PA14 and tir-1 ( yz68gf ) . Analysis of TIR-1 ( WT ) ::GFP and TIR-1 ( yz68 ) ::GFP suggests that UNC-43 regulates subcellular distribution of TIR-1 . While several mammalian TIR domain adaptor proteins have been shown to translocate to the plasma membrane following Toll-like receptor activation , SARM , the ortholog of TIR-1 , is activated in the brain by neural toxicity via a yet unidentified mechanism [55] , [56] . Our results raise the possibility that the cue associated with pathogenic bacterial infection triggers Ca2+ signaling to activate TIR-1 in the ADF neurons . It may be sensible that neurons and immune cells detect distinct molecular cues associated with infection thereby coordinating neuronal and physiological responses .
C . elegans strains were maintained at 20°C on NGM agar plates seeded with a lawn of Escherichia coli OP50 as a food source . WT animals were Bristol strain N2 . The Hawaiian isolate CB4856 was used in genetic mapping of the daf-19 and tir-1 mutations . The following existing mutant strains were used in this study: atf-7 ( qd22gf ) , atf-7 ( qd22 qd130lf ) , daf-6 ( e1377 ) , daf-19 ( m86 ) , dkf-2 ( pr3 ) , dyf-1 ( yz66 ) , eri-1 ( mg366 ) ;lin-15B ( n744 ) , ocr-2 ( yz5 ) , sek-1 ( km4 ) , tir-1 ( ok1052 ) , tir-1 ( tm3036lf ) , unc-43 ( e408 ) . Transgenic strains used in this study were: agIs219: Is[T24B8 . 5::gfp; ttx-3::gfp] [30] , CX3695: kyIs140[str-2::gfp; lin-15 ( + ) ] [57] , GR1333: yzIs71[tph-1::gfp; Rol-6 ( d ) ] [21] , Is[cat-1::gfp] [58] , JY222: ExZ042[ocr-2::gfp; Rol-6 ( d ) ] [4] , JY449: ExX002[str-2::gfp; Rol-6 ( d ) ] [59] . yz68 is a dominant mutation isolated from a genetic screen for mutants with enhanced GFP expression in ADF chemosensory neurons after ethyl methane sulfonate ( EMS ) mutagenesis of ocr-2 ( yz5 ) mutant carrying an integrated tph-1::gfp transgene as described previously [5] . Genetic mapping using single-nucleotide polymorphisms ( SNP ) of CB4856 localized yz68 to a contig of 1 . 43 map on the chromosome III . To identify the mutant gene , 174 genes located in the contig were individually inactivated in yz68 mutants by RNAi , and the clone F13B10 . 1 expressing double stranded ( ds ) -RNA of tir-1 suppressed the tph-1::gfp upregulation of the yz68 mutant . Sequencing yz68 genomic DNA revealed a G to A transition resulting in a cysteine426 to tyrosine substitution in the fourth ARM motif of the HEAT/Arm repeat region of TIR-1 . The amino acid altered in yz68 is in reference of the tir-1a isoform . yz69 and yz70 mutants are recessive mutations isolated from an EMS mutagenesis screen for mutants with dramatically reduced/absence of ADF tph-1::gfp . Analysis of the amphid morphology with fluorescence dye DiI revealed that ciliated neurons in yz69 and yz70 mutants were dye filling defective . CB4856 SNP mapping localized yz69 to a contig of 1 . 28 map units between the polymorphisms in the clones C18D1 and F44F4 on the chromosome II . Non-complementation assays with dye-filling mutants within the region indicated that both yz69 and yz70 were allelic to daf-19 . Sequencing the daf-19 gene of the mutants revealed in yz69 a G to A transition predicting a cysteine to tyrosine substitution at the conserved dimerization domain , and in yz70 a G to A transition predicting an opal mutation in the dimerization domain . The amino acid changes depicted in Figure 3A are in reference of daf-19a isoform . All constructs were generated by PCR . daf-19 ( g ) was a ∼14 . 8 kb genomic fragment amplified from the WT genome encompassing 2 . 9 kb 5′-upstream promoter sequence , exons/introns , and 574 bp 3′-UTR of the daf-19 gene . To express tir-1 and daf-19 in specific neurons , we fused tir-1 and daf-19 cDNA sequences individually to the gpa-13 promoter , which is expressed in three pairs of amphidal ciliated sensory neurons ADF , ASH and AWC . The gpa-13 promoter is expressed additionally in PHA and PHB phasmid neurons located in the tail [60] . tir-1 ( WT ) and tir-1 ( yz68 ) cDNA were amplified from cDNA mixture prepared from total RNA of WT and tir-1 ( yz68 ) animals , respectively , using primers corresponding to the tir-1a isoform . The cDNA of the daf-19c isoform was amplified from the plasmid PS0243 ( kindly provided by P . Swoboda ) . The 2 . 6 kb gpa-13 promoter sequence amplified from the plasmid PS0243 was fused to the sequences in the order of a cDNA , GFP and unc-54 3′UTR or a cDNA , mCherry and unc-54 3′UTR . To express DAF-19 in the intestine , the cDNA sequence of daf-19a isoform was fused to the 2 . 9 kb ges-1 promoter . For each construct , products from three independent PCR reactions were pooled to reduce potential PCR errors . The pooled PCR products were purified ( Qiagen ) and microinjected at the concentration of 50 ng/µl into worms . The plasmid containing either a dominant rol-6 gene ( Rol-6 ( d ) ) , elt-2::gfp or unc-122::RFP was co-injected as a transgenic marker . All RNA interference ( RNAi ) experiments were done in the background of eri-1;lin-15B , which enhances sensitivity to RNA interference in neurons [61] . RNAi assays were carried out by feeding worms E . coli HT1115 expressing dsRNA of a target gene or the control empty L4440 vector ( Ahringer RNAi library , University of Cambridge , England ) . RNAi clones were individually cultured overnight in Luria broth containing 100 µg/ml ampicillin , 500 µl of the bacterial culture were seeded onto agar plates containing NGM supplemented with 1 mM IPTG and 25 µg/ml carbenicillin to induce dsRNA expression , and incubated overnight at room temperature . For RNAi of tir-1 and nsy-1 , about 60 eggs were placed onto each plate and allowed to hatch , grow to adults and lay eggs . F1 progeny were transferred to a fresh plate , and tph-1::gfp in the ADF neurons of L4-stage worms of F2 generation quantified . For developmental RNAi of daf-19 , synchronized L1 worms were transferred to the plates ( day 0 ) , the worms were transferred to freshly prepared plates every day , and the expression of tph-1::gfp in ADF neurons or DiI staining of cilia morphology were analyzed on indicated days . DiI staining was done as previously described [5] . For each RNAi experiment , three independent trials each with three replicates were performed , and data from one representative trial presented . Whole-mount staining of worms with anti-5-HT antibody was performed as described previously [21] . The staining patterns were visualized via Alexa Fluor 594 or 488 conjugated secondary antibodies ( Molecular Probe ) under an AxioImager Z1 microscope equipped with proper filters , and images were captured using AxioCam MR digital camera ( Zeiss , Northwood , NY ) . To quantify the intensity of 5-HT immunoreactivity , images of ADF neurons in individual worms were captured under a 40× objective lens at a fixed exposure time of 3 ms with 100% UV exposure level . For each image , fluorescence intensity of a circular 10 pixels area within the ADF cell body was quantified using the ImageJ software ( National Institute of Health , Bethesda , Maryland ) . To exclude the background , fluorescence intensity over a circular 10 pixels area posterior to the ADF cell body in the same image was quantified , and the value of the background was subtracted from the value of the ADF area . The expression of a chromosomally integrated tph-1::gfp reporter in ADF neurons in living WT or mutant worms was evaluated by measuring GFP fluorescence intensity . Images of ADF neurons in individual animals were captured at a fixed exposure time . The external contour of each ADF neuron was delineated , and fluorescence intensity within the entire neuron was quantified by using the ImageJ software . L4-stage animals were examined , unless noted otherwise . Ideally mutants that reduce and that upregulate tph-1::gfp were assayed in parallel , thus the exposure time was designed to detect reduction as well as upregulation of tph-1::gfp . This setting was most reliable at the range of 1–3 fold higher than WT; consequently , we used ocr-2 mutant background to lower the basal tph-1::gfp to evaluate the changes of mutants with constitutively increased ADF tph-1::gfp , and the double mutants were compared with ocr-2 single mutants assayed on the same day . For quantifying tph-1::gfp intensity following PA14 treatment , one-day-old young adult worms were transferred to standard slow killing assay plates [62] seeded with either PA14 or control OP50 , incubated at 25°C for 6 hr , images of the ADF neurons were captured and GFP intensity quantified . For quantifying tph-1::gfp intensity in dauers , WT and mutants were induced to form dauers by dauer pheromones . Dauer pheromone and pheromone-containing plates were prepared as we have done previously [5] , following an established protocol [63] . 20–25 gravid worms from each strain were transferred to NGM plates supplied with 1 unit of dauer pheromone , allowed to lay eggs for 2–3 hr in a 25°C incubator , the adults were then removed from the plates , and dauers developed from hatched eggs on the plates were analyzed 72 hr later . For each strain the value of dauers was compared to that of L4 grown on the plates without pheromone and assayed on the same day . The expression of T24B8 . 5::gfp in the intestine was analyzed in two-day-old adult worms carrying the integrated agIs219[T24B8 . 5::gfp; ttx-3::gfp] transgene cultured on OP50 at 20°C . Images of the intestine were captured under a 10× objective lens at a fixed exposure time , and fluorescence intensity was quantified by measuring pixel intensity of three areas along the body of each animal as depicted in Figure 6B . Data represent the average of at least three trials unless specified otherwise . For each trial , 15–20 animals per strain per condition/treatment were analyzed and compared to the controls assayed on the same day . WT animals under the same conditions and treatments were analyzed for every experiment . Unpaired Student's t-test was used for comparisons between a mutant and WT and between two mutants or two treatments . The standard PA14 slow killing assays were performed as previously described [62] . Briefly , PA14 was cultured in King's broth overnight and the culture was seeded at the center of 3 . 5 cm diameter assay plates and incubated at 37°C for 20 hr followed by 20 hr incubation at room temperature . 40–50 L4 worms per strain were transferred onto each assay plate , incubated at 25°C and scored for dead or live every 8 hr . Worms were scored as dead if no response was detected after prodding with a platinum wire . daf-19 mutants tend to claw off the plate . So for each assay , more than 300 L4 daf-19 mutants were transferred to each plate , and live and dead animals on the agar surface were scored at indicated time points; dead animals on the wall of the plate not counted . For each strain , three replicates were analyzed for each experiment . To test the effect of RNAi of daf-19 , 4 to 6 gravid animals were grown on the RNAi plates as described above seed with either E-coli HT1115 harboring empty control plasmid L4440 or plasmids expressing RNAi against daf-19 gene . F1 progeny at the L4 stage were used to test the susceptibility to killing by PA14 . Total RNA from 100 one-day-old adults of WT and indicated mutant strains was extracted using Trizol ( Invitrogen ) , reserve transcribed to cDNA using the SuperScript III system ( Invitrogen ) , and the cDNA was used for qPCR analyses using the StepOnePlus machine ( Applied Biosystems ) and SYBR Green detection system ( Applied Biosystems ) in triplicated reactions . The primers for qPCR were designed using Primer Premier 5 . 0 ( Premier Biosoft ) ( Figure S5 ) . Values were normalized against the reference gene act-1 [37] . gpd-2 was analyzed as a second control showing no change relative to act-1 . Fold change was calculated using the delta Ct method [64] . qPCR analysis of L4 worms of those strains showed comparable results; data of the adults are presented . | Toll-interleukin receptor ( TIR ) –domain adaptor proteins are keys to activate signaling cascades inducing transcriptional responses to internal and external pathogenic signals in evolutionary disparate organisms . Despite lacking a homolog of the mammalian innate immunity transcriptional regulator nuclear factor-kappaB ( NF-κB ) , the nematode Caenorhabditis elegans responds to infections by activating TIR-1 signaling targets in the innate immune system and in neurons . Through a genetic screen for factors required for TIR-1 signaling to upregulate the serotonin biosynthesis gene tph-1 , we identified DAF-19 , an ortholog of regulatory factor X ( RFX ) transcription factors that were initially discovered in human immune cells . We show that DAF-19 concerts with ATF-7 , a member of the activating transcription factor ( ATF ) /cAMP response element-binding B ( CREB ) family of transcription factors , to upregulate tph-1 in the ADF chemosensory neurons and antimicrobial genes in the intestine in response to bacterial infection , reminiscent of RFX-CREB interaction in human immune cells . daf-19 mutants display heightened susceptibility to killing by the human pathogen Pseudomonas aeruginosa PA14 . Our studies suggest that RFX transcriptional regulation , which is essential for human adaptive immunity , has an ancient role in controlling serotonin biosynthesis and innate immunity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"molecular",
"neuroscience",
"genetic",
"mutation",
"gene",
"regulation",
"immunology",
"neuroscience",
"gene",
"function",
"molecular",
"genetics",
"neurotransmitters",
"gene",
"expression",
"biology",
"genetics",
"of",
"the",
"immune",
"system",
"genetic",
"screens",
"genetics",
"genetics",
"and",
"genomics"
]
| 2013 | RFX Transcription Factor DAF-19 Regulates 5-HT and Innate Immune Responses to Pathogenic Bacteria in Caenorhabditis elegans |
Hantaan virus ( HTNV ) infection in humans is a serious public health concern in Asia . A potent T cell activation peptide vaccine from HTNV structure protein represents a promising immunotherapy for disease control . However , the T cell epitopes of the HTNV restricted by the HLA alleles and the role of epitope-specific T cell response after HTNV infection remain largely unexplored . Five well-conserved novel CD8+ T-cell epitopes of the HTNV nucleoprotein restricted by the most popular HLA alleles in Chinese Han population were defined with interferon-γ enzyme-linked immunospot assay in 37 patients infected with HTNV during hospitalization . Two epitopes aa129–aa137 and aa131–aa139 restricted by HLA-A2 and B35 , respectively , were selected to evaluate the epitope-specific CD8+ T-cell response . HLA-peptide pentamer complex staining showed that the frequency of single epitope-specific CD8+ T cell could be detected in patients ( 95% confidence interval for aa129–aa137: 0 . 080%–0 . 208%; for aa131–aa139: 0 . 030%–0 . 094% ) . The frequency of epitope-specific pentamer+ CD8+ T-cell response was much higher in mild/moderate patients than in severe/critical ones at the acute stage of the disease . Moreover , the frequency of epitope-specific CD8+ T cells at acute stage was inversely associated with the peak level of serum creatinine and was positively associated with the nadir platelet counts during the hospitalization . The intracellular cytokine staining and the proliferation assay showed that the effective epitope-specific CD8+ T cells were characterized with the production of interferon-γ , expression of CD69 and the strong capacity of proliferation . The novel HLA class I restricted HTNV nucleoprotein epitopes-specific CD8+ T-cell responses would be closely related with the progression and the severity of the disease , which could provide the first step toward effective peptide vaccine development against HTNV infection in humans .
Hantaan virus ( HTNV ) , the prototype member of the genus Hantavirus of the family Bunyaviridae , was first isolated in Korea in 1978 and can cause a severe disease of hemorrhagic fever with renal syndrome ( HFRS ) in humans [1] . The genus Hantavirus is naturally maintained in persistently infected rodents and can be transmitted to humans via the inhalation of aerosols [2] , [3] . Recently , pathogenic Hantaviruses have emerged as an increasing threat to human health , as not only are they the causative agents of HFRS in Asia and Europe that are associated with HTNV , Seoul virus , Dobrava virus , and Puumala virus ( PUUV ) infections [4]–[6] , but also they can induce the severe hantavirus pulmonary syndrome ( HPS ) in North and South America , which is caused by Sin Nombre virus ( SNV ) and Andes virus serotypes [7] , [8] . Specifically , the outbreak news recently reported 6 cases of HPS among visitors to Yosemite National Park in California and two of the 6 cases were fatal . The HFRS is a fulminant infectious disease which is characterized by fever , hemorrhage , renal impairment , and thrombocytopenia [9]–[11] . More than 100 , 000 cases of HFRS are reported annually , with a case fatality rate of approximately 15% . China is a severe epidemic area , where the HFRS human cases account for 90% of the total global cases [12] . Several studies have suggested that the Hantavirus infection could induce a vigorous cellular immune response in humans [13] , [14] , including an expansion of the number of activated circulating CD8+ T cells [15]–[17] , an increase in the CD8+/CD4+ T-cells ratio [16] , the infiltration of CD8+ T lymphocytes in the kidney biopsies of PUUV-infected patients [18] , and the high frequencies of virus-specific memory CD8+ T lymphocytes that exist for a long time after the HTNV or PUUV infection [19]–[21] . It was reported that the T cell responses targeting the epitopes on viruses restricted by particular HLA alleles could contribute to the different outcome of the clinical course in many virus infectious diseases [22]–[24] . Our previous research found that the general T cell responses specific for HTNV nucleoprotein ( NP ) , which is highly immunogenic and conservative , might help reduce the risk of progression to acute renal failure [25] . However , whether the HTNV epitopes specific T-cell responses restricted by different HLA alleles would correlate with the clinical severity in HFRS is still unclear . Three cytotoxic T cell ( CTL ) epitopes of the HTNV ( 76–118 strain ) -NP , aa421–aa429 , aa334–aa342 and aa12–aa20 restricted by HLA-A1 , A2 . 1 , and B51 , respectively were identified a decade ago [19] , [26] . Recently , we have reported three novel CTL epitopes on HTNV-NP , including aa197–aa205 , aa245–aa253 , and aa258–aa266 , which were restricted by various HLA alleles , including A11 , A24 , and B7 [27] . In addition , HTNV-NP C-terminal polypeptides aa301–aa315 , aa355–aa369 , and aa415–aa429 could induce strong CD8+ T-cell responses [28] . Based on these , the overall aim of our experiments is to identify the novel HTNV epitopes restricted by the major HLA alleles in Chinese Han population and to investigate the function of HTNV epitope-specific T-cell responses , which will be helpful to establish a better program for the diagnosis and treatment of HFRS , and to develop an effective peptide vaccine against the HTNV .
Written informed consent was obtained from each HFRS patient or their guardians under a protocol approved by the Institutional Review Board of the Tangdu Hospital and the Fourth Military Medical University . The research involving human materials was also approved by the Ethical Review Board of the University , and the related information was used anonymously . A total of 37 adults presenting to the doctors with symptoms of fever , hemorrhage , effusion , and renal abnormalities and who were prospectively identified as HFRS were enrolled in this study from the Departments of Infectious Diseases at Tangdu Hospital of the Fourth Military Medical University ( Xi'an , China ) . The clinical diagnosis of HFRS was confirmed by the detection of specific immunoglobulin M ( IgM ) or IgG antibodies to HTNV in the patients' serum specimens . The severity degree of the HFRS disease could be classified as previously described [25]: ( 1 ) Mild patients were identified with mild renal failure without an obvious oliguric stage; ( 2 ) moderate for those with obvious symptoms of uremia , effusion ( bulbar conjunctiva ) , hemorrhage ( skin and mucous membrane ) , and renal failure with a typical oliguric stage; ( 3 ) severe patients with severe uremia , effusion ( bulbar conjunctiva and either peritoneum or pleura ) , hemorrhage ( skin and mucous membrane ) , and renal failure with oliguria ( urine output , 50–500 ml/day ) for ≤5 days or anuria ( urine output , <50 ml/day ) for ≤2 days; and ( 4 ) critical ones with ≥1 of the following symptoms during severe disease: refractory shock , visceral hemorrhage , heart failure , pulmonary edema , brain edema , severe secondary infection , and severe renal failure with either oliguria ( urine output , 50–500 ml/day ) for >5 days , anuria ( urine output , <50 ml/day ) for >2 days , or a blood urea nitrogen level of >42 . 84 mmol/L . The patients who had other kidney disease , diabetes , cardiovascular disease , hematological disease , autoimmune disease , viral hepatitis , and other liver diseases were excluded in this study . The number of patients with severity degree of mild , moderate , severe , and critical was 5 , 12 , 9 , and 11 , respectively . According to the clinical observations , the illness could be divided into five sequential stages: febrile , hypotensive , oliguric , diuretic , and convalescent . The phase within 8 days from the fever onset to early oliguric stage was usually defined as acute or early stage of the disease . The detailed characteristics of the patients enrolled in the present study were summarized in Table S1 . Four healthy adult volunteers , two women and two men , were selected as normal control donors for the present study .
We identified five novel CD8+ T-cell epitopes on HTNV-NP from four HFRS patients through a similar investigation protocol as before [27] . The single-positive 15-mer peptides which could primarily stimulate the CD8+ T-cell response were screened out in Figure 1A , indicating that these peptides contained the CTL epitopes . The definition of the epitopes recognized by the CD8+ T cells showed that the aa129–aa137 ( FVVPILLKA ) , aa131–aa139 ( VPILLKALY ) , aa247–aa255 ( LPDTAAVSL ) , aa167–aa175 ( DVNGIRKPK ) , and aa277–aa285 ( ETKESKAIR ) were the HTNV-NP CTL epitopes that can simulate strong IFN-γ responses ( Figure 1B ) . The in vitro peptide-specific pre-sensitized CD8+ T cells were successfully generated from all four donors , as defined by a flow cytometry analysis ( data not show ) . The HLA-matched and mismatched EBV-B cells used to confirm HLA restrictions and the HLA class I molecules of the patients were shown in Figure 2 . The nonamers recognized by these CD8+ T cells in the four donors were restricted by three different HLA class I alleles ( Table 1 ) . These novel nonamer epitopes were further supported by the binding motifs , mainly anchor residues at position 2 or position 9 . The epitope aa129–aa137 fits the HLA-A2 binding motif at the anchor residues position 2 ( leucine , valine or glutarnine ) . Epitopes aa131–aa139 and aa247–aa255 fit the HLA-B35 binding motif , including anchor residues at position 2 ( proline , alanine , or valine ) and the C terminus ( leucine , tyrosine , or methionine ) . Epitopes aa167–aa175 and aa277–aa285 fit the HLA-A33 binding motif , including anchor residues at position 2 ( alanine , isoleucine or valine ) and the C terminus ( arginine ) [29] . When comparing HTNV to other Hantaviruses , the sequences of the epitopes aa129–aa137 , aa131–aa139 , and aa167–aa175 were conserved well with 67%–100% concordance among Hantaviruses , whereas the epitopes aa247–aa255 and aa277–aa285 were less well conserved ( Table 2 ) . Since HLA-A2 is the most frequent allele ( 29 . 7% ) of HLA-A loci in the Chinese Han population and HLA-B35 is a major allele in HLA-B loci [30] , we focused the epitopes aa129–aa137 and aa131–aa139 restricted by HLA-A2 and HLA-B35 , respectively , and generated the HLA class I peptide pentameric complex for these two HTNV-NP epitopes . Twenty-five HLA-A2+ patients and eight HLA-B35+ patients with a different severity of HFRS were tested at early and late time points during hospitalization . Epitope aa129–aa137-specific pentamer+ CD8+ T cells in PBMCs could be detected in 11 of the 12 ( 91 . 7% ) patients with mild/moderate severity as compared with 7 of the 13 ( 53 . 8% ) performed on severe/critical patients ( Table S1 ) . ( Fisher's exact chi-square test , P = 0 . 073 ) . The frequency of CD8+ T cells that were specific for the HLA-A2-restricted epitope aa129–aa137 ranged from 0 . 010 to 0 . 700% ( CI95%: 0 . 080%–0 . 208% ) . For the HLA-B35+ patients , the frequency of the circulating CD8+ T cells that were specific for the epitope aa131–aa139 ranged from 0 . 010 to 0 . 185% ( CI95%: 0 . 030%–0 . 094% ) ( Figure 3 ) . We compared the epitope-specific CD8+ T-cell frequencies between patients with mild/moderate HFRS and those with severe/critical HFRS at the earliest available acute stage time point in the 18 patients with HLA-A2 and in the 8 patients with HLA-B35 from whom PBMCs had been collected . The statistical analyses showed that the frequencies of the epitope-specific CD8+ T cells were much higher in the mild/moderate HFRS patients than that in the severe/critical patients ( Mann–Whitney U test , for HLA-A2 restricted epitope: P = 0 . 007 , for HLA-B35 restricted epitope: P = 0 . 021 ) ( Figure 4A ) . There was no difference in the frequency between the two groups at the late stage of the disease ( data not shown ) . Then we compared the frequencies of the epitope-specific CD8+ T-cell response between the acute stage and the late stage ( diuretic and convalescence stage ) during the illness in ten HLA-A2+ patients and in seven HLA-B35+ patients , and the results showed that the frequency of the epitope-specific CD8+ T cells in the acute stage was higher than that in the late stage in patients ( Mann–Whitney U test , for HLA-A2 restricted epitope: P = 0 . 041 , for HLA-B35 restricted epitope: P = 0 . 009 ) ( Figure 4B ) . The analysis of the associations between the HTNV-NP epitope-specific CD8+ T-cell responses and the clinical parameters showed that the frequency of epitope-specific CD8+ T cells at the acute stage was inversely associated with the peak level of serum creatinine ( Spearman correlation test , for HLA-A2 restricted epitope: P = 0 . 030 , r = −0 . 511; for HLA-B35 restricted epitope: P = 0 . 021 , r = −0 . 786 ) and positively associated with the nadir of platelet counts ( for HLA-A2 restricted epitope: P = 0 . 018 , r = 0 . 549; for HLA-B35 restricted epitope: P = 0 . 010 , r = 0 . 833 ) during the clinical course of the HFRS ( Figure 4C ) . Interestingly , it seemed that there were two subsets of CD8+ T cells in PBMCs with high or low mean fluorescence intensity , both of which could respond to the specific epitopes presented by the pentamers ( Figure 3 ) . The further staining showed that both CD8hi and CD8lo T cells were CD8αβ heterodimer expression ( Figure S1 ) . The much higher ratio of CD8lo/CD8hi T cell-numbers in the acute stage would decline at late stage of HFRS . Both the CD8lo and CD8hi T cell subsets could produce IFN-γ when stimulated with specific epitopes ( Figure 5A ) . However , a stronger proliferative capacity of epitope-specific CD8hi T cell subset than that in the CD8lo T cell subset at the acute stage was observed , as shown by the pentamer+ cells in the upper left quadrants of the dot plots ( Figure 6A lower panel ) . The quantity of epitope-specific CD8+ CD69+ T cells producing IFN-γ at acute and late time points were performed in Figure 5A , where the examples of four HLA-A2+ patients with indicated severity showed that a decline tendency could also be found in the kinetics of IFN-γ quantity ( upper panel ) , and the peptide-induced up-regulation of CD69 was detectable at both time points in the PBMCs of the patients ( lower panel ) . The frequency of IFN-γ produced by the CD8+ CD69+ T cells was much higher in patients with mild/moderate severity than in severe/critical ones at the acute stage of HFRS ( Mann–Whitney U test , for aa129–aa137: P = 0 . 037; for aa131–aa139: P = 0 . 043 ) ( Figure 5B–5C ) . Despite the low number of total activated epitope-specific CD8+ T cells , a high proliferative potential of the pentamer+ CD8+ T cells was observed ( Figure 6A ) . For epitope aa129–aa137 , the proliferative capacity of the epitope-specific CD8+ T cells was stronger in mild/moderate patients than that in severe/critical patients ( Mann–Whitney U test , P = 0 . 013 ) ( Figure 6B ) .
In this study , we defined five novel HTNV-NP CTL epitopes restricted by the major HLA class I molecules in Chinese Han population; we provided , for the first time , a quantification of the HTNV epitope-specific IFN-γ–secreting CD8+ T cells in HFRS patients; we analyzed the kinetics changing , the activation , and proliferation capacity of the epitope-specific CD8+ T cells and evaluated the associations between the epitope-specific CD8+ T-cell frequencies and the different outcomes of the HFRS severity . Generally , the residues on these novel epitopes we defined are well conserved , especially among the Old World Hantaviruses that cause HFRS . Specifically , the HTNV and Dobrava virus share the identical sequences of three epitopes we defined , although the endemic areas of these two viruses are far from each other . We speculate that the sequences of well-conserved epitopes may locus mainly at the conserved region among the genus Hantaviruses . Particularly , a well-conserved epitope aa131–aa139 with little variation in the HLA anchor residues restricted by HLA-B35 was defined , in accordance with both the sequence and HLA restriction of the epitope identified on the NP of the SNV and Andes virus previously [31] , [32] , which make our results more credible . The quantitation of peptide-specific CD8+ T cells demonstrated that the epitopes we defined could bind to the corresponding HLA class I molecules to form the HLA-peptide complex , which could stimulate epitope-specific CD8+ T-lymphocyte responses . Compared with our previous general study [25] , we focused on the HLA restricted single epitope-specific CD8+ T-cell responses and found that there was an obvious inverse association between the magnitude of the epitope-specific CD8+ T-cell responses in the acute stage and the severity degree of the HFRS , which was in accordance with our recently findings that at the acute stage of disease , patients in severe/critical group were found to have higher viral loads than those in mild/moderate group [33] , suggesting that functional HTNV epitope-specific CD8+ T cells at the acute stage of HFRS would be important for the clearance of the virus . Furthermore , we found there was a tendency of decline of the epitope-specific CD8+ T-cell frequencies from acute stage to convalescence , which was inconsistent with our previous results that the gradually increased frequencies of HTNV NP-specific T-cell responses during the course of the disease was found with IFN-γ ELISPOT assay [25] . There are several critical methodological differences between these two studies . However , the Finnish group found when studying HFRS caused by PUUV that IFN-γ ELISPOT assays using PBMCs obtained in the acute stage of the disease hugely underestimated the frequency of peptide-specific T cells compared to HLA/peptide tetramer staining probably due to activation induced apoptosis of these peptide-specific T cells during incubation with the peptides for the assay [20] . Thus the conclusions would be taken by caution in our previous study [34] . The decreased frequencies of HTNV-NP epitope-specific CD8+ T cells were consistent with the findings that HTNV RNA load could only be detected in plasma of HFRS patients in febrile/hypotensive and oliguric stage and gradually declined to an undetectable level with the progress of the disease [33] . Additional analysis found that the levels of single epitope-specific CD8+ T-cell response at the early stage of the disease associated with clinical parameters . The peak level of serum creatinine and the nadir of platelet counts during the course of HFRS usually could provide a better prediction of disease severity . The higher frequency of the epitope-specific CD8+ T cells at the acute stage of the patients , the lower peak level of serum creatinine and the higher nadir of platelet counts during hospitalization , which indicated that the level of the HTNV-NP epitope-specific CD8+ T-cell responses at the acute stage could predict the different outcomes of the disease . However , the association between the NP-specific CD8+ T-cell response and the disease severity still needs to be investigated , since the major immunodominant structure proteins of HTNV that could induce the CD8+ T-cell response is uncertain . Furthermore , the studies conducted on the function of the virus-specific CD8+ T-cell responses in different viruses , usually with specific HLA restrictions , are currently controversial [22] , [35] , [36] . Similar with our finding , a study conducted on the Andes virus-infected patients found that among the HLA-B35-positive patients , mild disease outcome seemed to be associated with stronger responses toward the Gn-carboxyterminus than that in patients with severe HPS [32] . Whereas the study on SNV showed that the intense responses of the epitope-specific CD8+ T cells restricted by HLA-B3501 contribute to the severe outcome of HPS [23] . Therefore , the different HLA allele restrictions and different viruses would be important factors influencing the T cell responses and the outcome of the diseases that should be taken into account in such anti-virus cellular immune response studies . Another finding from the current data is that the frequency of the HTNV epitope-specific pentamer+ CD8+ T cells was low , regardless of the severity of the HFRS patients . It is not surprising for the result because only the NP epitope-specific CTL response was detected here . The low frequencies may be due to the relatively low contribution of NP-directed T-cell responses , in comparison to the surface glycoprotein , which is the other important structure antigen of HTNV and can induce much stronger CTL responses than NP , as indicated in the studies of other Hantaviruses [23] , [32] , especially the report on SNV infections , in which the frequency of epitope-specific CD8+ T cells has been reported to be as high as 44 . 2% [23] . In fact , during the millennia of severe Hantavirus infections , persons with susceptible HLA alleles may have been eliminated from certain human populations , thus skewing the distribution of alleles in the population . Since the HLA-A2 and -B35 in our present study are the major alleles in Chinese Han population , but not the alleles with susceptibility in HFRS patients [37] , [38] , whereas the HLA-B3501 allele is a risk factor for severe HPS and relatively common allele among HPS patients induced by the New World Hantaviruses [23] , [39] , we speculate that the different genetic susceptibility should also be taken into account to explain the differences of the frequency of T-cell response among Hantaviruses . Although our data demonstrated that HTNV-specific CD8+ T-cell responses directed against the individual epitope could be detected with the pentamer staining , a high proportion of subpopulation CD8lo T cells presented in the early stage of HTNV infection . It is interesting to evaluate whether both these subsets of CD8+ T cells have similar behaviors that contribute to the anti-viral effects . According to the studies on other virus infectious diseases , polyfunctional CD8+ T cells may play different roles in the pathogenesis of the disease [40] , [41] . Compared with epitope-specific CD8hi T cell subset , the CD8lo T cells with the similar ability to secrete IFN-γ and the milder proliferation capability at the acute stage of the disease might be represented with the same peptide specificity , but with a poly-cytokines secretion profile or a variety of phenotypes , which might contribute to the clearance of the virus and the milder severity of the disease . Due to limitations in the PBMC samples , we were unable to further test for differential cytokine secretion profiles for both the subsets and the cause of this CD8lo T cell subset , but we propose that both of them would have a similar effect function with more or less distinction that need to be confirmed next . An overall analysis revealed that the epitope-specific CD8+ T cells up-regulated CD69 after exposure to peptides , representing an activated phenotype and showed obvious capacity to secret IFN-γ and proliferation in acute infection . Similar with many other virus infectious diseases [42] , [43] , epitope-specific CD8+ T-cell proliferation generally correlated with effective immunity , as the capacity of proliferation was slight in severe/critical patients , which represents a further step in the understanding of the impairment of CD8+ T cell functions in severe/critical HFRS patients . In summary , such a thorough study on the identification of HLA Class I-T cell epitopes of HTNV and evaluation of the function of single epitope-specific CD8+ T-cell response on the clinical outcomes have not been reported before in the field of HTNV investigation . Our results might provide new insights into understanding the relationship between single epitope-specific CD8+ T-cell response , CD8+ T-cell functional characters , and disease control in acute zoonotic HTNV infections in humans , which could be a rationale to explore immunotherapy as an adjunctive therapy in people with HFRS and will help to speed up the novel vaccine design process against the HTNV infection . Since the first step toward a peptide vaccine is epitope mapping of the HTNV structure proteins according to the most frequent HLA alleles in Chinese Han population , the only research on the T-cell epitopes and their responses to NP is just a fraction of the research work . The epitopes and the immunogenicity of the HTNV glycoproteins , which could elicit even higher detectable T-cell responses in HFRS patients , are still needed investigation to make a comprehensive explanation of the cellular immunity functions after HTNV infection . The combination of the studies on both HTNV-NP and glycoprotein will be more valuable and helpful to understand the immune interaction between the HTNV and the host . Besides , whether our results could be generalized to T-cell responses to peptides presented by other HLA alleles and whether broadening or magnifying the CD8+ T-cell response in patients with severe infection is possible still need to be investigated . | Hantaan virus ( HTNV ) , the prototype of the Hantavirus genus , is a rodent-borne pathogen that causes human hemorrhagic fever with renal syndrome ( HFRS ) with a mortality rate of approximately 15% in Asia . Since effective prevention is not available currently and the non-specific symptoms at the early stage of the disease always lead to the delay of visiting to hospital or misdiagnosis , alternative vaccinations against HTNV are of priority to overcome the problem . We defined five novel HTNV nucleoprotein CD8+ T-cell epitopes restricted by the most popular HLA alleles in Chinese Han population . For the first time , we quantitated the HTNV epitope-specific CD8+ T-cell frequency during HTNV infection and evaluated the correlations between the CD8+ T-cell response and the different outcomes of the HFRS severity . We also found that effective HTNV nucleoprotein epitope-specific CD8+ T-cell responses were characterized by the interferon-γ secretion with a strong capacity of activation and proliferation . Our results add weight to understanding the important role of epitope-specific CD8+ T-cell responses in the disease control after acute zoonotic HTNV infections in humans and provide a rationale foundation to speed up the process of peptide vaccine development . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"viral",
"hemorrhagic",
"fevers",
"infectious",
"diseases",
"hemorrhagic",
"fever",
"with",
"renal",
"syndrome",
"adaptive",
"immunity",
"immunity",
"virology",
"neglected",
"tropical",
"diseases",
"biology",
"microbiology",
"host-pathogen",
"interaction",
"viral",
"diseases"
]
| 2013 | HLA-A2 and B35 Restricted Hantaan Virus Nucleoprotein CD8+ T-Cell Epitope-Specific Immune Response Correlates with Milder Disease in Hemorrhagic Fever with Renal Syndrome |
Scientists have long sought to understand how vascular networks supply blood and oxygen to cells throughout the body . Recent work focuses on principles that constrain how vessel size changes through branching generations from the aorta to capillaries and uses scaling exponents to quantify these changes . Prominent scaling theories predict that combinations of these exponents explain how metabolic , growth , and other biological rates vary with body size . Nevertheless , direct measurements of individual vessel segments have been limited because existing techniques for measuring vasculature are invasive , time consuming , and technically difficult . We developed software that extracts the length , radius , and connectivity of in vivo vessels from contrast-enhanced 3D Magnetic Resonance Angiography . Using data from 20 human subjects , we calculated scaling exponents by four methods—two derived from local properties of branching junctions and two from whole-network properties . Although these methods are often used interchangeably in the literature , we do not find general agreement between these methods , particularly for vessel lengths . Measurements for length of vessels also diverge from theoretical values , but those for radius show stronger agreement . Our results demonstrate that vascular network models cannot ignore certain complexities of real vascular systems and indicate the need to discover new principles regarding vessel lengths .
Networks that supply resources are essential to the maintenance and growth of many natural and engineered systems . These resource-distribution networks are pervasive throughout biology . Examples include the tracheal system in insects , xylem networks in plants [1] , foraging trails of ant colonies [2 , 3] , and cardiovascular systems in animals [4–6] . Because the cardiovascular system delivers resources and energy to the body , its structure at least partly determines rates of growth and metabolism [4 , 7 , 8] . Moreover , the cardiovascular system plays a role in many diseases—such as heart disease , stroke , inflammation , and malignant tumor growth [9–11] . The pervasiveness and importance of material transport in biology motivates the search for basic principles that help shape resource-distribution networks in general and the cardiovascular system in particular . The theory of the structure of vascular networks has roots in the early 20th century [5 , 7 , 12] . More recent theories predict properties of the entire vascular network by assuming a hierarchical structure and include those of West , Brown and Enquist [4] ( henceforth the WBE model ) , Banavar et al . [13 , 14] , Dodds [15] , Huo and Kassab [16] , and others [1 , 17–21] . These theories assume or predict how vascular structure , such as the radius of vessels , changes as the network branches from aorta to capillaries . While early theories focused on vessel radius , recent theories also incorporate how vessel length changes across the network [4 , 14–16] . Many also assume symmetric branching—child vessels have identical properties [4 , 16 , 19 , 20] . In symmetric , hierarchical models , knowledge of both radius and length enables derivations of how metabolic rate varies with body size across species [4] and how organismal and tumor growth rate change with body size [9–11] . Several theories predict vascular structures will be found to be self-similar—some aspect of the network can be viewed as a rescaled copy of the whole [22]—across specific ranges of spatial scale [4 , 23] . Self similarity necessarily leads to relationships and distributions that are characterized by power laws , whose exponents we call the scaling exponents [24] . Self similarity can either be a strict or statistical property . Each new chamber of the nautilus is a larger exact copy of the previous chamber , while along coastlines shorter segments exhibit rescaled statistical properties of the longer segments [25] . Self similarity in nature suggests scale-free principles that constrain structure . An organism may need to maintain a certain shape at all stages of growth , use a common developmental program at all scales , or cope with physical processes with no preferred scale—such as energy minimization or turbulence in fluids . Self similarity greatly simplifies many calculations , with applications from cardiac physiology [26] to allometric scaling [4] . Real vasculature is known to deviate from hierarchical , symmetric models in many ways , leading to criticism and debate about leading models [27 , 28] . Without reliable data , it is impossible to determine whether these deviations can be ignored , so that existing theories can accurately predict newly observed features such as curvature in scaling relationships [19] . Price et al . [29] recently decried the lack of data for individual vessel segments that are needed for tests of scaling theory . Direct measurements and tests may reject existing principles while laying the groundwork for the discovery of new patterns and principles in vascular architecture . In addition , measurements can parameterize existing equations to obtain more exact predictions for metabolic rates and growth curves , and can be used to examine the natural variation in the parameters across species , tissues , and tumor types . Vessel segment data preserves the asymmetry , reticulation , tortuosity , and other features of real vasculature , and quantitative data about these features may inform future models . Furthermore , extreme values or distinct patterns of variation may be signatures of pathologies that could eventually be used as diagnostics . Recent work involving direct measurements of plant architecture has begun to realize this potential [21 , 30 , 31] . Comprehensively characterizing vascular structure and obtaining reliable estimates for vascular scaling exponents requires large numbers of measurements across orders-of-magnitude in spatial scale—ranging from 0 . 004 mm to 15 mm for vessel radii in humans . Measurement is complicated because vascular systems are intertwined with tissues throughout the body at a wide range of spatial scales [32] . Even gross morphological measurements have historically taken impressive and time-consuming efforts and required invasive methods such as casting [33–36] . Zamir [36] and Kassab [35] constructed explicit descriptions of small regions of vascular systems—such as the coronary artery—by perfusing fixed specimens with a silicone or acrylic polymer , dissolving tissue away and examining each vessel segment under a microscope . These approaches enabled the measurement of tens of thousands of vessels from fixed specimens that were then used to test and develop vascular system models [35 , 36] . However , the process of casting may enlarge or damage vessels , and little of this raw data is publicly available for analysis . In addition , many more measurements across the range of scales are needed to identify the principles that shape vascular architecture . Different physical principles may dominate at different scales , and mapping out different regimes will require large amounts of data at each scale . In the WBE model [4] , the dominant mechanism of energy loss for blood flow in the arterioles ( radius ≪ 1 mm ) is viscous dissipation , but near the heart ( vessel radius ≫ 1 mm ) pulsatile flow and reflection of pressure waves along vessel walls dominates [26] . The transitions between these regimes are neither well-understood theoretically nor described empirically [8] . The pulsatile regime—the focus of our measurements—has greater variation in the number of branching orders and size of vessels across species and is thus the primary determinant of scaling of metabolic and other vital rates with body size [8] . Large amounts of vascular data across all relevant spatial scales are contained within existing angiographic images ( e . g . , MRI and X-Ray ) . These images are obtained non-invasively , thus avoiding problems of damage to vasculature and allowing for the possibility of longitudinal studies . The latent data within these images represents a tremendous opportunity . All that is needed is a reliable and automatic method for extracting vascular data from angiographic images . We have developed novel software that imports 3D images , creates a topologically and spatially explicit map of the blood vessel network , and measures the radius , length , and volume of all visible vessels . We have applied this software to 20 magnetic resonance angiograms of living humans to obtain 3015 data points that range in radius from 0 . 6–6 . 8 mm , representing an order of magnitude . This range corresponds to large vessels in the pulsatile flow regime relevant to allometric scaling [8] . We then analyzed these data based on the four distinct methods ( Eqs 1a–4 ) for measuring vascular scaling exponents described below .
Vascular scaling exponents encapsulate how radius and length of vessels change across the network . Virtually all scaling relationships for local or global properties can be expressed in terms of these vascular scaling exponents . Consequently , we view these scaling exponents as forming the foundation of most modern biological scaling theory and make them the primary focus of our analysis . We here describe four distinct methods for calculating vascular scaling exponents . Because the radius of the vessel plays a primary role in determining both the flow rate and resistance to blood flow through the vessel , theories for the vascular scaling exponent for vessel radius often focus on the power to pump blood from the heart to the capillaries . It is argued [4] that this power will have been minimized by natural selection to allow as much power as possible to be available for foraging , growth , and reproduction . One classical approach minimizes power loss of blood flow due to viscous dissipation and due to cost of blood volume in order to derive that flow rate , Q . , depends on the cube of the vessel radius , r3 . Combining this result with conservation of fluid flow at a branching junction yields Murray’s law , r p 3 = ∑ i r c , i 3 , where rp is the radius of the parent vessel segment and rc , i is the radius of the ith child ( distal ) segment . Another classical approach for the cardiovascular system is to minimize wave reflections in pulsatile flow , as Womersley and West et al . have done [4 , 6 , 26] . This approach leads to area-preserving branching ( or Da Vinci’s Rule ) , so that the sum of the cross-sectional areas ( ∝ r2 ) of child vessels equals the cross-sectional area of a parent vessel at a branching junction . In the WBE model , reflections dominate for large vessels while dissipation dominates for small vessels ( ≪ 1 mm ) . Moreover , in the WBE model , a volume-servicing argument [4] is used to derive an analogous relationship for vessel lengths , l p 3 = ∑ i l c , i 3 , while Huo and Kassab assume the same relationship but allow the exponent to vary from length-preserving ( exponent of 1 ) to volume-servicing ( exponent of 3 ) . Optimization has been a common approach to developing vascular models throughout the past century , but it has been highly debated as to which properties are optimized and what are the tradeoffs among them [4 , 27 , 28 , 37–39] . For instance , Banavar et al . [40] optimize for efficient transport within three-dimensional bodies and Dodds [15] minimizes network volume as required to continually supply metabolites within a body . Indeed , different principles and assumptions lead to a variety of relationships between the flow rate and vessel radius . Consequently , we express a generalized form of Murray’s law r p 1 / a = ∑ i r c , i 1 / a . ( 1a ) in which we define the vascular scaling exponent , a , for vessel radius to be consistent with the notation of Price et al . [1] . The analogous generalization for vessel length is l p 1 / b = ∑ i l c , i 1 / b ( 1b ) where b is the vascular scaling exponent for vessel length . To ease computation , many models further assume that vascular networks are strictly self-similar and symmetrically branching—child vessels all have identical properties within a branching level . In this case , scale factors and associated scaling exponents can be defined for each branching level , k , which represents the number of branching junctions from the heart to that vessel . Following the notation of the WBE model , the scale factors are β = rk+1/rk and γ = lk+1/lk . For dichotomous branching , we can solve for β and γ using Eqs 1a and 1b , to find β = r k + 1 r k = 2 - a and γ = l k + 1 l k = 2 - b . ( 2 ) Furthermore , for these idealized networks , the frequency distributions of radius and length follow power laws with scaling exponents 1/a and 1/b . Because there are N = 2k vessels of radius r = r0 βk , Eq 2 give the two power-law relationships N = ( r / r 0 ) - 1 / a = ( l / l 0 ) - 1 / b . ( 3 ) Similarly , for any vessel , its radius and length are related to the number of downstream endpoints ( e . g . capillaries ) , Nd , by r ∝ N d a and l ∝ N d b . ( 4 ) Eqs 1a–4 constitute four methods of calculating vascular scaling exponents . Each method relies on different types and levels of information . First , for each branching junction , the generalizations of Murray’s law ( Eq 1a ) and volume servicing ( Eq 1b ) can be numerically solved for the exponents a and b by Newton’s method . The solution provides a direct local measurement of the scaling exponents at every junction that we call conservation-based scaling exponents . The value is undefined if a child vessel has radius or length greater than its parent . Second , for each parent-child pair of vessel segments , the ratio of vessel radius and length can be calculated . By Eq 2 , these scale factors can be used to compute a second local measure—our ratio-based scaling exponents . Third , across all vessels and junctions , empirical distributions of radii and lengths can be fitted to power laws to produce what we term the distribution-based scaling exponents , as in Eq 3 . Fourth , across all vessel segments , log-log regressions of radii and lengths versus the number of downstream endpoints can be performed to derive regression-based scaling exponents , following Eq 4 . These latter two methods each provide single values for the vascular scaling exponents , a and b , across the whole network , and they do not rely on information about the geometry of individual branching junctions . In the literature , these four methods for measuring scaling exponents are often used interchangeably [1 , 29 , 41 , 42] . However , these are only proven to be identical for symmetrically-branching , strictly self-similar networks . Furthermore , it is unknown whether measurements of vascular scaling exponents using these four methods ( Eqs 1a–4 ) will produce values that are approximately similar or significantly different . If they differ , this raises questions about which of these four measures of vascular scaling exponents , if any , best corresponds to the scaling relationships predicted by ideal networks or observed empirically for metabolic and growth rates .
We attempt to solve Eq 1a numerically at each branching junction . We label the parent at each branching junction by traversing the vascular network starting at the vessel segment of greatest radius , which we assume is a segment in the aorta . If all children are smaller than the parent , the equation is guaranteed to have exactly one solution , and numerical convergence to the solution is fast . Otherwise , the equation may have zero , one , or two solutions . Such cases may represent real anatomical variation , or misidentification of the parent vessel due to errors or ambiguous topology such as the Circle of Willis . We consider only the simplest case , in which children are smaller than their parent , because otherwise the solutions are difficult to interpret in the context of existing vascular network models . All child radii are smaller than the parent in 82% ( 222 ) of junctions , and all child lengths are smaller in only 35% ( 94 ) of junctions . We show distributions of these measures of the vascular scaling exponents , a and b , across branching junctions in Fig 1 . The arithmetic mean and associated 95% confidence intervals for these conservation-based measurements of a and b are presented in Table 1 . Topological information allows us to compute β and γ directly for each parent-child pair of vessel segments . Our dataset contains 703 pairs of parent and child segments with dichotomous branching . A small proportion of the β and γ values may be over-estimated due to misidentification of the parent-child relationship ( see Methods ) . This bias would produce underestimates in a and b , but because misidentification is very infrequent , the magnitude of this bias is expected to be within the measurement error . The distribution of β and γ is displayed in Fig 2 , and the arithmetic mean and associated 95% confidence intervals of the ratio-based vascular scaling exponents calculated using Eq ( 2 ) are shown in Table 1 . For symmetrically branching , self-similar networks , the frequency of radius and length measurements follow power-law distributions . We did a linear fit to the log-log transformed histograms of radius and length measurements ( the log of Eq ( 3 ) ) using SMA regression . We derived empirical 95% confidence intervals by resampling with different bin sizes and cutoff values for the tail ( see Data Fitting ) . The fits are shown in Fig 3 . The scaling exponents obtained from our fits are given in Table 1 . By taking the logarithm of Eq ( 4 ) , we can estimate the vascular scaling exponents , a and b , by performing regressions of the logarithm of the number of downstream tips ln Nd against the logarithm of radius , ln r , and the logarithm of length , ln l , respectively [41] . Regression lines are shown in Fig 4 . The measured slopes , which are the estimates of the vascular scaling exponents , and associated 95% CI are shown in Table 1 . As in all analyses based on experimental data , measurement errors affect uncertainties in quantities calculated from the raw data . Consequently , we investigate the sensitivity of our calculated scaling exponents and entire analysis to the choice of threshold intensity in our algorithm as well as to Gaussian noise in the image quality . An intensity threshold is used in our algorithm to select the voxels from the image that describe the shape of the vessel lumen . Lower thresholds reveal more vessels , so for our analysis above , we used the minimum threshold that produced reliable segments , as described in Software and Algorithm and S1 Text . Because the threshold affects the boundaries of the vessel lumen , the vessel radius also depends on the threshold . For each image there is a minimum acceptable threshold that we used above , and for our sensitivity analysis , we also chose a maximum threshold to be the largest value for which at least 30 vessel segments are visible . For our plots in Fig 5 , we normalized the threshold to range from 0 and 1 . That is , 0 is the threshold used for the results above , and 1 is the maximum threshold for which at least 30 vessel segments are visible . For each normalized threshold increment of 0 . 05 between 0 . 00 and 1 . 00 , we ran our entire analysis and calculated the four scaling exponents for radius and length ( Fig 5 ) . The values of the scaling exponents at a normalized threshold of 0 . 00 recapitulate Table 1 . Our results remain qualitatively similar as we increase the threshold and include fewer segments . However , at higher normalized thresholds , we can no longer resolve differences between some exponents that are resolvable at normalized threshold 0 . 00 , as expected from the law of large numbers . Noise in images may also cause errors in the identification of vessels or in estimates of the radius or length of vessels . We conducted a sensitivity analysis similar to the above by adding Gaussian noise that varied in magnitude from 0 . 47 ( the measured baseline level of noise in the foreground of one image ) to 4 . 7% ( 10 times the baseline noise level ) of the maximum voxel intensity . Results ( S1 Fig ) show no significant changes in our results with higher levels of noise . As another measure of uncertainty , we located vessels with radius estimates that differed with threshold despite high reliability in vessel identification ( vessel endpoints were similar across at least 5 threshold values ) . We located 12 vessel segments from 9 different patients that matched this criterion . Across these 12 vessels , the mean radius estimate varied from 0 . 9 to 6 . 7mm , but the coefficient of variation ( = standard deviation/mean across thresholds ) ranged only from 0 . 02 to 0 . 08 . Moreover , the coefficient of variation was uncorrelated with radius ( Pearson correlation = 0 . 07 ) , suggesting that individual vessel radius measurements are precise to roughly ±10% ( 2 times the coefficient of variation ) regardless of vessel size . For a specific threshold , we calculate how the measured vessel radius differs from the mean of the vessel radius across all thresholds . At each threshold , this difference tended to be in the same direction ( mostly positive or mostly negative ) across the 12 vessels we measured . Moreover , the magnitude of the difference was proportional to vessel radius ( most R2 > 0 . 99 ) . Consequently , the ratio of vessel radii at any specific threshold is roughly equal to the ratio of the means of the vessel radii across thresholds and also equal to the ratio of vessel radii at any other specific threshold . That is , within the plausible range of thresholds that we explore here , the calculated scaling ratios are largely independent of the choice of threshold value , implying that the choice of threshold does not create any bias in the calculated scaling exponents or ratios . Of course , our sensitivity analysis cannot exclude all possible sources of error or bias . In the Methods sections , we discuss how our results might be affected by other possible sources of error , such as tree topology identification , the patient population , small vessel censoring , vessel lumen misidentification , skeleton line selection , centerline quantization , and vessel segment misattribution . The positional accuracy of MRI is high , so errors arise due to classification or interpretation of voxels . Because the threshold parameter and noise in the image primarily control how we classify voxels—the first step of analysis—these are the major determinants of subsequent errors . In our analysis , as presented in Fig 5 and S1 Fig , no systematic biases are observable , so we conclude that our results are highly robust to the largest and most notable sources of uncertainty . The estimated values of vascular scaling exponents obtained using our four different methods are all presented in Table 1 . All pairs of measures are statistically significantly different ( Welch’s t-test , P < 0 . 01 ) except for the ratio-based and regression-based a ( P = 0 . 18 ) . Values of a based on conservation rules at branching junctions , scale factors for parent-child pairs , and regression of ln r versus ln N are all between a = 1/2 ( the WBE prediction for large vessels ) and a = 1/3 ( Murray’s law , the Banavar et al . prediction [14] , and the WBE prediction for small vessels ) , and the remaining distribution-based a includes a = 1/3 in its 95% CI . The conservation-based exponent a is not statistically-significantly different from a = 1/2 . Different measures for b range from 0 . 17 to 1 . 40 and all are statistically significantly different ( P < 0 . 01 ) from each other and from the volume-servicing and area-servicing values of b = 1/3 and 1/2 respectively . It is notable , however , that the difference between regression-based and distribution-based measurements of b is no longer resolvable at normalized thresholds higher than 0 . The distribution-based and regression-based exponents lie between area-servicing and length-preserving . These discrepancies between measures suggest that vessel segment lengths are poorly modeled by strictly self-similar and symmetrically branching networks .
Our software acquired direct measurements of a large number of connected vessel segments from in vivo angiography . We calculated vascular scaling exponents in these data using four methods to directly compare values from real vascular networks with each other and with theoretical values from the WBE model , Murray’s Law , Banavar et al . [14] , and Huo and Kassab [16] . Intriguingly , our results lead to contrasting conclusions for the changes in vessel radius and length across scale . For the vascular scaling exponent a that quantifies changes in the radius , the conservation-based and ratio-based estimates are closer to a = 1/2 than a = 1/3 . These estimates support work on large vessels by West et al . , Zamir and Banavar [4 , 14 , 43] , in contrast to Murray’s law , which does not distinguish large and small vessels . West et al . derive that the dominant source of power loss for large vessels ( estimated to be r ≫ 1 mm ) is the reflection of pressure waves at branching junctions , while for small vessels , power loss is dominated by viscous dissipation between blood and the vessel walls . Because of the resolution of our MRI volumes , we are able to extract data mostly for vessels with a radius greater than 1 mm , corresponding to the large vessel regime in the WBE model . Consequently , our results for the scaling of radii are supportive of the area-preserving branching of large vessels , corroborate recent findings for plants [41] , and reject the possibility that Murray’s law might apply , either generally or on average , to junctions of large vessels . This result demonstrates that minimizing energy-dissipation due to blood flow does not capture the guiding principles that shape the vascular system across all scales . Future studies using higher resolution angiography ( e . g . , micro-CT ) to obtain data for small vessels are needed to test Murray’s law and the WBE prediction for small vessels and to determine if minimizing energy dissipation is a relevant principle at any scale . For the vascular scaling exponent , b , for vessel lengths , the discrepancies between predicted and estimated values are more difficult to reconcile and interpret . None of the four measures of b agree with each other or provide support for volume- or area-servicing branching , while only the regression-based method provides support for length preservation . Within the WBE model , b is predicted to be 1/3 based on an argument that the vascular network must be volume-servicing for the entire body [4] . This volume-servicing argument has been questioned on theoretical grounds [28] , and here we provide empirical evidence that volume-servicing or any other conservation law for length does not hold locally at branching junctions . Indeed , for the conservation-based exponents , 65% of branching junctions violate the model so severely that exponent values are undefined . The volume-servicing argument is supposed to apply across at least the vast majority of scales and is a key element of the WBE explanation for the 3/4 allometric scaling relationship between metabolic rate and body size . The breakdown between this argument and the real vascular networks we measured may occur because , contrary to the WBE argument , the length of a vessel segment is not a reliable indicator of the volume it services . The correlations between length and number of downstream endpoints , a proxy for volume serviced , is very low compared to the same correlation for radius ( 0 . 2 versus 0 . 7 ) . Considering only the largest vessels , this is not surprising . The ascending aorta is only a few centimeters in length and services most of the body , while the carotid artery is much longer ( at least 10 cm in length ) and services only half the head . Our results imply that either modification of the volume-servicing argument is needed or some new principle yet to be discovered guides the distribution of vessel lengths as the vascular network branches throughout the body . These new developments could lead to corrections to the power-law predictions of the original WBE theory that may agree better with recent findings of “curvature” in the allometric relationship [19] . Beyond the differences discussed thus far , vessel lengths and radii also differ in their distributions for vascular scaling exponents and scale factors . Measurements of a and β exhibit a strong central tendency ( Figs 1 and 2 ) , while the scale factor for length , γ , has a highly skewed distribution with typical values that are not well-described by the mean . Thus , a derivation implicitly based on a mean-value approximation may be successful for predicting vessel radii but fail to predict scaling relationships involving vessel lengths . Thus , while hierarchical symmetric models may fail outright to adequately describe vessel lengths , the discrepancies between vessel radii in real networks and idealized models , such as the WBE model , may only result in minor corrections to model predictions . This may help explain the success of the WBE model in predicting a wide range of phenomena . Differences in results for vessel radius and length could be tied to different strengths of the constraints on vessel geometry . Radii and length distributions have previously been observed to differ in the external branching of plants and leaves [41 , 44] . One explanation for this is that viscous power loss depends much more sensitively on vessel radius ( as a 4th power , ∝ r4 ) than on vessel length ( linearly , ∝ l ) . Thus , the strength of selection for optimal vessel radius is much stronger than for optimal vessel length , implying evolution has more often sacrificed vessel length when negotiating tradeoffs in anatomy . Another potential explanation is that vessel radii are self-similar due to a local constraint at each branching junction , whereas vessel lengths may be constrained only at larger scales—organs and organisms—that more accurately capture how the vascular network needs to span and feed a spatially inhomogeneous body . Disagreement about the value of the length scaling exponent between our four methods indicates that assumptions of the simplest model must be violated so strongly as not to hold even approximately . That is , strict self similarity , symmetric branching or both must be strongly violated for the real vascular networks we measured . Our data reveal pervasive asymmetry in both radius and length between child vessels . How far the results for symmetric networks generalize to asymmetric networks has been explored very little [45] . The differences we observe between different measures of vascular scaling exponents could be explained by the inability of existing theories to account for the asymmetry of real vascular networks . Developing a theory to account for asymmetric branching may be challenging . For instance , accounting for asymmetry would require at least two scale factors for radii ( e . g . , βbig and βsmall ) and two for lengths ( e . g . , γbig and γsmall ) . These additional scale factors and associated scaling exponents would necessarily change our analysis and our estimates for the ratio-based scaling exponent , and would potentially change our interpretation of the distributions of β and γ . Rather than thinking of distributions of β and γ , we would think of joint distributions of βbig and βsmall , for example . For similar reasons , our interpretation and analysis of the frequency distribution of radius and length could be altered , thus affecting the estimates of the distribution-based exponents as well . Angicart outputs β and γ values for each vessel pair , and can provide the detailed information required for future studies of multiple scale factors for length or radius and asymmetric branching . All of the models discussed in this paper ignore any reticulation or loops in the vessel topology , in contrast to recent work on leaf venation networks [46–49] . Our analysis also follows this assumption . However , loops are known to occur anatomically in healthy ( Circle of Willis ) and diseased ( tumors , arteriovenous malformations ) tissue . Extensions to our software and to theory could address this issue . Such an extension could be used to investigate abnormal tumor vasculature [9] , or allow new theory to be developed to explain the normal anatomical function of reticulation . There are also other spatial aspects that have received theoretical attention , such as branching angle , that our software is already capable of recording . Many more tests could be performed with data on microvasculature . For instance , Huo and Kassab [16] have published scaling relationships for how crown volume and length change with stem radius . Testing these requires knowledge of the full crown , down to the microvasculature . Similarly , Dodds [15] makes predictions for virtual vessels that coincide with real vessels only at the smallest scales . We developed new software and applied it to MRI of human head and torso to obtain one of the most detailed datasets for examining branching architecture in vascular networks . In addition , we conducted a comprehensive data analysis that uses both local and global methods to measure scaling exponents . Together , this new software , data , and analysis provides valuable information for answering fundamental questions about vascular system morphology . The public release of our imaging software , angicart , should enable researchers to ground future vascular network theories in empirical data . The software facilitates comparison across spatial scales and between studies by operating uniformly on all tomographic imaging methods . Because imaging can be done non-invasively , our method affords the opportunity to record all spatial information of in vivo vasculature through time or across development . We explain four different methods to estimate vascular scaling exponents from spatially-explicit data . Although researchers use and sometimes interchange these four methods , we found that all four methods can lead to different results , and that for scaling exponents for vessel lengths , these differences can be dramatic . This result is in stark contrast to theoretical calculations for idealized , symmetric networks that predict all methods will give identical values . We advise caution when interpreting different methods and estimates as the same scaling exponent because this could lead to misperceptions and disagreements among studies . For instance , regression-based estimates are the most common across levels of biological organization while distribution-based estimates are used for forests [42] , so comparing these estimates to each other must be done with care . The differences we observe call for a new understanding of the relationships between the local geometry of vessels and the global properties of vascular systems . New theory should be developed to accommodate the anatomical variation and asymmetric branching we observe in real networks .
After local institutional review board approval and written informed consent had been obtained , 20 consecutive adult patients with clinically suspected supraaortic arterial occlusive disease were prospectively enrolled to evaluate new MRI methods for the study of carotid atherosclerosis [50] . Although other data was recorded in that study , we use only the images . In that study less than 0 . 5% of observed vessel segments had notable luminal narrowing , so we conclude patient selection and enrollment did not affect our results . We acquired contrast-enhanced magnetic resonance angiograms ( CEMRA ) of the upper torso , neck , and head in the 20 human subjects ( N = 20 ) using a 3 Tesla Siemens Trio scanner ( Siemens Medical Solutions , Erlangen , Germany ) . The data acquisition details have been previously described [50] . In brief , the CEMRA images were acquired after an antecubittal vein injection of gadolinium based contrast agent ( Gd-DTPA , Magnevist , Bayer Shering Pharma AG , Berlin , Germany ) . The image volumes have dimensions that are typically close to 380 × 640 × 128 voxels , with each voxel nearly isotropic and between 700 × 700 × 800 μm and 800 × 800 × 900 μm . The resolution and imaged volume are typical of high-quality 3T MRI . The point-spread-function for MRI is known to be precise and equivalent to the programmed pixel size [51] . In practice , the geometric accuracy is known to be sub-millimeter [52] . The vessel networks in each image are clearly visible due to the sharp image contrast provided by the presence of the contrast agent , which makes the blood appear bright relative to dark non-blood tissues . We averaged each 2 × 2 × 2-voxel cube of adjacent voxels into a single 1 . 4–1 . 8 mm voxel to remove noise , reduce processing time , and match conventionally-acquired resolutions . This reduced noise-induced errors without substantially changing the number of vessel segments represented . In two of our 20 image volumes , segmentation failed because bright , non-blood tissues were present very close to the blood volume , so that no threshold value excluded all non-blood objects as described in S1 Text . We did not record any vessel segments from these image volumes . We saw no relationship between failure of segmentation and vascular system geometry . We created a free , open source software package—angicart—to read tomographic images of vascular networks , to automatically decompose a vessel lumen into vessel segments , and to measure the geometry and topology of the segments ( Fig 6 ) . The software and data used in this study are available on the internet ( https://github . com/mnewberry/angicart/ ) under a GNU Public License . Our software starts by classifying voxels in a 3D image as part of the vessel lumen if they are within the largest connected group of voxels that exceed an intensity threshold [30 , 53] , as with other level set or thresholding methods . We use a manual binary search to select the threshold value that best matched visual identification of vessels . The result is a 3D binary image , called the network mask . Next , we use spatial criteria to find the endpoints of the vessel network , where the vessels become too small to detect . Given these endpoints , we find the centerline and branch points of vessels by skeletonization [54 , 55] ( See S1 Text ) . We implement skeletonization using an erosion technique—successively removing voxels until no more are removable without disconnecting the endpoints in the mask [56] . The voxels that remain after erosion lie within approximately 1 voxel-width of the true centerline . This information allows us to partition the network mask into segments by attributing each voxel to the segment whose centerline is closest to it . We record how vessels are connected and measure the length of each segment’s centerline and the volume of each segment . Following a geometric argument ( see S1 Text ) , quantization error leads us to overestimate length . We compute radius as r = V / π l . Radii are therefore underestimated on average due to the quantization error in length measurements . This bias affects the accuracy of individual length and radius measurements , but does not bias estimates of scaling exponents ( relative measures ) as long as the percent error does not change systematically with vessel size , which it does not . As a final filter of possibly misclassified vessels , we omit vessels in which more than 20% of the voxels lie further than ( r + 1 ) from the centerline , or whose total volume is less than 4 voxels . Further details of each step are presented in the S1 Text . We determined the conservation-based node scaling exponents by solving Eq ( 1a ) numerically using Newton’s method implemented in OCaml [57] and iterated until the sum of powers was within 0 . 00001 of 1 . We estimated the regression-based scaling exponents using Standard Major Axis ( SMA ) regression of the natural log of radius ( length ) against the natural log of the number of downstream endpoints [58] . We used SMA regression because the variability and uncertainty in the y-axis ( vessel radius or length ) is as large as the variability and uncertainty in the x-axis ( number of downstream endpoints ) . Furthermore , SMA is appropriate because our goal is to obtain the best estimate of the scaling exponents ( slopes ) and not the best prediction of y given x . We estimated distribution-based scaling exponents by fitting the tail of the probability distribution of radius and length to a power law . That is , we binned log-transformed data and determined the slope of the log of probability density versus the log of radius and length using SMA regression [58] . We used 20 bins and discarded 5 and 7 initial bins of radius and length respectively . Blood vessels near the resolution limit of MRI may not be visible . Although dimensions measured from observed small vessels are used in our other methods , counts of small vessels are unreliable due to censoring . Thus , we discarded initial bins in order to exclude vessel sizes where non-uniform censoring of values might occur . We computed standard errors by varying bin size and the number of initial bins discarded ( up to ±3 each ) and using the middle 95th percentile of these values . By binning our data and fitting the power law using SMA regression , we avoid problems that can arise when using maximum-likelihood estimators to fit our power-law distributions . Specifically , the maximum-likelihood estimators are derived with specific assumptions and support choices such as smooth and continuous or discrete and integer , as in Clauset et al . [59] . In contrast , our distributions may be somewhere in between: continuous with an increased likelihood to take values near certain points , such as powers of β times the aorta radius . Our SMA regression on bins of simulated vessel data produced stable estimates with relatively little bias in comparison to fits based on published maximum-likelihood estimators . | Vascular networks distribute resources and constrain metabolic rate . Founded on a few key principles , biological scaling theories predict characteristic patterns for vascular networks as they branch from large to small vessels . These theories also predict seemingly unrelated phenomena , such as size limits on mammals . However , vascular networks are difficult to measure because there are billions of vessels that range in size from meters to micrometers . To test the foundations of biological scaling theories , we developed software that quickly measures thousands of in vivo vessels based on MRI . Data for vessel radii match predicted patterns but lengths do not . Our work suggests the need for new theoretical principles and should facilitate comparisons across organisms , spatial scales , and healthy and diseased tissue . | [
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion",
"Methods",
"and",
"Materials"
]
| []
| 2015 | Testing Foundations of Biological Scaling Theory Using Automated Measurements of Vascular Networks |
Migrating cells are guided in complex environments mainly by chemotaxis or structural cues presented by the surrounding tissue . During transmission of malaria , parasite motility in the skin is important for Plasmodium sporozoites to reach the blood circulation . Here we show that sporozoite migration varies in different skin environments the parasite encounters at the arbitrary sites of the mosquito bite . In order to systematically examine how sporozoite migration depends on the structure of the environment , we studied it in micro-fabricated obstacle arrays . The trajectories observed in vivo and in vitro closely resemble each other suggesting that structural constraints can be sufficient to guide Plasmodium sporozoites in complex environments . Sporozoite speed in different environments is optimized for migration and correlates with persistence length and dispersal . However , this correlation breaks down in mutant sporozoites that show adhesion impairment due to the lack of TRAP-like protein ( TLP ) on their surfaces . This may explain their delay in infecting the host . The flexibility of sporozoite adaption to different environments and a favorable speed for optimal dispersal ensures efficient host switching during malaria transmission .
Optical analysis of cell migration can be performed in simple cell culture assays , in reconstituted 3D-environments , tissue explants or in living animals [1] , [2] . Clearly , the migratory patterns of cells have evolved in the context of the environment they maneuver in . Many cells , such as white blood cells , neurons , metastatic tumor cells or sperm cells as well as bacteria rely on a chemotactic response system to guide them towards their target [3] . Because chemotactic signals are usually subjected to large noise levels , in all of these cases cells follow stochastic search strategies based on different variants of a Brownian random walk [4] . However , a complex pattern of movement inside for example skin or lymphoid tissue can also arise in the absence of chemotactic signals due to stochastic elements in the movement-generating processes or structural constraints in the environment [5] . For example , naive T and B cells were shown by intravital two-photon microscopy to move inside lymphoid organs on independent and apparently random 3D-paths [6] , [7] . Although intravital imaging is arguably a very powerful tool to observe cells in their natural environment , analysis of in vivo data can be hampered by a number of factors such as the micro-architecture of an organ , which can potentially mask a weak chemotactic effect [5] , [8] . Here we analyze motile Plasmodium sporozoites , which are the forms of the malaria parasite injected into the dermis during a mosquito bite [9] , [10] . They move rapidly in seemingly random paths and can invade blood or lymphatic vessels [10] . Sporozoite motility is essential for establishing an infection within the host [11] . It relies on an actin-myosin motor , which is placed just beneath the plasma membrane of the parasite and is linked to the substrate by trans-membrane proteins of the TRAP ( thrombospondin-related anonymous protein ) family to the substrate [12] . Different TRAP family proteins appear to play distinct roles during sporozoites adhesion and motility [13]–[21] . While we begin to understand the molecular basis of parasite motility , little is known about how Plasmodium sporozoites reach the blood vessels after they are injected into the dermis by a mosquito . In principle sporozoites could either follow chemotacic cues [22] or be guided by their physical environment . Curiously , sporozoites migrate on near perfect circular paths on 2D substrates , which extend to spirals in 3D gels [22] , [23] . To migrate within the skin after a mosquito bite presents a formidable challenge to the parasite as the mosquito bite amounts to a strong disruption of the tissue structure ( including haemorrhages ) and as neutrophils rush to the side of infection to clear introduced pathogens [24]–[26] . It is thus not apparent how sporozoites could successfully detect and follow a chemotactic signal in such a situation . The aim of this study is therefore to assess if structural constraints in the skin are sufficient to create the characteristic motility patterns of parasites and guide them through the skin . We reasoned that micro-fabricated substrates featuring pillars of well-defined spacings and feature sizes , often used to investigate the mechanical activity of large vertebrate cells or molecular networks [27]–[30] could serve as adequate obstacle arrays to deflect motile sporozoites , which can be placed in-between the pillars . Applying a wide range of different obstacle geometries we found that parasites can be easily deviated from their circular motion observed on planar substrates . Imaging wild type sporozoites in the absence or presence of actin depolymerizing drugs and sporozoites lacking the TRAP family protein TLP ( TRAP-like protein ) [20] , [21] followed by quantitative analysis of their movement patterns , allowed us to define and experimentally probe key parameters of sporozoite movement . Depending on the distance of the obstacles we found parasites predominantly moving in linear or meandering patterns , preferences also observed in vivo in different skin sites .
To quantitatively assess the behavior of sporozoites inside their natural environment after transmission we carried out in vivo imaging of parasites expressing cytoplasmic GFP [31] inside the skin of the ear pinnae and the tail of living mice ( Figure 1 ) . In both sites sporozoite movement appeared to follow different patterns . While sporozoites in the ear rarely moved on linear paths , those injected into the skin of the tail were frequently seen to move for tens of seconds over several dozen of micrometers on almost linear trajectories ( Figure 1A , B ) . Next we classified parasite movement according to the changing angles between subsequent positions of parasite tracks as movements in a linear , circular or meandering fashion ( see Material and Methods ) . Additionally , parasites were categorized as non-motile when they were attached to the substrate and did not travel forward . Since parasites can vary their movement pattern over time we did not classify each single parasite but plotted the fraction for each pattern over the whole population . This analysis revealed that the fraction of non-motile sporozoites and circular moving sporozoites in the two distinct bite sites were similar at around 30% and 12% of the total observed time points from 589 and 164 sporozoites examined in the ear and tail , respectively ( Figure 1C ) . The fraction of sporozoite movements that occur in a linear fashion was higher in the tail compared to the ear ( P<0 . 001 ) ( Figure 1C ) . Conversely , the fraction of meandering movement was higher in the ear than in the tail ( P<0 . 001 ) ( Figure 1C ) . Sporozoites moving in the tail moved about 20% faster than those in the ear ( Figure 1D ) . We next determined key parameters of sporozoite dispersal . As already apparent from in vivo projections ( Figure 1A , B ) , individual sporozoites moving inside the skin of the ear maintained meandering motility for longer periods of time than sporozoites inside the skin of the tail ( Figure 2Ai ) . Similarly , sporozoites moving in a linear fashion in the tail move both longer and further before changing their direction than sporozoites inside the ear ( Figure 2Aii ) . To determine whether parasite movement follow a random walk pattern [8] , we plotted and analyzed their mean square displacement ( MSD ) over time ( Figure 2B–E , Figure S1 ) . MSD of sporozoites moving in the ear becomes flatter after about 3 minutes of movement indicating a restricted random migration as would be expected from dispersal in the presence of obstacles ( Figure 2B ) . In fact most biological systems do not obey normal diffusion due to internal processes or external factors like obstacles that influence their motion patterns , leading to so called sub-diffusion [32] . The MSD plot for movement in the tail shows a slight increase over time implying a more linear migration as shown in Figure 1C and 2B . Since a linear migration pattern should result in a MSD curve following a 2nd order polynomial function , the nearly perfect fitting ( R2 = 0 . 99 ) to the tail data confirms the predomination of linear patterns ( Figure S1B ) . Other patterns should follow a first-order regression , which was indeed shown by the nearly perfect 1st order fitting to parasite migration in the ear where meandering patterns dominate ( Figure S1B ) . Furthermore , the sporozoites in the tail disperse even more rapidly than would be expected from the difference in speeds ( Figure 1D , 2B ) . Parasites moving in a linear fashion also showed the longest mean free path length ( MFPL ) , i . e . the average distance covered in one type of pattern before changing into another type of pattern ( Figure 2C , Figure S1A–D ) . As a consequence , parasites delivered to the tail covered longer distances than parasites in the ear when moving in a linear fashion . Moreover , weighting the MFPL with the percentage of the observed specific motility pattern showed that parasites moved longer distances in the tail ( ∼34 µm ) compared to parasites moving in ear tissue ( ∼13 µm ) before changing their motility pattern ( Figure 2C ) . To confirm our quantitative imaging analysis , we finally calculated key parameters describing dissemination from the origin , the motility coefficient M [8] and the actual scaling exponent α [33] from the MSD plots ( Figure S1G–I ) . Alpha is dependent on the possibility and frequency for cell deflection influenced by the presence of obstacles ( Figure S1H ) . M is a measure for how fast cells displace from their starting positions during a random walk process ( Figure S1I ) . As expected , M was higher for sporozoites moving inside the dermis of the tail than in the dermis of the ear ( Figure 2D ) reflecting the higher speed of parasites in the tail . Parasites moving inside the tail show an α larger than 1 , due to their dominant linear movement pattern in this tissue ( Figure 1C , 2E ) . Parasites moving inside the ear show an α smaller than 1 ( Figure 2E ) . The latter reflects a sub-diffusion behavior of sporozoites indicating that their motility might be influenced by the presence of deflecting obstacles [34] . To test if deflection of sporozoites contributes indeed to their overall migration paths , we utilized micro-fabricated pillar substrates as defined obstacle arrays . These were generated with individual pillar diameters of 10 µm and pillar-to-pillar distances from 2 to 6 µm ( Figure 3A , Figure S2A , B ) . At least 600 sporozoites ( from at least triplicate experiments ) were tracked and analyzed for each different obstacle array if not stated otherwise . Visual inspection showed that decreasing pillar-to-pillar distance altered the movement pattern of the sporozoites ( Figure 3B ) . Pillar contact and sporozoite speed was uniformly distributed over a parasite population with normal length ( 10 . 5 to 13 . 5 µm ) . Sporozoites placed between pillars spaced 6 µm apart mainly moved in circles around the pillars . In a non-constrained environment like on glass around 50% of the observed parasites turn in circles . On arrays with 6 µm or 7 µm pillar-to-pillar distance the percentage of circling sporozoites was even increased to around 60% . However , with decreasing pillar-to-pillar distances , fewer parasites go in circles with a higher occurrence of complex movement patterns increasing from ∼20% ( 34 of 164 for 6 µm ) up to ∼70% ( 124 of 172 for 2 µm ) ( Figure 3C , S2C ) . Curiously , the number of linear movements increased from around 20% to around 30% when the pillar-to-pillar distance was reduced from 6 to 5 µm ( P<0 . 001 , n = 642 total sporozoites for 6 µm and n = 671 for 5 µm ) but then sharply dropped when the distance was further decreased to 4 µm or 3 µm ( n = 638 for 4 µm and n = 52 of 651 for 3 µm ) ( Figure 3C ) . In contrast , the number of meandering trajectories increased from around 5% to around 50% with decreasing pillar-to-pillar distance . Only ∼20% of sporozoites did not move ( Figure 3C , Table S1 ) . The highest median speed and fastest dispersal was found for sporozoites moving mainly on linear tracks on substrates with 5 µm pillar-to-pillar distance ( 5 µm arrays ) ( Figure 3D , Figure 4A , Figure S2D ) . Interestingly , sporozoites moving in 3 and 4 µm arrays showed a MSD plot comparable to sporozoites moving in the skin of the ear pinnae ( Figure 2B and 4A , Figure S2E , F ) . Similar to the situation in the skin ( Figure 2C ) , parasites moving in a linear fashion in pillar arrays also showed the longest mean free path length ( MFPL ) ( Figure 4B , S2D ) . Weighting the MFPL with the percentage of the observed specific motility pattern ( Figure 4B ) showed that parasites in 5 µm arrays had an overall longer MFPL ( ∼28 µm ) than those in 3 µm arrays ( ∼10 µm ) ( Figure 4B , S2D ) . The motility coefficient M decreased with decreasing pillar-to-pillar distance with values from 1 . 2 µm2/s to 0 . 9 µm2/s ( Figure 4C ) due to higher mean velocities of sporozoites moving in a linear fashion in 6 µm and 5 µm arrays ( Figure 3C , 3D and 4C ) . Similarly , α also decreased with pillar spacing , being larger than 1 . 5 in 5 µm arrays and less than 0 . 8 in 3 or 4 µm arrays ( Figure 4D ) . As the increasing density of obstacles lead to an α<1 for parasite dispersal , this indicated that the presence of obstacles caused a sub-diffusion behavior of sporozoites . We next probed how the speed of sporozoites is related to their dispersal in different arrays . To modulate speed we applied cytochalasin D ( cyto D ) , which inhibits motility in a dose-dependent fashion [35] . Importantly , the application of increasing concentrations of cyto D affected the average speed of the parasites by mainly increasing the proportion of non-moving parasites while decreasing all movement patterns uniformly ( e . g . increase of around 20% for 10 nM cyto D , n = 131 ) ( Figure 5A , 5B ) . As expected with decreasing speeds drug treatment also caused a decrease in MSD ( Figure 5C ) . We next plotted the MSD at certain time points against the average speed with which sporozoites were moving ( Figure 5D ) . This showed a linear dependency at early times of movement and at low speeds but the MSD appeared to reach a plateau at about 1 . 3 µm/s . This value corresponds closely to the average speed of sporozoites observed in vitro as well as in vivo [10] , [13] , [23] , [35] . As there is always a large range of sporozoite speeds within a population of parasites we refined our analysis by plotting the average speed for individual sporozoites from this experiment against the MSD at the end point ( Figure 6A ) . The MSD over speed plot reaches a plateau phase for sporozoites moving in 3 µm arrays , while it increases linearly for those moving in 5 µm arrays ( Figure 6A , B ) . Importantly the fraction of sporozoites moving in a circular or meandering fashion does not increase with increasing sporozoite speed ( Figure 6C ) and can therefore not cause the observed plateau phase ( Figure 6A ) . In 5 µm arrays the linear increase of the MSD to speed plot is due to the increasing proportion of linear movers at higher speed ( Figure 6D ) . This suggests that a difference in MFPL could cause this change . Indeed , the MFPL in 3 µm arrays decreases with increasing speed , while in 5 µm arrays it increases ( Figure 6E , F ) . From this set of data it can be speculated that in 3 µm but not in 5 µm arrays sporozoites achieved an optimal compromise between energy expenditure for speed and dispersal in the environment . Finally , we wondered if we could explain in vivo observations on efficient malaria transmission by comparing dispersal behavior of wild type ( WT ) and mutant sporozoites in pillar arrays . Sporozoites lacking the proteins SPECT-1 , SPECT-2 and CelTOS , the surface phospholipase PbPL and the TRAP family adhesin TLP ( TRAP-like protein ) have all been shown to be impaired in their capacity to cross the dermis [20] , [21] , [24] , [36]–[38] . All mutant parasite lines display a deficiency in migration through cells ( transmigration ) . However , only for one mutant sporozoite line , tlp ( - ) this deficiency could be correlated with impaired gliding locomotion in vitro [19] . tlp ( - ) sporozoites were shown to move at a slightly lower rate on their circular trajectories than WT sporozoites and detached more frequently from the substrate [19] . To investigate tlp ( - ) sporozoites in a quantitative manner in obstacle arrays , we first generated a tlp ( - ) P . berghei line ( strain ANKA ) that expresses green fluorescent protein in the parasite cytoplasm under control of the EF1alpha promoter [39] ( Figure S3A , S3B ) . Sporozoites from isolated clones from this GFP-tlp ( - ) line displayed the same in vitro motility features as the previous non-fluorescent tlp ( - ) sporozoites ( strain NK65 ) , such as reduced average speed and speed distribution over a parasite population [19] , [21] ( Figure S3C , S3D ) . GFP-tlp ( - ) sporozoites also showed similar phenotypes in transmission experiments ( Table 1 ) . Importantly , GFP-tlp ( - ) sporozoites injected by mosquito bite or subcutaneously showed a delay of about one day in the emergence of blood stages , while sporozoites injected intravenously did not ( Table 1 ) . This finding further strengthens the hypothesis that tlp ( - ) sporozoites show a defect in crossing the skin to invade blood capillaries [19]–[21] . We then tracked over 300 GFP-tlp ( - ) parasites for each investigated pillar array ( from at least three independent experiments ) . When monitored in the obstacle arrays GFP-tlp ( - ) sporozoites showed a similar overall change of migration patterns as WT sporozoites . The numbers of circling and meandering sporozoites in obstacle arrays were directly and inversely correlated to the pillar-to-pillar distance , respectively ( Figure 7A ) . As reported for motility on obstacle-free , two-dimensional glass slides [19] , about 10% fewer GFP-tlp ( - ) sporozoites were attached to the substrate and consequently fewer sporozoites were moving in the respective patterns ( Figure 7A , e . g . for 3 µm array an increase of non motile parasites from around 20% ( n = 651 total sporozoites ) to around 33% ( n = 342 ) . Of these , fewer GFP-tlp ( - ) sporozoites were moving for long periods of time being often interrupted by non-motile phases . Analysis of Stop ( speed<0 . 3 µm/s ) and Go ( speed>0 . 3 µm/s ) phases revealed an increase of Stop periods for GFP-tlp ( - ) parasites ( Figure S4A , B ) , which were twice as long in time as those for WT parasites ( Figure S4A , C ) . The Stop phases did not alter significantly between the different motile modes . Like WT sporozoites , GFP-tlp ( - ) sporozoites also moved fastest in obstacle arrays with 5 µm pillar-to-pillar distance ( Figure 7B ) but at slightly lower overall speeds than wild type sporozoites . Accordingly their mean square displacement was lower than that of wild type sporozoites ( Figure 4A , 7C , Figure S3E ) . Few GFP-tlp ( - ) sporozoites migrated for more than 3 minutes as they detached more frequently than the wild type . The MSD of GFP-tlp ( - ) sporozoites was reduced by 40–50% compared to the wild type ( Figure 7C ) . Curiously , however , the MSD over speed plot revealed that GFP-tlp ( - ) sporozoites moving at the same average speed as wild type sporozoites dispersed less ( Figure 7D , Figure S3F ) . This finding shows that average speed is not the sole variable determining sporozoite dispersal in complex environments . Finally , we tested whether the impaired migration revealed in vitro can also be observed in the dermis of ear and tail of mice . In vivo imaging of GFP-tlp ( - ) sporozoites ( near = 274 , ntail = 92 ) from 20 transmission experiments showed that the overall fraction of non motile parasites is increased by around 12% compared to the WT ( Figure 1C , 7E ) . Accordingly , the fraction of motile parasites was uniformly decreased with a consistently higher percentage for linear or meandering movers in tail and ear , respectively . The MSD of motile GFP-tlp ( - ) parasites was significantly decreased compared to the WT ( Figure 2B , 7F ) . The MSD curves , however still followed the characteristic 1st or 2nd order polynomial function for parasite migration in ear and tail ( Figure 7F ) . Finally , both speed and MFPL of migrating GFP-tlp ( - ) sporozoites were also slightly reduced compared to WT ( Figure 1D , 2C , Figure S3G , H ) for sporozoites migrating in the tail but not in the ear . Interestingly , while WT sporozoites migrating in the tail were faster then those in the ear , no such difference was found for GFP-tlp ( - ) sporozoites ( Figure S3G ) .
It was previously suggested that sporozoites isolated from the mosquito hemolymph but not salivary gland-derived sporozoites were attracted to extracts of salivary glands [22] . However , whether naturally transmitted sporozoites , i . e . those from the salivary glands , are chemotactic or not is currently not known . We first investigated migration in different in vivo environments . Sporozoites migrating in the dermis of the tail more often showed movement in linear trajectories than sporozoites migrating in the skin of the ear . Conversely , sporozoites in the dermis of the ear pinnae showed more random migration patterns ( Figure 1 ) indicating increased restriction ( Figure 2 ) . The observed difference between the migration patterns of sporozoites in different tissues may arise from chemical cues . Alternatively the patterns might be due to a difference in tissue architecture or be derived from a mix of both influences . The architecture of the tail skin is indeed known to be different from that of all other tissues [40] . Thus motility could be influenced differently by varying interactions of sporozoites with cells , fibers or other structures in the respective tissue environments of the ear and tail . Similarly , it was recently shown that thymocytes exhibit varying motion patterns within specific regions in the medulla , but not in the cortex of the thymus [41] . As the calculated key dispersal parameters from in vivo imaging suggest that sporozoites follow a sub-diffusion type of random walk , we hypothesized that sporozoites can be guided through the environment by being deflected from obstacles . To test this , we utilized micro-patterned obstacle arrays [27]–[30] . This analysis showed that sporozoites in some arrays adopted complex movement patterns that partially resembled those seen in the skin ( Figure 1–4 ) . Analysis of sporozoites moving in obstacle arrays revealed strikingly similar MSD , MFPL and alpha values to sporozoites moving in the skin of the ear pinnae , for 3 or 4 µm arrays or the tail , for 5 µm arrays ( Figure 1–4 ) . Interestingly , the same trend is also observed for speed , which is highest for sporozoites moving in the tail and those moving in 5 µm arrays . These data together show that movement patterns of motile cells can be influenced by changes in the distances of obstacles . In the case of sporozoites , differences in in vivo migration trajectories can thus be partially reconstructed in vitro using obstacle arrays . We suggest that changes in parasite trajectories are rather due to the reflection of the motile sporozoites by different obstacles then to an actively altered migration pattern , which might be expected in order to optimize the chance of finding a blood capillary . As there are no chemical gradients present in the obstacle arrays , our data suggest that the complex patterns of sporozoites observed in vivo could also be explained by the presence of structural constraints imposed by the environment . However , our data does not allow us to exclude possible contributions to the modulation of parasite migration in vivo by chemical stimuli or 3D structures like collagen fibers , which are missing in the obstacle arrays . Parasite gliding could additionally be modulated by responses towards stimuli based on e . g . immune responses or gradients build up over longer time periods . On the other hand , in vivo imaging did not reveal a clear target-oriented rush of the parasites towards blood capillaries . The pillar experiments thus clearly indicate the importance of considering the role not only of chemical signals but also of physical constraints for sporozoite migration . Future research is necessary to elucidate if chemo- or durotaxis based signals could additionally influence sporozoite migration . While fast-moving mammalian cells like lymphocytes migrate at less than 0 . 1 µm/s , Plasmodium sporozoites move at 1–2 µm/s in the skin , a remarkably fast speed for substrate-dependent locomotion [6] , [10] . Our data allowed us to probe if this rapid sporozoite movement is optimized for an efficient spread in the skin or if sporozoites just move as fast as they possibly can . This question cannot be answered by analysis of in vivo data , as only few fast moving sporozoites can be imaged and tracked for a long enough time in vivo . Also currently no methods are available to rapidly and conditionally manipulate sporozoite speed in vivo . As an alternative and based on the good correlation between sporozoite movement patterns in vivo and in pillar arrays we could manipulate sporozoite speed in vitro ( Figure 5 , 6 ) . Modulating the speed with the actin-disrupting drug cyto D revealed that , as expected , dispersal is directly dependent on the average speed of the migrating parasite population ( Figure 5 ) . Plotting the displacement ( MSD ) of every single parasite against its average speed revealed a linear relationship at low speed for 3 µm arrays ( Figure 6A ) . Unexpectedly , at higher average speed ( between 1 . 5 and 2 . 5 µm/s ) the spread of parasites reached a plateau phase in 3 µm arrays , while it continued to increase in pillar arrays that favored a linear migration pattern ( 5 µm arrays ) ( Figure 6B ) . As no major shift in movement pattern was observed for faster moving sporozoites , the majority of sporozoites gliding at higher speed in 3 µm arrays still displayed a meandering pattern ( Figure 6C ) . The observed plateau phase is therefore likely due to the shorter mean free path length of fast moving parasites , which encounter more obstacles during the same period of time as slow moving ones ( Figure 6C ) . Thus , in environments such as the skin where parasites undergo meandering migration , it can be speculated that the parasite would need considerably more energy in order to further increase its spread if moving faster than 1 . 5 µm/s . Indeed , in vivo ∼90% of the sporozoites move slower than 2 µm/s ( Figure 1D ) , suggesting that sporozoites have adapted to the environmental constraints that limit the mean free path length . We therefore speculate that sporozoites have evolved to reach an optimal level of dispersal relative to energy expenditure for speed . We also investigated parasites lacking the plasma membrane protein TLP , which appears to be important for sustained rapid motility on circular tracks in vitro [19] , for exiting the skin and penetrating and/or growing within the liver [20] , [21] . These parasites dispersed much less in all tested obstacle arrays than WT parasites due to a shift to a larger number of non-motile parasites and a slower migration speed ( Figure 6 ) . However , unlike for parasites treated with cytoD , the percentage of non-motile phases increased due to more frequent and longer resting phases of motile parasites . This is likely due to the adhesion impairment of tlp ( - ) sporozoites [19] . Strikingly , even tlp ( - ) parasites moving at the same average speed as WT parasites showed a smaller MSD ( Figure 7B , 7D ) . This can be explained by the different range of instantaneous speeds for the two different populations ( Figure 3D , 7B ) . GFP-tlp ( - ) sporozoites show a broader range of e . g . 0 . 5 to 2 . 8 µm/s compared to 0 . 8 to 2 . 3 µm/s for wild type sporozoites ( Figure 3C , 7B ) . This shift is likely due to the change in the adhesive capacity of the parasites in the absence of TLP [19] . The high maximum instantaneous speeds of GFP-tlp ( - ) sporozoites in the 3 and 4 µm arrays fall into the plateau phase of the MSD over speed plot and thus limits their contribution to dispersal ( Figure 6A ) . This shows that the average speed is not always directly linked to parasite dispersal and suggests that a careful dissection of parameters describing motility is necessary for the understanding of a parasite line showing a subtle motility deficient phenotype . This also suggests , that the adhesion capacity of sporozoite surface proteins , as exemplified by TLP , influences their efficient dispersal in complex environments . Finally , the decreased dispersal rate could contribute to the significant increase in the time needed for tlp ( - ) sporozoites to cause a blood stage infection if injected by mosquitoes or into the subcutaneous tissue ( Table 1 ) . This would suggest that the motility defects seen in tlp ( - ) sporozoites compared to WT reveal that TLP functions mainly in crossing the skin barrier after transmission . To probe this hypothesis we performed in vivo imaging of GFP-tlp ( - ) parasites . These parasites were less motile and showed only a slight reduction in speed but dispersed less than wild type parasites ( Figure 7 ) . This shows that the mutant line is migration impaired in vivo as well as in vitro . In comparison , however ( Figure 7A–D and 7E–F ) , the reduction of dispersal of tlp ( - ) sporozoites was not so pronounced in vivo as on pillar arrays . In arrays with obstacles spaced 5 or 3 µm apart the MSD was decreased by around 50 to 70% ( from around 10*103 µm2 to around 3*103 µm2 for 5 µm and from around 1*103 µm2 to around 0 . 5*103 µm2 for 3 µm arrays , respectively ) ( Figure 7C ) . In the skin a decrease of around 25% in MSD was revealed ( e . g . from around 1*103 µm2 to around 0 . 8*103 µm2 for ear ) ( Figure 7F ) . The less distinct influence of the loss of TLP for in vivo migration could be explained by the difference between the quasi-2D environment of the pillar substrate and the complex 3D environment of the dermis . For example , a 3D surrounding could better compensate for adhesive defects as parasites detaching from one side could adhere at the other side almost simultaneously , which is obviously not possible to the same degree in pillar arrays . Additionally , factors induced by the inflammatory response to the mosquito bite or the mosquito saliva were not present in pillar arrays . We currently do not know if these factors influence parasite motility . Also , the stiffness of a PDMS obstacle array is clearly different to the more elastic skin environment . Concerning TLP function , we can also not rule out that this protein also functions in blood vessel entry . However , despite these shortcomings , we could successfully use obstacle arrays to reveal differences in parasite migration patterns that resemble in vivo migration much better than those observed on plain 2D surfaces , where sporozoites simply glide in circles . Micro-patterned obstacle arrays can be used to deepen our understanding of how parasites travel trough the skin in order to reach the capillary system and establish infection . The characterization of a genetically defined parasite mutant further illustrates how micro-patterned obstacle arrays , in combination with transmission experiments , could be used to rapidly screen for motility defects during host switch . Whole cell assays , such as the one described herein , in combination with diverse small compound libraries offer a potential for anti-malaria lead substance discovery [42] , [43] . Importantly , these arrays could also be adapted for use with other cells to similarly check for defects in migration behavior of a variety of cell types and pathogens .
All animal experiments were performed concerning FELASA category B and GV-SOLAS standard guidelines . Animal experiments were approved by German authorities ( Regierungspräsidium Karlsruhe , Germany ) , § 8 Abs . 1 Tierschutzgesetz ( TierSchG ) . We first designed a master mask using a home-written program called Phyton and converted it with the mask-writer software DWL-66 ( Heidelberg Instruments ) . The individual pillars were arranged in hexagonal patterns to assure the distances of 2 to 6 µm between pillars in any direction ( Figure S2A ) . The master mask was produced by ML&C Jena . The PDMS micropillar substrates were then made by photolithographic and replicate molding techniques as described in [29] . To allow wetting of the hydrophobic pillar substrates they were first treated chemically with Extran , an alkaline solution ( Merck , Germany ) in a 1∶10 dilution in H2O for 20 minutes while gently shaken . Afterwards the structures were washed three times for 10 minutes in distilled water . A silicon flexiPERM chamber ( GreinerBioOne ) was used for imaging parasites . The chamber surrounds the micropillar substrate and maintains a stable environment without flow ( Figure S2B ) . For the motility studies on pillar substrates the sporozoites in RPMI containing 3% BSA ( bovine serum albumine ) were added to the flexiPERM chamber surrounding one pillar field . For the cytochalasin D treatments the parasites were preincubated with the different drug concentrations for 15 minutes . Fluorescent P . berghei ( NK65 ) or GPF-tlp ( - ) ( ANKA ) sporozoites [31] ( see section: Generation of GFP-tlp ( - ) sporozoites ) were produced and prepared as described [35] , [44] . Comparison of parasite motility in vitro showed no difference between NK65 and ANKA strain parasites ( Figure S3 ) . Imaging of sporozoite motility on the different obstacle arrays was performed on an inverted Axiovert 200M Zeiss microscope using the GFP filterset 37 ( 450/510 ) at room temperature . Images were collected with a Zeiss Axiocam HRM at 1 Hz using Axiovision 4 . 8 software and a 10× , 25× or a 40× Apoplan objective lens ( NA = 0 . 25 for 10× ) . A single DIC ( differential interference contrast ) image of the substrate was taken before and after the analysis to merge the PDMS pillars with the time-lapse series of motile sporozoites recorded in the GFP channel . For the motility analysis , a time lapse of one image per second for a total of 300 frames was taken . For each obstacle array at least three different region of interest ( ROIs ) were recorded . For each substrate the experimental setup was repeated several times on different days with sporozoites harvested between 17–21 days after mosquito infection . Thus the motility parameters of at least 600 sporozoites for each pillar array were recorded . The DIC image and the GFP movie were overlaid and analyzed with ImageJ ( http://imagej . nih . gov/ij/index . html ) . For visualization , the pillars were illustrated in red and the maximum intensity projections of the parasite trajectories in green ( Figure 3B , Figure S2B ) . NMRI mice ( Charles River ) were anaesthetized by intraperitoneal injection of 70 µl ketamin/xylazin ( 200 µl/50 g ) and the ear pinnae or the tail were exposed to infected mosquitoes inside a gauze-covered beaker as described [10] , [13] . Mosquitoes were infected either with wild type ( NK65 ) GFP fluorescent parasites [31] or GFP-tlp ( - ) parasites ( ANKA ) ( see section: Generation of GFP-tlp ( - ) sporozoites ) . For each experiment mosquitoes were allowed to probe for 1–3 minutes at the exposed site before the mouse was placed on the microscope table of an inverted microscope . Parasites inside the skin were filmed using a GFP filterset either on an inverted widefield Axiovert 200M Zeiss microscope ( using Axiovision 4 . 6 software and a 25× LCI Plan-Neofluar objective ( NA 0 . 8 ) ) or a PerkinElmer UltraView spinning disc confocal unit on an inverted Nikon TE 2000-E microscope ( using PerkinElmer Ultraview Software and a 20× objective ( PlanFluor multi-immersion , NA 0 . 75 ) . Images were acquired at 0 . 2 or 1 Hz . The mouse was kept at a stable temperature using a preheated incubation chamber or a heating blanket . The plasmid used to generate the tlp ( - ) parasites [21] has been transfected into a Plasmodium berghei ANKA strain expressing cytoplasmic GFP [39] using standard transfection methods [45] . Clonal lines were obtained by limited dilution into 15 recipient NMRI mice . Genotyping of recombinant parasites was performed by gDNA extraction and integration PCR with the same primer combination as used by Heiss et al [21] ( Figure S3A ) . Mosquitoes were infected with GFP-tlp ( - ) parasites and sporozoites isolated from infected mosquitoes as described [44] . To validate that the tlp gene was deleted in GFP-tlp ( - ) parasites transcript detection was performed by RT-PCR from total RNA obtained from gradient-purified schizonts or salivary gland-associated sporozoites ( Figure S3B ) . GFP-tlp ( - ) and , as controls , WT parasites were used for RNA isolation and reverse transcription . cDNAs were synthesized from 2 µg of total RNA using Retroscript ( Ambion ) . Reference pools ( − ) were obtained in the absence of reverse transcriptase . For detection of TLP transcripts we used primers PbTLPfor_1 and PbTLPrev_1 [21] and Pbaldolase primers ( Aldolase_for 5′ TGTATTTAAAGCTTTACATGATAATGG 3′; Aldolase_rev 5′ TTTTCCATATGTTGCCAATGAATTTGC 3′ , expected size: ∼450 bp ) for transcript controls . 4–6 C57Bl/6 mice were infected either intravenously ( i . v . ) with 100 WT ( ANKA GFP ) or GFP-tlp ( - ) sporozoites ( ANKA ) , subcutaneously ( s . c . ) with 1 , 000 parasites or by bite ( bite ) using 10 or 5 mosquitoes per animal . The infected mosquitoes were pre-selected based on their GFP signal in the midgut one day prior to the transmission experiment and kept in a standard insectary incubator ( Sanyo incubator ) without the normally present sugar pad . Giemsa-stained smears from tail blood were monitored daily . The prepatent period is the time until the first detection of an erythrocytic-stage parasite in Giemsa-stained blood smears after infection . The parasites were either tracked manually using the ImageJ manual tracking plugin or with an automated tool , ToAST [35] . Designation of movement as either circling or complex was first performed manually by differentiation of circling parasites and parasites that were not turning in stable circles ( = complex pattern ) . For an automated analysis we generated a MATLAB script , which distinguishes between attached , circular or linear moving or meandering parasites according to the angular change during motility with the following constraints: motile sporozoites that changed their angle between 0 and 12 degrees from one frame to the next were classified as moving in a linear fashion . For angles between 12 and 45 degrees and >45 degrees the sporozoites were classified as moving in a circular and meandering fashion , respectively . Parasites moving slower than 0 . 3 µm/s were assigned as non motile . Using the MATLAB script , parasites were automatically classified and the speed , the mean square displacement ( MSD ) , the actual scaling exponent α and the motility coefficient M were determined for motile parasites . Random walk patterns were analyzed by plotting the average mean square displacement from the origin of the motility against time [8] . The MSD was calculated for motile parasites moving faster than 0 . 3 µm/s [35] and in linear and meandering patterns , excluding circling parasites ( Figure S1C , S1D ) . The slope of the linear regression on a logarithmic MSD plot corresponds to α while M can be derived from the offset [33] ( Figure S1G ) . Linear regression was performed using MATLAB . The visualization of the trajectories was also performed in MATLAB . 589 sporozoites in the ear and 164 sporozoites in the tail tissue were tracked and analyzed from over 25 transmission experiments ( Figure 1 ) . At least 600 sporozoites were tracked and analyzed for each different pillar array ( Figure 3 ) . Between 150 and 200 sporozoites were counted for the manual analysis ( Figure S2C ) . For each drug treatment we tracked and analyzed more than 150 parasites ( Figure 5 ) . We tracked over 300 GFP-tlp ( - ) parasite for each investigated pillar array . GraphPad Prism was used for graphing and statistical analysis . Mean and standard deviations were plotted for each graph if not stated otherwise and Students t-test was performed . For not normally distributed unpaired data ( in vivo data ) the Mann-Whitney test was used ( Figure 1D , 3D , 7B , S3G ) . The range of whiskers plots indicates the 2 . 5–97 . 5% percentile , the box includes 50% of all values and the horizontal bar shows the median . The threshold alpha of the p value was set to 0 . 05 ( * ) , 0 . 01 ( ** ) and 0 . 001 ( *** ) . | Guidance of motile cells plays an important role during the life of a multi-cellular organism from early embryogenesis to the intricate interactions of immune cells during an infection . These migrations , like those of pathogens , can be directed by both chemical and physical cues . The malaria parasite needs to migrate immediately after being injected into the skin of the host by a mosquito bite . The malaria parasite forms deposited in the skin are called sporozoites . These must penetrate the dermis of the host to reach and enter a blood vessel . It is not clear if the sporozoites follow chemical cues or rely on the physical context of the environment . We show here , using in vivo imaging that sporozoites migrate along different paths in different skin environments . Introducing a novel assay for the study of cell migration in general we show that these in vivo paths can be largely recreated in vitro by placing sporozoites in a micro-patterned environment . This shows that environmental constraints are sufficient to guide sporozoite migration in the skin dermis . We further speculate that sporozoites have evolved to migrate at the fastest speed possible for efficient dispersal and show that a parasite lacking a surface protein has substantial defects in tissue dispersal and thus cannot efficiently infect the host . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
]
| [
"medicine",
"cell",
"motility",
"biological",
"data",
"management",
"parasitic",
"diseases",
"parasitology",
"zoology",
"infectious",
"diseases",
"cell",
"adhesion",
"biology",
"vectors",
"and",
"hosts",
"biophysics",
"physics",
"cell",
"biology",
"malaria",
"computational",
"biology",
"molecular",
"cell",
"biology"
]
| 2011 | Environmental Constraints Guide Migration of Malaria Parasites during Transmission |
Stoichiometric balance , or dosage balance , implies that proteins that are subunits of obligate complexes ( e . g . the ribosome ) should have copy numbers expressed to match their stoichiometry in that complex . Establishing balance ( or imbalance ) is an important tool for inferring subunit function and assembly bottlenecks . We show here that these correlations in protein copy numbers can extend beyond complex subunits to larger protein-protein interactions networks ( PPIN ) involving a range of reversible binding interactions . We develop a simple method for quantifying balance in any interface-resolved PPINs based on network structure and experimentally observed protein copy numbers . By analyzing such a network for the clathrin-mediated endocytosis ( CME ) system in yeast , we found that the real protein copy numbers were significantly more balanced in relation to their binding partners compared to randomly sampled sets of yeast copy numbers . The observed balance is not perfect , highlighting both under and overexpressed proteins . We evaluate the potential cost and benefits of imbalance using two criteria . First , a potential cost to imbalance is that ‘leftover’ proteins without remaining functional partners are free to misinteract . We systematically quantify how this misinteraction cost is most dangerous for strong-binding protein interactions and for network topologies observed in biological PPINs . Second , a more direct consequence of imbalance is that the formation of specific functional complexes depends on relative copy numbers . We therefore construct simple kinetic models of two sub-networks in the CME network to assess multi-protein assembly of the ARP2/3 complex and a minimal , nine-protein clathrin-coated vesicle forming module . We find that the observed , imperfectly balanced copy numbers are less effective than balanced copy numbers in producing fast and complete multi-protein assemblies . However , we speculate that strategic imbalance in the vesicle forming module allows cells to tune where endocytosis occurs , providing sensitive control over cargo uptake via clathrin-coated vesicles .
Protein copy numbers in yeast vary from a few to well over a million[1 , 2] . Expression levels , along with a protein’s binding partners and corresponding affinities , are critical determinants of a protein’s function within the cell . In the context of multiprotein complexes–especially obligate complexes such as the ribosome–it is thought that protein concentrations are balanced according to the stoichiometry of the complex . This is referred to as the dosage balance hypothesis ( DBH ) [3–5] . Here , we expand this hypothesis to a network wide level , where proteins participate in multiple distinct complexes as well as transient interactions . In these more complex networks ( Fig 1A ) , balance can be defined as having just enough copies of each protein to construct a target vector of complex abundances , with no proteins ( or protein binding sites ) in significant deficiency or excess . This generalized definition of balance reproduces the expected result for obligate complexes , where , for example , the ARP2/3 obligate complex ( Fig 1B ) would be balanced if all subunits had equal copy numbers . For obligate complexes , dosage balance means that there are no leftover subunits , as these would be a waste of cell resources . However , even for proteins in non-obligate complexes a number of deleterious effects could be caused by imbalance . An overexpressed core or “bridge” subunit may sequester periphery subunits , paradoxically lowering the final number of complete complexes[5 , 6] . Excess proteins may be prone to misinteractions , also called interaction promiscuity , with nonfunctional partners . Numerous studies have identified proteins with high intrinsic disorder as sensitive to overexpression[7–9] , and these proteins have low , tightly regulated native expression levels[10 , 11] indicating that misinteraction propensity and abundance are related . Underexpression carries its own dangers: a single underexpressed subunit will become a bottleneck for the whole complex . In addition , weakly expressed proteins are noisier[12] and thus less reliable for the cell . Male ( XY ) animal cells are known to employ “dosage compensation” mechanisms to increase the expression of X-chromosomal genes to be on par with female cells[13 , 14] , though for other genes it is the female cell that cuts expression levels in half[15] , indicating that the cell preserves an optimized set of expression levels . But optimized does not necessarily mean balanced . Imbalance may be necessary for functional reasons: signaling networks utilize underexpressed hubs to regulate which pathways are active as a given time[16] . Recent models show imbalance can be beneficial to complex assembly when affinity and kinetics are taken into account[17 , 18] . A study of over 5 , 400 human proteins by Hein et al . found that strong interactions forming stable complexes are correlated with balance , but weak interactions are not , which may mean that the network as a whole is not balanced [19] . Finally , the concept of dosage balance being an optimal set of protein copy numbers generally relies on the assumption that proteins reach an equilibrium state of complex yield . Most processes in the cell do not occur at equilibrium and therefore deviations from balance could be beneficial in non-equilibrium models . Here , we test the hypothesis that protein expression levels are significantly biased towards balance , even for complex PPINs that include weak and transient interactions . This first required us to develop a method to quantify stoichiometric balance in any arbitrary PPIN , given known binding interfaces and some observed copy numbers , which we call Stoichiometric Balance Optimization of Protein Networks ( SBOPN ) . Copy number correlations thus are evaluated beyond direct binding partners to the more global network of interactors . We then can quantify the consequences of imbalance relative to perfect balance according to two criteria: 1 ) the deleterious consequences and cost of forming misinteractions , and 2 ) the potentially beneficial control of specific functional outcomes by modulating which complexes , given known binding affinities , actually assemble . Applied to the 56-protein , manually curated , interface-resolved CME PPIN [20] , two of its sub-networks , as well as the ErbB PPIN[16] , we find that stoichiometric balance in observed copy numbers is often significant , and observed imbalances , particularly of underexpressed proteins , could provide tuning knobs for functional outcomes . The first consequence of imbalance we evaluate , misinteractions cost , has an indirect effect on function by allowing unbound proteins to bind to non-functional partners , sequestering components and thus affecting formation of specific complexes . They are believed to play a role in dosage sensitivity[7 , 8 , 21] , and avoiding them has been shown to be an evolutionary force limiting protein diversity[22 , 23] , expression levels[24 , 25] , binding strengths[26] , and protein network structure[23 , 27] . Misinteractions , not being selected for by evolution , are weak and generally unstable , but there are far more ways for N proteins to misinteract ( order N2 ) than bind to their few functional partners ( order N ) [22 , 23] . Cells have evolved a variety of mechanisms to increase specificity , such as allostery[28 , 29] , negative design[30 , 31] , compartmentalization[22] , and temporal regulation of expression[32] . Copy number balance would be another such mechanism , as protein binding sites would saturate their stronger-binding functional partners . The second and ultimately more direct consequence of imbalance we evaluate is that changes to copy numbers control which specific and functionally necessary complexes can form . When the central clathrin protein is knocked out in cells , for example , clathrin-mediated endocytosis ( CME ) is terminated , as clathrin is functionally irreplaceable[33] . The plasma membrane lipid PI ( 4 , 5 ) P2 is also essential for CME , as it is required for recruiting the diverse cytosolic clathrin-coat proteins to the membrane to assemble vesicles[34] . Many clathrin-coat proteins , however , can be knocked out without fully terminating CME[35] . As the CME network illustrates ( Fig 1 ) , most of these proteins have multiple domains mediating interactions involving both competitive and non-competitive interactions . Adaptor proteins ( proteins that bind to the membrane , to transmembrane cargo , and often to clathrin as well ) exhibit redundancy in their binding partners that can partially explain how knock-outs to one protein can be rescued by the activity of related proteins . With simulation of simple kinetic models , we can then test these hypotheses , including for the non-equilibrium production of vesicles at the membrane . Although these models are far too simple to recapitulate the complexities of CME in vivo , they are nonetheless useful in highlighting potential bottlenecks in assembly due to copy numbers or binding affinities . Quantifying balance in protein networks can thus lead to new insights , as unbalanced proteins may serve as assembly bottlenecks , or maintain alternate cellular functions outside of the network module being analyzed[18] . Dosage balance is also important for understanding dosage sensitivity[4 , 21] , a phenomenon where overexpression of a gene is detrimental or even lethal to cell growth . Studies estimate ~15% of genes in S . cerevisiae to be dosage sensitive[9 , 36] , but the negative effects of gene overexpression have been observed in several eukaryotic species including maize[4] , flies[37] , and humans[38–40] . Studying balance at a network-wide level is challenging because it requires resolved information about the interfaces proteins use to bind . A protein that binds noncompetitively with two partners requires equal abundance to its partners . But if the binding is competitive–i . e . the same interface is used to bind two different partners–the protein’s abundance must equal the sum of that of its partners to have no leftovers ( Fig 2 ) . Classic protein-protein interactions networks ( PPINs ) lack this resolution , but recent studies have begun to add this information , creating what we refer to as interface-interaction networks ( IINs ) [16 , 20 , 41] . An IIN tracks not just protein partners but also the binding sites that proteins use to bind . Our study of stoichiometric balance in larger , interface resolved PPINs is organized in the Results section in three parts . In the first part , we define a metric for quantifying stoichiometric balance and how noise in protein expression levels can be approximately accounted for . We apply our algorithm SBOPN to the CME PPIN [20 , 41] and the ErbB PPIN [16] , highlighting which proteins are over- and underexpressed relative to perfect balance . Although this analysis excludes temporal expression and binding affinity , it provides a starting point for the analysis of these features in the subsequent parts . In the second part , we switch to generalized interface-interaction network ( IIN ) topologies and network motifs to focus exclusively on how our first evaluation criteria , the cost of misinteractions under imbalance , is worse for strong binding proteins and for network topologies that resemble biological networks . In the third part , we return to the interface-resolved CME PPIN to evaluate the observed degree of stoichiometric balance in two smaller sub-networks of the CME network: the 7-subunit ARP2/3 complex and a simplified , nine protein , clathrin-coat forming module . In these sub-modules , we now can also evaluate our second criteria and assess how observed copy numbers influence proper multi-protein assembly given known binding affinities of interactions . Our simulations of ( non-spatial ) kinetic models demonstrate that stoichiometric balance does , in fact , improve multi-protein assembly relative to observed copy numbers , even for the nonequilibrium clathrin-coat assembly module . We speculate that the observed imbalances in clathrin adaptor proteins could offer a mechanism for making the vesicle formation process more tunable , since adaptor proteins are responsible for selecting cargo for endocytic uptake , which is the ultimate purpose of CME .
For a multi-subunit complex such as the ribosome or ARP2/3 complex ( Fig 1B ) , all subunits bind together non-competitively to assemble a functional complex . Stoichiometric balance is simply having enough of each subunit to form complete complexes , with no subunit in excess . But quantifying balance in a general protein-protein interaction network is more challenging because some proteins will bind competitively , using the same interface for multiple interactions . Such proteins will need a higher concentration in order to saturate their functional partners ( Fig 2 ) . Thus , to establish stoichiometric balance in a PPIN the binding interfaces must be known . In previous work we analyzed several interface-resolved PPINs , including the 56-protein clathrin-mediated endocytosis ( CME ) network in yeast [20 , 41] ( Fig 1A ) , and the 127-protein ErbB signaling network in human cells[16] . To balance a network , a number of desired complexes may be assigned to each edge and then the number of required interface copies directly solved for . This is constrained with a starting set of copy numbers , C0 , otherwise the solution would be arbitrary . However , the inclusion of multiple interfaces per protein introduces a new constraint: interfaces on the same protein should have the same copy number . This constraint often makes nontrivial solutions ( i . e . when none of the proteins are set to zero ) impossible ( see Methods ) . Therefore , we treat it as a soft constraint , using a parameter “α” to balance its influence . A high α allows more variation of interface copy numbers on the same protein ( Fig 2C ) . We constructed and minimized an objective function using quadratic programming ( Methods ) , which produces a new , optimally balanced set of copy numbers , Cbalanced . For any given interface-resolved PPIN , there can be multiple locally optimized solutions of balanced copy numbers . In Fig 2C we illustrate solutions found by our algorithm SBOPN using the copy numbers of Fig 2B as C0 . If we apply our algorithm to Fig 2A , which is an already balanced network , it simply recovers the input copy numbers , such that Cbalanced = C0 , regardless of α . Because our algorithm minimizes distance from C0 to Cbalanced , the optimal solutions produce both under and overexpressed proteins . The benefit of this method is that the distance between C0 and Cbalanced gives you a relative estimate of how “balanced” C0 already is , and thus a metric from which to evaluate the significance of balance in the observed copy numbers . Using real copy numbers taken from Kulak et al . [2] , Creal , as C0 , we calculated both chi-square distance ( CSD ) and Jensen-Shannon distance ( JSD ) between Creal and Cbalanced ( Methods ) . The former metric looks at differences between absolute values and penalizes high deviations more strongly than low deviations , whereas the latter converts both vectors to distributions and measures the similarity between them . We do not expect any networks to have Creal that is already perfectly optimized , such that Creal = Cbalanced . To establish the significance of both distance metrics , we generated 5 , 000 sets of random C0 vectors , sampled from a yeast concentration distribution . We then measured the CSD and JSD from C0 to Cbalanced for each of these random copy number vectors . If Creal is balanced , its distance metrics should have a significant p-value relative to yeast copy numbers selected randomly from the yeast distribution . The C++ code for our SBOPN algorithm and example input and output files may be downloaded at https://github . com/mjohn218/StoichiometricBalance . In this second part , we investigate how the cost of imbalance , measured solely in terms of misinteractions , depends on general properties of proteins , including binding affinity and number of binary partners . In a stoichiometrically balanced network , proteins will be driven to saturate their stronger-binding functional partners . Any “leftover” proteins , however , may misinteract , or form non-functional complexes that , while weak , are combinatorically numerous . In the second part , “Imbalance increases misinteractions dependent on the network topology and binding affinities of proteins” , we only studied binary , competitive interactions . But proteins often bind noncompetitively into higher complexes , and they may interact weakly and thus form few complexes , in which case imbalance may have functional benefits [17 , 18] . Furthermore , the above models looked at equilibrium results , whereas many biological systems exhibit non-equilibrium dynamics . We created kinetic models of two modules from the CME network with observed imbalances: the ARP2/3 complex and a simplified vesicle forming protein subset . Simulating higher complex formation is challenging because of the exponentially large number of possible species , so we used NFSim[48] , a stochastic solver of chemical kinetics that is rule-based , enabling an efficient tracking of higher-order complexes as they appear in time .
The metrics and SBOPN algorithm we have developed objectively determine whether a protein is under or overexpressed relative to not only its direct binding partners , but to a larger network including partners of partners . This global evaluation is thus sensitive to the size of the network , but directly captures how the multiple binding interfaces of a protein can control its competition for binding partners . In the interface-resolved CME network , we have shown evidence of imperfect , but statistically significant stoichiometric balance . However , the original 56-protein network was overall unbalanced due to the high overexpression of the actin binding protein cofilin . The size of the network clearly matters , in the small modules , we are statistically out-of-balance , but on a larger scale , still in balance . Outliers are emphasized in smaller networks . At the same time , leaving out additional partners can provide some explanation for the observed imbalance . Imbalance may also indicate possible missing interactions in the network . Despite the simplicity of our metric , our method was still able to highlight both correlated concentrations and proteins that violate balance for functional reasons , such as the kinase PRK1 . Furthermore , the observed balances can suggest possible mechanisms of assembly , for example , that can then be studied using kinetic modeling , as we did here . What our results emphasize is that correlations are highly important: functionality can be obliterated with significant imbalance , and misinteractions can also be overwhelming with significant imbalance . Although we only applied our stoichiometric balance analysis to the 56 protein CME network , two smaller modules of this network , and the 127-protein ErbB network , these networks are significantly larger than the obligate complexes previous studied for copy number balance[5 , 6] . Our networks also contain a much larger variety of binding interaction strengths and competitive and non-competitive interactions . As we showed above , balance depended on the protein network’s underlying IIN . While it would be beneficial to repeat this analysis on a larger network , there is a paucity of manually curated IINs in the literature . There are various larger automatically constructed IINs , constructed with homology modeling[77 , 78] , but our previous work found these automatic IINs suffer from various inaccuracies and differ significantly from manually curated IINs in topology[41] . The SBOPN method only accounts for the binding interface network structure and observed copy numbers . A missing feature of our stoichiometric balance metric is that proteins within a network can be expressed with both spatial and temporal variation . For a small binding network this is not a major concern , since proteins in the same complex tend to be co-expressed[79] and co-localized so they may bind . But as network size is scaled up , the probability of all proteins being equally present reduces . Such temporal and spatial variations could be taken into account in the construction of the network , leaving out proteins that are not functional at the same time . A natural extension to our measure of stoichiometric balance would be to also account for binding affinities of interactions in addition to the binding interface network structure and observed copy numbers . Our results here and previous studies[19] indicate that balance should be more tightly constrained for strong binding proteins . However , one benefit to leaving affinities out of the measurement is that biochemical data is in even more limited availability than binding interface data . Our existing metric can thus be much more easily applied to a variety of networks . Furthermore , by picking out highly correlated expression levels , our method can then indicate which interactions might be quite strong , or vice-versa , which may be transient or weak . In this study we used yeast copy numbers from Kulak et al . because it was the most comprehensive . The other three studies we used for comparison did not cover all 56 proteins in our network . However , for the proteins we could compare , we found significant discrepancies between relative abundances . Light chains are weakly expressed in other studies , for example[1 , 49 , 50] . A few possible reasons for this exist . The first is that fluorescence data is inherently noisy . Experimentalists must deal with background noise , interference with protein localization due to the large fluorescent tags , and cross interactions with other proteins[80] . The second is that cell lines can accrue mutations over time that decrease or increase gene expression , a phenomenon observed with HeLa cells[81] . Finally , cells may alter gene expression for regulatory reasons , so the environment in which cells are grown may alter gene expression . We do not expect the cell to perfectly optimize the yield of all of its many assemblies . Each network we have evaluated here is ultimately part of a larger , global cellular network . Perfectly optimizing isolated , local modules does not appear to be a significant pressure for the cell , particularly when a sufficient balance , such as we observe for the vesicle-forming module , maintains functionality . Additionally , these processes , such as in the vesicle forming model discussed below , typically do not occur at equilibrium . Therefore , the concept of minimizing ‘leftover’ proteins based on expected equilibrium complexes formed is a simplification . Correlations in copy numbers are nonetheless often significant relative to randomly assigned copy numbers . We found that copy number imbalance can lead to misinteractions and the features of biological IINs ( power-law-like degree distribution , square and hub motifs , sparseness ) typically have less misinteractions under balance copy numbers but more misinteractions under imbalance . These networks thus should require more tightly controlled balance to avoid misinteractions . But misinteractions are of course not the only pressure on copy numbers . For multi-protein assembly in an obligate complex ( ARP2/3 ) and in a minimal model of vesicle formation for CME , we found that the functional cost of imbalance was dominated more by its impact on determining specific functional complexes than avoiding misinteractions . Nonetheless , the fact that misinteractions can decrease vesicle formation , by sequestering away adaptor proteins into large aggregates , shows that misinteractions are worse than simply having an excess of free proteins . If this result can be generalized , it may have important implications for mechanistic modeling of biological systems , as misinteractions or system error is rarely taken into account . Although the functional effects of copy number balance are usually discussed in the context of number of complete complexes at equilibrium , we have shown that non-equilibrium dynamics can be affected as well . While the clathrin heavy chains and light chains were balanced with each other , they were overexpressed compared to their adaptor proteins , and this limited the frequency of vesicle formation . Although we found that perfectly balanced copy numbers therefore improved vesicle formation frequency compared to observed copy numbers , we speculate that specific imbalances could still be selected for evolutionarily . There are various possible reasons for this imbalance: the function of endocytosis is cargo uptake , and there is a cargo loading process before endocytosis occurs . [75 , 76] Hence to maximize function , controlled endocytosis around high-cargo areas of the membrane may be preferably to frequent , spontaneous endocytosis , and the adaptor proteins can serve as an intentional bottleneck in the process . Clathrin , which cannot directly bind to the membrane , may be kept at a high expression in the cytosol so that there are enough triskelia to quickly form a vesicle no matter where the endocytic site occurs . However , the observed underexpression could also be because there are other adaptor proteins not included in our model , or because clathrin interactions have weaker affinities than interactions between adaptor proteins and must saturate them . Finally , the predictions of our minimal vesicle-forming model are ultimately limited by the approximations we made to simulate the clathrin coat assembly and vesicle formation . Our model vesicles formed about 10 times faster than is observed in vivo . To fully capture the dynamics of this complex process , an ideal model would include all the proteins in our CME network ( Fig 1 ) , and include both the known biochemistry of binding interactions and the physics and biomechanics of membrane bending and scission . In yeast , the cytoskeleton is needed to help induce membrane budding , after which energy-consuming proteins such as dynamin scission off the vesicle from the plasma membrane for transport into the cell [76 , 82] . However , such a modeling approach does not exist , due to the computational limitations of simulating such large complexes and membrane remodeling , and the lack of biochemical data . Based on the model we did construct , however , there are some more specific limitations . The first is that while rule-based modeling is a convenient way to model complex formation , some theoretical aggregates may be impossible due to steric hindrance . Our model predicted that a vesicle of 100 triskelia could contain ~1900 additional proteins . Assuming each vesicle is a sphere with 100nm diameter , the allowable surface area per adaptor/scaffold protein would only be ~17nm2 , which is too small to accommodate the excluded volume of the large , disordered regions of proteins such as ENT1 and 2[83] . Second , we did not include cooperatively in our model . Molecules localized in the same aggregate do not interact at a faster rate in conventional rule-based modeling . Clathrin triskelia weakly polymerize , as noted above , but the aggregation effect of the adaptor proteins–especially the SYP1/EDE1 complex–localizes triskelia close together , allowing them to bind strongly . In future work we will consider effects of cooperativity on assembly , as well as construct more detailed spatial and structural models of the vesicle forming process .
A stoichiometrically balanced network has the copy numbers of each interface matched to the copy numbers of all pairwise complexes it participates in ( Fig 2 ) . Balanced copy numbers are obtained by assigning a number of desired complexes to each edge in the interface binding network . The balanced copy numbers of each interface can then be calculated from the equation: Ax=C ( 2 ) Where “A” is a binary matrix with Nint rows ( one for each interface ) and Medge columns ( one for each pairwise interaction ) . Ai , j = 1 if the interface i is used in the interaction j , or 2 if a self-interaction , and 0 otherwise . “x” is the vector of desired pairwise complexes ( Medge x 1 ) , and “C” is the number of interface copy numbers ( Nint x 1 ) . In Fig 10 we illustrate this procedure for a small toy network . If desired pairwise complexes , x , is specified , interface copy numbers , C , can directly be solved for using Eq 1 , but if interface copy numbers , C , are specified , x will not , in general , have an exact or nontrivial solution unless C is balanced . This is because all entries of x must be >0 or some other minimum value , as negative copies cannot exist . This produces a hard constraint on x . Given a vector C , an optimal solution to x must be solved for using quadratic programming rather than linear least-squares . Our goal is to select for an optimal x given an input set of copy numbers “C0” . This is a soft constraint on the optimal x , because the input C0 may not be balanced . Once an optimal x is found , forward solving Eq 1 will in general not perfectly recover C0 . C0 can constrain all interfaces or a subset of them . To constrain a protein is to constrain all interfaces on it . We introduce a third constraint on the optimal x: the copy numbers of interfaces on the same protein should be equal . This often makes nontrivial solutions impossible ( Fig 10 ) , so it is also a soft constraint . Combining all of these constraints , the optimal desired number of complexes “x” can be found by minimizing the equation: minx[α ( Ax‑C0 ) TZ ( Ax‑C0 ) + ( Ax ) TH ( Ax ) ] , x≥0 ( 3 ) Where each variable is defined as follows: A: Nint x Medge matrix defining which interfaces are used in which interaction , i . e . pairwise complex . x: Medge x 1 vector of desired pairwise complex copy numbers C0: Nint x 1 vector of constrained copy numbers . Z: Nint x Nint diagonal matrix that selects which interfaces are constrained . Entries = 1 if the interface is constrained and = 0 otherwise . If all interfaces are constrained , Z equals the identity matrix . H: Nint x Nint permutated block diagonal matrix with positive and negative entries such that H*C = 0 if interfaces on the same protein have equal copy numbers . Each block corresponds to a protein ( Fig 10 ) . α: 1x1 scaling parameter which determines the relative weight of the C0 soft constraint vs the equal interfaces soft constraint . For any vector x , Eq 2 produces a positive scalar value . The equation was minimized using the OOQP ( object-oriented quadratic programming ) 0 . 99 . 26 package for C++[84] . Quadratic programming is necessary due to the constraint of x≥0 . Eq 2 can be converted into a quadratic equation of the form 12xTQx+dTx+r ( 4 ) Using Q = 2αATZA + 2ATHA dT = -2αC0TZTA r = αC0TZC0 “r” can be ignored by the solver when minimizing the equation since it is a constant term . Once xmin is found via Eq 3 , the optimized interface copy numbers can obtained by forward solving A*xmin = Cbalanced , int . Interfaces on the same protein will not necessarily have equal copy numbers due to the competing constraints of Eq 2 ( Fig 2C ) . We can assign a single copy number to each protein by averaging over all interface copy numbers on that protein to give Cbalanced , a vector of protein copy numbers . These values were used when calculating which proteins were over or underexpressed in the networks . Distance from C0 to Cbalanced was used as a metric to determine relative balance ( see below ) . For the yeast CME network , C0 was used to constrain all 56 proteins ( Z = Identity matrix ) because copy numbers from Kulak et al . were available[2] . For the ErbB signaling network , only 115 out of 127 proteins with available expression level data were constrained . 100 of these proteins were constrained with HeLa copy number estimations from Kulak et al . [2] , while estimated copy numbers for 15 additional proteins were added from four additional studies[19 , 51–53] , leaving 12 proteins with unknown expression data . See S2 Table for all values . Using the optimized copy numbers , Cbalanced , we can then ask , how close are the original , biologically observed copy numbers to these optimally balanced values ? If the original copy numbers are already perfectly balanced , then they will match the optimal copy numbers . If they are imperfect , then the two distributions will differ . We use two metrics to quantify the distance between the observed and optimized concentrations: chi-square distance ( CSD ) ∑i ( Xi−Yi ) 2 ( Xi+Yi ) ( 5 ) and Jensen-Shannon Distance ( JSD ) after converting both vectors ( X and Y ) to distributions ( x and y ) 12 ( DKL ( x‖z ) +DKL ( y‖z ) ) ( 6 ) Where z = ( x+y ) /2 and DKL is the Kullback-Leibler divergence DKL ( x‖y ) =∑ixilogxiyi ( 7 ) For cases where Z≠I ( i . e . not all interfaces were constrained ) only distance between constrained interfaces was measured . Binding for the five 3- or 4-node network motifs; triangle , chain , square , 4-node hub , and flag; was simulated using the Gillespie algorithm[47] . Besides the specific binary interactions , nonspecific interactions were allowed at a strength determined by an “energy gap” between binding energies , though in practice we defined the ratio nonspecific KD to specific KD by factors of 10 . This corresponded to a linear difference in free energies via the equations: KD , specific=c0e−ΔE1KBT KD , nonspecific=c0e−ΔE2KBT KD , specificKD , nonspecific=e− ( ΔE1−ΔE2 ) KBT The networks were simulated under various initial concentrations . The steady-state ratio of Eq 1 was recorded , where Nnonspecific is the number of nonspecific binary complexes , Nspecific is the number of specific binary complexes , and Nfree is the number of free proteins . Ratios were averaged across 5 , 000 runs . To generate surface plots , two proteins were chosen to be variable while the remaining proteins were given fixed copy numbers . Because the flag motif produced asymmetric plots , two different choices of variable proteins were used . ( S3 Fig ) Surface plots were generated using Matlab . We calculated sensitivity by determining the principal component of the surface plot data ( i . e . the vector of greatest variance ) and measuring the percent change in ratio from the optimum along this vector . For better comparison , we normalized distance along the surface plots via dividing the abundance of the variable proteins by the abundance of the fixed proteins . Motifs with purely noncompetitive interactions were not considered , because the interface network would then consist entirely of pairs , such as the IIN for Fig 1B . The balance is simple for pairs: all interfaces have the same copy numbers . We limited our analysis of Results part 2 , “Imbalance increases misinteractions dependent on the network topology and binding affinities of proteins” , to small competitive motifs where we could enumerate all possible complexes and study effects of concentration variation systematically . For the large network analysis we used the 500 networks from Johnson et al , J Phys Chem B 2013[27] . 25 sets of 10 networks each were randomly generated using two parameters: number of nodes ( 90 , 110 , 125 , 150 , 200 ) , keeping the number of edges fixed at 150; and the preferential attachment exponent “γ” from Goh , 2001[85] . γ = 0 corresponds to a binomial , Erdos-Renyi network , whereas γ = 1 corresponds to a power-law or “scale-free” network . Values of 0 , 0 . 2 , 0 . 4 , 0 . 6 , and 0 . 8 were used . Finally , a local topology optimization algorithm that decreased the frequency of chain and triangle motifs and increased hub motifs was applied to each network , for 500 networks in total . All networks assume competitive ( binary ) binding . Rather than assign an arbitrary specific and nonspecific KD for the networks , we used the relative binding energies determined for each network in the source paper . This was determined by a physics-based Monte Carlo optimization scheme of amino acid residues , as described in Johnson , 2011[23] . The minimum energy gap between specific and nonspecific interactions could be measured as a relative metric of the network’s propensity for misinteractions . Because the binding strengths were relative , we could alter the average binding strength to determine the effects on misinteractions . This was varied between 7 values of 1 nM to 1 mM , using factors of 10 . Finally , to obtain results more comparable to the simple networks , we also ran simulations where each specific interaction had KD = 100 nM and each nonspecific interaction had KD = 100 μM . Networks were simulated to steady state using the Gillespie algorithm[47] under five differing sets of copy numbers ( CNs ) for free proteins: equal CNs for each protein , random CNs sampled from a yeast protein concentration distribution ( performed 20 times ) and three forms of balanced CNs using the network architecture . Any set of CNs without leftovers–i . e . having exactly enough proteins to create a certain number of specific complexes–is considered “balanced” , and thus there are infinite solutions . The first balanced set assumed an equal number of each type of specific complex , which results in protein CNs proportional to the protein’s number of partners . The remaining balanced CNs were determined by finding “x” to minimize a simplified form of Eq 2: minx ( A*x‑C0 ) T ( A*x‑C0 ) ( 8 ) Here there is only one interface on each protein , and all the proteins are constrained , so there is no need for a Z matrix , the α scaling parameter , or the second term . C0 is either equal copy numbers or randomly sampled copy numbers . After xmin is found via quadratic programming ( see above ) , the balanced CNs are obtained by forward solving Cbalanced = A*xmin . To measure nonspecific complex formation , a modified ratio was used: Cost ( C0 ) =2Nnonspecific ( C0 ) 2Nspecific ( C0 ) +Nfree ( C0 ) ( 9 ) to compare total individual proteins in each bound or unbound state , rather than number of unbound or bound states . To measure sensitivity , the ratio under unbalanced CNs ( C0 ) divided by the ratio under balanced CNs ( Cbalanced ) was calculated . A higher ratio indicates higher sensitivity to CN balancing . The kinetic model was simulated using the stochastic simulation method ( the Gillespie algorithm ) . Binding interactions were encoded via the rule-based language BioNetGen and simulated via the Network Free Simulation ( NFSim ) software [48] . Trimer cooperativity was modeled by increasing the rate of the third reaction if three members of a correct trimer were held together by two reactions . For example , if A is bound to B is bound to C , and a binding between A and C is possible , that reaction rate was set to be arbitrarily high . Reaction rates were arbitrary , but interactions with the core subunit ARC19 were set to be ~10 fold stronger than interactions between periphery subunits , as this increased yield . Yield was measured via the equation Yield=NdesiredNdesired+Nundesired ( 10 ) Where Ndesired is the number of proteins in complete complexes ( equal to seven times the number of complex complexes ) and Nundesired is the number of proteins in incomplete or misbound complexes . Completely free proteins were ignored . A subnetwork of nine proteins–clathrin heavy chain ( CHC1 ) , clathrin light chain ( CLC1 ) , SLA2 , ENT1/2 , EDE1 , SYP1 , and YAP1801/2 –was defined based on known binding interactions ( Table 1 ) . Because the existence of multiple interfaces , allowing noncompetitive binding , results in a large number of possible species we simulated our model using the Network Free Simulator ( NFSim ) [48] . Binding dissociation constants were obtained from the literature , including for protein-lipid binding . For simplicity , the heavy chains were already assumed to be in trimer form , and ENT1/2 was combined into a single protein as the binding partners were the same . Binding constants were pulled from the literature . ( Table 1 ) The cell membrane and the cell cytoplasm function as different compartments with different volumes , but NFSim is not integrated with BioNetGen’s compartment language . We bypassed this problem by doubling the number of rules: besides the main rule for each reaction , an additional rule stated that if both proteins are on the cell membrane then the kon rate should be increased according to the membrane volume . Cell membrane ‘volume’ was determined by multiplying the membrane surface area by a factor 2σ = 2 nm to capture the change in binding affinities between 3D and 2D ( see S1 Text ) . Since our primary goal was to measure clathrin recruitment to the membrane , any complex on the membrane with at least 100 triskelia ( a complex of three CHC1 and three CLC1 ) was considered a “vesicle” and deleted at a high rate kdump . Proteins in the vesicle were then added back to the cytoplasmic pool at a rate krecyc , which was set to be equal to kdump to indicate fast recycling . However , we clarify that even fast recycling is not instantaneous , and that proteins are added back one at a time rather than all at once . Fast vesicle formation thus could still drain the pool of adaptor proteins . Misinteraction strengths were determined by calculating the geometric mean of the dissociation constants of each interface , as this provided a KD based on the arithmetic mean of the binding energies . KD , mean=KD , 1KD , 2…KD , nn=e−ΔE1KBT∙e−ΔE1KBT∙…e−ΔE1KBTn=e− ( ΔE1+ΔE2+…ΔEn ) KBTn=e− ( ΔE1+ΔE2+…ΔEn ) nKBT The KD of a misinteraction between two interfaces was set to be: fKD , mean , 1KD , mean , 2 ( 11 ) where f = 10 , 000 ( weak misinteractions , corresponding to an energy gap of ~9 . 21 ) or 1 , 000 ( stronger misinteractions , energy gap of ~6 . 91 ) Network maps were generated using Cytoscape[86] and RuleBender[87] . Plots were generated in MATLAB . C++ code for the network balancing algorithm SBOPN is available at https://github . com/mjohn218/StoichiometricBalance , and may be applied to any interface-resolved network . The CME and ErbB networks are provided as example inputs . | Protein copy numbers are often found to be stoichiometrically balanced for subunits of multi-protein complexes . Imbalance is believed to be deleterious because it lowers complex yield ( the dosage balance hypothesis ) and increases the risk of misinteractions , but imbalance may also provide unexplored functional benefits . We show here that the benefits of stoichiometric balance can extend to larger networks of interacting proteins . We develop a method to quantify to what degree protein networks are balanced , and apply it to two networks . We find that the clathrin-mediated endocytosis system in yeast is statistically balanced , but not perfectly so , and explore the consequences of imbalance in the form of misinteractions and endocytic function . We also show that biological networks are more robust to misinteractions than random networks when balanced , but are more sensitive to misinteractions under imbalance . This suggests evolutionary pressure for proteins to be balanced and that any conserved imbalance should occur for functional reasons . We explore one such reason in the form of bottlenecking the endocytosis process . Our method can be generalized to other networks and used to identify out-of-balance proteins . Our results provide insight into how network design , expression level regulation , and cell fitness are intertwined . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"chemical",
"characterization",
"vesicles",
"protein",
"interactions",
"protein",
"interaction",
"networks",
"cell",
"processes",
"membrane",
"proteins",
"network",
"analysis",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"cell",
"binding",
"assay",
"computer",
"and",
"information",
"sciences",
"network",
"motifs",
"proteins",
"proteomics",
"binding",
"analysis",
"cell",
"membranes",
"endocytosis",
"biochemistry",
"cell",
"biology",
"secretory",
"pathway",
"biology",
"and",
"life",
"sciences"
]
| 2018 | Stoichiometric balance of protein copy numbers is measurable and functionally significant in a protein-protein interaction network for yeast endocytosis |
Pseudomonas syringae is a phylogenetically diverse species of Gram-negative bacterial plant pathogens responsible for crop diseases around the world . The HrpL sigma factor drives expression of the major P . syringae virulence regulon . HrpL controls expression of the genes encoding the structural and functional components of the type III secretion system ( T3SS ) and the type three secreted effector proteins ( T3E ) that are collectively essential for virulence . HrpL also regulates expression of an under-explored suite of non-type III effector genes ( non-T3E ) , including toxin production systems and operons not previously associated with virulence . We implemented and refined genome-wide transcriptional analysis methods using cDNA-derived high-throughput sequencing ( RNA-seq ) data to characterize the HrpL regulon from six isolates of P . syringae spanning the diversity of the species . Our transcriptomes , mapped onto both complete and draft genomes , significantly extend earlier studies . We confirmed HrpL-regulation for a majority of previously defined T3E genes in these six strains . We identified two new T3E families from P . syringae pv . oryzae 1_6 , a strain within the relatively underexplored phylogenetic Multi-Locus Sequence Typing ( MLST ) group IV . The HrpL regulons varied among strains in gene number and content across both their T3E and non-T3E gene suites . Strains within MLST group II consistently express the lowest number of HrpL-regulated genes . We identified events leading to recruitment into , and loss from , the HrpL regulon . These included gene gain and loss , and loss of HrpL regulation caused by group-specific cis element mutations in otherwise conserved genes . Novel non-T3E HrpL-regulated genes include an operon that we show is required for full virulence of P . syringae pv . phaseolicola 1448A on French bean . We highlight the power of integrating genomic , transcriptomic , and phylogenetic information to drive concise functional experimentation and to derive better insight into the evolution of virulence across an evolutionarily diverse pathogen species .
Many Gram-negative bacteria attach to host cells and translocate effector proteins into them via type III secretion systems ( T3SS ) . Such systems are necessary for pathogenesis , are horizontally transferred across species , and are accompanied by dynamically evolving repertoires of type III effector ( T3Es ) genes [1] , [2] . The T3SS is essential for Pseudomonas syringae pathogens to thrive in plant tissues . P . syringae represents an excellent example of the plasticity of T3E repertoires [3] . Despite a collectively broad host range for the species , individual isolates of P . syringae typically display pathogenic potential on a limited set of plants and either elicit immune responses , or simply fail to thrive on other plant species . Strains can be isolated from diseased plants , as epiphytes from healthy plants [4] , and from various environmental sources [5] . The hrp/hrc group I T3SS is essential for P . syringae pathogens to cause disease on plants [1] , [6] . The genes that encode the hrp/hrc T3SS and accessory proteins are clustered in a conserved pathogenicity island in P . syringae [7] . The genes for the associated T3Es can be scattered across the genome , often in association with mobile elements indicative of horizontal transmission [8]–[10] . Each strain's T3E repertoire ranges from 15–30 genes sampled from at least 57 different families and these collectively modify host cell biology to suppress immune response and favor bacterial proliferation and dispersion . However , the action of individual T3E proteins can be recognized by plant host disease resistance proteins , and this triggers immune responses sufficient to limit pathogen growth [11] . The conflicting selective pressures to retain a collection of T3E sufficient to suppress host defenses without triggering effector-specific immune responses [11] drives diversity in the suites of T3Es in plant pathogenic P . syringae isolates [3] . Transition from saprophytic to epiphytic or pathogenic lifestyle requires significant transcriptional reprogramming . Expression of genes encoding the P . syringae T3SS structural components and the associated T3E suite is controlled by the ECF-type sigma factor HrpL [12]–[14] . The expression of hrpL is induced in bacteria that encounter the leaf environment [13] . Subsequently , HrpL binds to promoters carrying a “hrp-box” consensus sequence to up-regulate the expression of the corresponding gene ( s ) [12]–[15] . Previous studies in P . syringae identified proteins that are neither T3Es nor structural components of the T3SS ( hereafter , non-T3Es ) , but are HrpL-regulated [3] , [16]–[19] . Non-T3Es coordinately regulated with the T3SS and its substrates were also found in other T3SS-expressing plant pathogens such as Erwinia amylovora [20] , Ralstonia solanacearum [21] , Xanthomonas campestris pv . vesicatoria [22] , [23] and Pectobacterium carotovora [24] . Some HrpL-regulated non-T3E genes affect virulence on host plants in the well-studied strain P . syringae pv . tomato DC3000 ( PtoDC3000 ) ; these include the corR regulator of coronatine toxin production [18] , [25] . Notably , CorR expression is not HrpL-regulated in other strains , such as P . syringae pv . glycinea PG4180 [26] . Multi-Locus Sequence Typing ( MLST ) separates plant pathogenic P . syringae into at least 5 distinct phylogenetic groups [3] , [27] . The fifth group , represented initially by P . syrinage pv . maculicola ES4326 , was recently renamed P . cannabina pv . alisalensis ES4326 [28] . Many P . syringae genome sequences are now available , including three closed genomes from isolates representing major pathogen clades [29]–[31] , and ∼120 additional draft sequences . Newly sequenced genomes also trace P . syringae disease outbreaks across the globe and over time [3] , [32]–[39] attesting to the continued importance of the species . Recently , isolation and sequencing of saprophytic and epiphytic strains provided insight into a subgroup from group II that carries a non-canonical T3SS [40] . To date , transcriptome analyses using high throughput short read cDNA sequencing ( RNA-seq ) have been applied only to PtoDC3000 , providing a well-curated reference gene annotation , but not specifically informing studies of the HrpL regulon [41]–[44] . In this study , we defined the HrpL regulon of six distinct strains of P . syringae with complete or draft genomes using RNA-seq coupled with the GENE-counter software package [45]–[47] . We sought primarily to compare the diversity of non-T3E HrpL-regulated genes between strains and secondarily to determine if there were additional type III effectors not found in our DNA-based analyses [3] . We detect non-T3E genes regulated directly or indirectly by HrpL . Those directly regulated by HrpL are distributed throughout the P . syringae clades in a mosaic pattern . However , most are either absent or not HrpL-regulated in MLST group II . We demonstrate that a novel cluster of non-T3E genes is required for P . syringae pv . phaseolicola 1448A virulence . We also identified two novel T3E families from a previously understudied clade . Our study reveals the mechanisms for gene recruitment into , and loss from , the key virulence regulon in P . syringae , and provides a roadmap for future functional studies .
We defined the HrpL regulons of P . syringae pv . phaseolicola strain 1448A ( Pph1448A ) , P . syringae pv . lachrymans strain 107 ( Pla107 ) representing MLST group III; P . syringae pv . syringae strain B728a ( PsyB728a ) , P . syringae pv . japonica strain MAFF 301072 PT ( Pja ) representing MLST group II; P . syringae pv . tomato strain DC3000 ( PtoDC3000 ) representing MLST group I and P . syringae pv . oryzae strain 1_6 ( Por ) , belonging to the relatively poorly studied clade , MLST group IV [3] , [27] . The native hrpL gene from each isolate was cloned downstream of an arabinose-inducible promoter for controlled , high-level expression in the strain of origin . Isogenic strains carrying either the appropriate hrpL construct , or an empty vector ( EV ) as negative control , were grown with arabinose to induce the expression of the cloned hrpL gene in a minimal medium [19] . Expression of the native hrpL was repressed by addition of peptone to the media [48] . Figure S1 depicts our experimental pipeline and control validation . We generated Illumina cDNA libraries from two biological replicates of each strain . Because our goal was to compare transcript abundance more than to improve annotation of transcribed genes , we used a simple cDNA method to minimize the RNA processing steps where transcripts could be lost . Therefore , we did not enrich for 5′ ends or distinguish transcript orientation . Transcript abundance was compared between isogenic HrpL and EV samples using GENE-counter [45] . Similar to other RNA-seq analysis methods like EdgeR or DESeq [49] , [50] , GENE-counter determines differential expression . While EdgeR and DESeq use the standard negative binomial distribution , GENE-counter relies on the negative binomial p distribution which better accounts for the over-dispersion observed in mRNA-seq data [51]–[53] . We bootstrapped the GENE-counter output for each isolate ( Materials and Methods ) to control for noise introduced by sample normalization . Between 1 . 6 and 5 . 6 million unambiguous reads per sample ( mapping to only one location in the reference genome ) were used for our analyses ( Table 1 ) . The sequencing depth ranged from 9 . 5 to 16 . 2 times the genome size , with the exception of the PsyB728a samples , which we sequenced to higher coverage ( Table 1 ) . On average 93 . 5% of the total number of annotated coding genes in a genome were covered by at least one read in at least one sample ( Table 1 ) . Bootstrapped-GENE-counter analysis established a median read count for every sample , a median q-value and a B-value , for every gene covered by at least one read in one biological replicate ( Table S1 ) . Genes not covered by any unambiguous reads are not represented in our GENE-counter output . The B-value represents the percentage of bootstraps in which a particular gene was called differentially expressed . We further considered only genes with B-values≥50% . Like all “significance thresholds” the B-value cut-off is somewhat subjective . We selected a B-value of 50% to apply to all genomes because this threshold captured 95% of the known HrpL-regulated genes identified in our control genome , PtoDC3000 , with a median q-value greater than 0 . 05 . We identified between 59 to 192 genes differentially expressed across the strains ( Table S1 ) . For all strains , the large majority of the differentially expressed genes were up-regulated ( between 53 and 180 genes , Table 2 ) . These genes mainly encode T3SS components and known T3Es . Surprisingly , we identified few HrpL-down-regulated genes ( Table S2 ) : ranging from none in Pph1448A to 45 in Pla107 . Genes called down-regulated in our analysis had relatively low q-values , reflecting low differences in read coverage between HrpL and EV samples . Lan et al . 2006 and Ferreira et al . 2006 identified down-regulated genes in their microarray studies for PtoDC3000 . However , almost no overlap was found between the list of down-regulated genes from previous studies and ours , indicating that the down-regulated genes identified are most likely neither biologically , nor statistically robust , and thus unlikely to be biologically relevant . In contrast , there was stronger overlap between our HrpL-induced genes and those shared between these earlier studies ( see below ) . Down-regulated genes were therefore not further analyzed . Finally , we manually inspected and curated all genes with B-values greater than or equal to 50% to define the HrpL regulon for each strain ( Table 2 , Table S3; Table S4; Materials and Methods ) . To evaluate the reproducibility of our method , we compared the read coverage within and between biological samples for all PtoDC3000 genes covered by at least one read in our normalized GENE-counter data set . Biologically replicated samples had highly correlated results ( R2 = 0 . 93 between EV replicates 1 and 2; R2 = 0 . 96 between conditional expression replicates HrpL1 and HrpL2 , Figure 1A , lower panels ) . Comparing HrpL and EV replicates from two biological replicates , the majority of the data points correlate and cluster around the trend line ( Figure 1A , upper panels ) . The outlier data points in red represent genes defined as differentially up-regulated by GENE-counter and having a B-value≥50% . We plotted the log of the median q-value of each PtoDC3000 gene defined to be differentially up-regulated ( before manual curation ) and their corresponding B-values ranked from smallest to largest ( Figure 1B ) . As expected , genes with highly significant q-scores also have high B-values . Several genes not previously reported to be HrpL-regulated ( marked in red ) had more significant q-value scores ( 3 . 8E-02 ) than avrE ( marked in blue ) , a well-characterized conserved HrpL-regulated type III effector [54] . We analyzed the same PtoDC3000 RNA-seq data set using either the complete PtoDC3000 genome sequence [30] or a draft PtoDC3000 genome sequence [36] as references . The draft genome sequence covers 85% of genes at over 90% of their length [36] . Using either the complete or the draft genome as a reference resulted in similar sequencing depths ( Table 1 ) . Using the draft genome as a reference , GENE-counter identified 124 HrpL-upregulated genes out of the 133 found using the complete PtoDC3000 genome ( Table 2 ) . Most of the genes that were not identified as differentially expressed using the draft genome were missing from the draft genome ( data not shown ) . The high correlation between the Log ( median q-value ) of genes in the two data sets ( Figure 1C ) indicates that our method will effectively identify the majority of genes of the HrpL regulon from P . syringae isolates for which only a high quality draft genome is available . To further validate our pipeline to define HrpL–regulated genes , we compared our manually curated list of 110 PtoDC3000 HrpL-regulated genes ( Table 2 ) to HrpL-regulated genes identified by three previous studies: one promoter probe study using an arabinose-inducible hrpL gene and two custom microarray analyses which compared expression between wild type and hrpL deletion mutant strains [16] , [17] , [19] . These studies produced largely overlapping , but not identical , lists of putatively HrpL-regulated genes ( Table S5 ) . Our PtoDC3000 HrpL-regulated gene set included 57 out of the 66 genes previously identified as HrpL-regulated in at least two of the previous studies ( Table S5 ) , even though our induction and analysis methods differed from these studies . 96 of the 110 genes we identified were also found to be HrpL-regulated in at least one of the previous studies [16] , [17] , [19] or were downstream genes in HrpL-regulated operons ( Table 2 ) . Overall , we found 91% of the previously identified HrpL-regulated genes in PtoDC3000 . Our analysis also identified 14 novel HrpL-regulated genes ( Table 2 ) ; six out of eight tested were confirmed to be HrpL-regulated using qRT-PCR ( Table 3 , see below ) . Notably , four of the nine missing genes were not present in our laboratory strain , which has lost part of the PtoDC3000 plasmid A . One gene , shcA ( PSPTO_5353 ) was found differentially expressed in our analysis but had a B-value less than 50% . Further , GENE-counter discards RNA-seq reads that map non-uniquely to more than one location in the genome , and HrpL-regulated duplicated genes account for three missing PtoDC3000 genes: T3E genes hopAM1-1 ( PSPTO_1022 ) and hopQ1-2 ( PSPTO_4732 ) , and the non-T3E gene plcA2 ( PSPTO_B0005 ) ( Table S5 ) . Finally , hopK1 ( PSPTO_0044 ) , was covered by RNA-seq reads but the differences in expression in HrpL and EV treatments were not statistically significant ( Table S1 , S5 ) . Two previous studies focused on the identification of HrpL-regulated genes in Pph1448A [18] , [19] and identified 43 HrpL-regulated genes comparing expression between wild type and hrpL mutants . We identified 35 ( ∼80% ) . Four of the missing eight genes were covered by reads but not found significantly differentially expressed , hopAK1 ( PSPPH_1424 ) , a gene encoding a MarR transcriptional regulator ( PSPPH_1519 ) , avrRps4 ( PSPPH_A0087 ) , and hopAS1 ( PSPPH_4736 ) . Those four genes had a median read coverage ranging from 100 to 1000 , indicating that the absence of differential expression in our analysis is not due to weak or undetectable levels of expression . One , PSPPH_2294 is a pseudogene . PSPPH_1525 encoding a putative effector related to Ralstonia Hpx30 [55] , PSPPH_A0009 and A00075 encoding truncated hopW1 are duplicated and had very low to no read coverage ( Table S5 ) . Our GENE-counter analysis pipeline results are consistent with previous transcriptional studies , reinforcing the validity of our methods . Additionally , we identified robustly HrpL-induced genes that were not previously identified . We identified between six and 32 genes previously not known to be HrpL-regulated in each strain with corresponding q-values ranging from E-02 to E-54 ( Table 2 , Table S3 ) . Some of these are shared across strains . We could not identify a consensus upstream hrp-box in the promoters of several , and suggest that these could be indirectly activated by HrpL . We performed qRT-PCR using samples derived from strains expressing HrpL in the pBAD system and confirmed 19 of 23 tested ( Figure S2 ) . Additionally , we confirmed HrpL-dependent expression of 19 genes out of 20 tested , by comparing wild type expression with expression in a hrpL deletion mutant in hrpL-inducing minimal medium ( Table 3 , Figure S3 ) . We observed a high correlation between RNA-seq data and either qRT-PCR profiling method , especially for genes with a q value>E-03 ( Table 3 , Figures S2 , S3 ) . In sum , we identified the majority of previously identified HrpL-regulated genes in two well-studied strains and we confirmed wild type HrpL regulation for nearly all of the newly identified members of this key virulence regulon . Most of the known T3E and candidate T3E genes in our tested strains and those previously defined by similarity and/or functional criteria were included in the HrpL regulons we defined in our RNA-seq analyses ( Figure S4 ) . Most of strains used in this study had previously been screened for novel type III effector genes by functional translocation assays with the exception of Por and Pja [3] , [19] . Therefore , we searched the Por and Pja HrpL regulons for potential novel effector genes based on the criteria of having an identifiable upstream hrp-box sequence and no homology to previously identified T3E families . We chose six Por genes ( Porcurated_02784 , 04644 , 04640 , 03530 , 02145 , and 04371 ) to investigate as potentially encoding novel T3Es . Pja also carries a gene homologous to Porcurated_04644; but only the Por allele was tested . All six putative T3E were tested for their ability to be translocated via a native T3SS using an established assay [56] ( Materials and Methods ) from PtoDC3000D28E , an “effector-less” PtoDC3000 strain [57] . Only PtoDC3000D28E carrying Porcurated_02784-Δ79avrRpt2 or Porcurated_04640-Δ79avrRpt2 triggered a Hypersensitive Response ( HR ) in Col-0 ( Figure 2A ) . We verified that HA-tagged versions of all six T3E candidates were expressed in PtoDC3000D28E indicating that lack of HR in our translocation assay was unlikely due to a lack of protein accumulation ( Figure 2B ) . No HR was observed in the rps2 mutant , indicating that the response was avrRpt2-specific and not the result of toxicity . These two new P . syringae effectors will henceforth be referred to as HopBH1Por and HopBI1Por according to proposed T3E naming guidelines [58] . None of the 19 P . syringae strains for which we previously performed comparative genomic analysis encode either hopBH1 or hopBI1 [3] . However , each can be found in P . syringae strains isolated from various sources ranging from non-symptomatic plants to snow [33] , [35] , [40] , [59]–[62] ( Figure S5 ) . Amino acid sequence alignments suggest that HopBH1 is a bi-modular effector exhibiting sequence conservation within its C-terminal domain and sequence diversity toward its N-terminal half ( Figure S6 ) . In the non-pathogenic strain Psy642 , the putative HopBH1 protein appears to have been disrupted by a frameshift mutation , leading to two putative open reading frames designated as ORF29-30 [40] . Phylogenetic analysis of strains carrying either hopBH1 and/or hopBI1 indicates that both effector genes occur with a mosaic distribution across the P . syringae , but are absent from the phylogenetic group III [3] , [27] ( Figure S5 ) . Neither HopBH1 nor HopBI1 contain known protein folds , nor do they display sequence or structural homology to proteins of known function . The composition of the HrpL regulon across strains was surveyed by functional classification based on protein annotation and sequence homology determined by BLASTP ( Table S6 ) . As summarized in Figure 3 and Table S7 , PtoDC3000 and Por possess the largest and most diverse HrpL regulons among the sampled strains , while the Group II strains Pja and PsyB728a have the smallest . We are confident that the less complex HrpL regulons are not a sampling artifact , because the data collected from Pja has a transcriptome depth similar to the other strains , and the PsyB728a HrpL regulon was sampled at relatively high depth compared to our other transcriptomes . We conclude that HrpL regulons vary in size and composition across the P . syringae phylogeny . We observed variable HrpL-dependent expression for several highly conserved non-T3E genes present in all six strains ( Table S6 ) . We identified polymorphisms in the hrp-box sequences from two of these ( Figure 4A ) . In the first case , new HrpL-regulated genes we identified , PSPTO_2130 , PSPPH_1906 and Lac107_00061530 , are orthologs that encode a DNA-binding response regulator . HrpL-dependent induction was confirmed by qRT-PCR ( Table 3 , Figure 4B ) . Orthologous genes are also present in Pja , PsyB728a , Por ( Pjap_00016990 , Psyr_1940 , and Porcurated_00527 , respectively ) but were not identified as differentially expressed ( Table S1 ) . PSPTO_2130 and all of its orthologs have conserved hrp-box motifs . However , the promoters of the orthologs from Por and all other group II strains contain single nucleotide polymorphisms ( in red , Figure 4A ) in the consensus hrp-box sequence . Our RNA-seq data suggested that expression of these polymorphic alleles was not HrpL-dependent , a finding confirmed by qRT-PCR performed with both of our HrpL-regulation experimental tests ( Figure 4B , Figure S7A ) . PSPTO_2130 and its orthologs are part of a putative operon composed of four genes , PSPTO_2128-2131 ( Figure S8A ) . Unusually , the hrp-box sequences were located within the first ORF of the putative operons of PSPTO_2130 and its orthologs . We monitored HrpL-dependent expression using qRT-PCR of all genes from PSPTO_2131 to 2128 from three strains ( Figure S8B , C , D ) . In none of these strains was the first gene of the operon , containing the putative hrp-box , differentially expressed . By contrast , HrpL-dependent expression was observed for genes downstream of the predicted hrp-box , including coding sequences , PSPTO_2130 and PSPPH_1906 , in all but the group II reference strain PsyB728a ( Figure S8 ) . Deletion mutants in PtoDC300 and Pph1448A of PSPTO_2130 and PSPPH_1906 did not display any growth defect on Arabidopsis accession Col-0 or French bean cultivar Tendergreen ( susceptible to PtoDC3000 and Pph1448A , respectively ) ( data not shown ) . Thus , the role of PSPTO_2130 and its orthologs in virulence remains unclear . In the second case , PSPTO_2105 and its orthologs , which encode a putative ApbE-family protein , are highly conserved across P . syringae and are HrpL-regulated in Pph1448A , Pla107 , PtoDC3000 and Por but not in the group II strains PsyB728a or Pja ( Table S5 , S6 ) . qRT-PCR ( Figure 4C , Figure S7B ) support our RNA-seq data . PSPTO_2105 is required for full virulence of PtoDC3000 on Arabidopsis [18] . We also observed significantly reduced virulence when we tested two independent deletion mutants of the Pph1448A ortholog PSPPH_1855 for growth on the native host , French beans ( Figure S9 ) . Every group II strain analyzed has variations in the otherwise well conserved hrp-box sequence in at least two positions ( Figure 4A ) . Collectively , these data demonstrate that promoter erosion within the hrp-box is a mechanism to remove genes from the virulence regulon . Both PsyB728a and Pja appear to have relatively small HrpL regulons; both belong to the MLST group II . To address whether this was a general feature of group II strains , and to address the distribution of the genes that we identified experimentally across the phylogeny , we extended our investigation of non-T3E HrpL regulon diversity to BLAST homology searches of 44 sequenced Pseudomonas spp . strains [3] , [35] , [63] . Our non-T3E gene search set included genes likely to be directly HrpL-regulated , derived from either previous studies [19] , [64] or this study . From our study , these included genes we experimentally confirmed for HrpL-dependent expression , genes that encoded proteins found not to be translocated , or genes unlikely to encode a translocated product by annotation . We removed T3SS genes and known T3Es ( Figure 5 ) . Most of the directly HrpL-regulated non-T3E genes we identified are absent from group II strains , but distributed across strains from groups I and III . Some are present in the previously described group IV and V , as well as the novel MLST groups VII , IX , X ( Berge et al . , personal communication , Figure S5 ) for which we had limited sampling . Further , the promoters of group II homologs of Porcurated_02977 , 01635 are divergent , and lack canonical hrp-boxes ( data not shown ) . Thus , not only do group II strains possess lower numbers of known T3E genes on average than the other phylogroups , group II strains also possess fewer non-T3E genes in their HrpL regulon suggesting a potential shift in virulence mechanisms of this clade [3] . Both Pph1448A and Pla107 contain avrD , a gene required for synthesis of syringolides , small molecules sufficient for HR on soybean cultivars expressing the Rpg4 disease resistance gene [65]–[67] . avrD is a non-T3E gene , as defined above ( Figure 5 ) , and its expression in E . coli is sufficient for production of syringolides [65] . RNA-seq analysis identified a series of orthologous , HrpL-regulated genes directly downstream of avrD in both Pph1448A and Pla107 ( Table S3 , S6 ) . In Pph1448A , those genes are arranged in two clusters composed of PSPPH_A0112-A0110 and PSPPH_A0109-A0106 , which are flanked by transposable elements ( Figure 6A ) . While most of these genes seem to encode hypothetical proteins , PSPPH_A0112 , A0109 , A0108 , A0107 encode putative enzymes: a phosphoglycerate mutase , a sulfotransferase , an amino transferase , and an oxidoreductase respectively . We confirmed the HrpL-dependent expression of PSPPH_A0112 , A0110 , A0109 , and A0107 ( Table 3 , Figure S2 , and S3 ) . This operon is typically found as a presence/absence polymorphism; when present , it is almost always downstream from avrD ( Figure 6B ) . PSPPH_A0111 corresponds to a 99 bp sequence present in Pla107 , P . syringae pv . mori ( Pmo ) , P . syringae pv . glycinea R4 ( PgyR4 ) , P . syringae pv . tomato T1 ( PtoT1 ) , and P . syringae pv . actinidiae ( Pan ) but not annotated as an ORF , thus it is not represented in the graphical representation of the conserved neighborhood region ( Figure 6B ) . In P . syringae CC1629 , this putative operon appears to have been disrupted by insertion of a transposable element . In P . syringae pv aesculi 0893_23 ( Pae ) this locus is not entirely sequenced . To determine whether avrD is part of an operon with PSPPH_A0112-A0106 , we used RT-PCR to confirm that the intragenic regions between avrD/PSPPH_A0112 and between PSPPH_A0110/PSPPH_A0109 were transcribed in wild type Pph1448A but either very weakly or not at all in the ΔhrpL mutant ( Figure 6C ) . We generated two independent deletion mutants for avrD and PSPPH_A0107 ( ΔavrD #1 and 2 , ΔPSPPH_A0107 # 1 and 2 , respectively ) and tested their growth on French bean cv . Tendergreen ( Figure 6D ) . All mutants displayed reduced growth compared to wild type Pph1448A ( Pph ) , indicating that both avrD and PSPPH_A0107 are required for full virulence on cv . Tendergreen . We confirmed that the HrpL-dependent expression of several downstream genes was not disrupted by mutations ( Figure S10 ) . However , PSPPH_A0112 , A0107 and A0106 were consistently slightly up-regulated in avrD mutants compared to the wild type . The intact remaining hrp-box is closer to PSPPH_A0112-A0106 in the avrD mutants , which could account for increased transcript levels . Additionally , these data could explain why the avrD mutants displayed a reduced growth defect compared to the ΔPSPPH_A0107 mutants ( Figure 6D ) . We speculate that these non-T3E genes are involved in the synthesis of a secondary metabolite ( s ) required for virulence of Pph1448A .
We identified HopBH1Por and HopBI1Por , defining two novel effector families . Both have a mosaic phylogenetic distribution across P . syringae [35] , [40] , [63] ( and an unpublished strain , TLP2 , JGI taxon ID: 2507262033 ) . Both are present in CC1513 and CC1629 , two other strains belonging to the MLST group IV . They appear to be absent from sequenced MLST group III strains . HopBH1 has a bi-modular structure . The ∼170 amino-acid N-terminus is divergent compared to the relatively well conserved ∼250 amino acid C-terminal domain across HopBH1 alleles ( Figure S6 ) . The HopBH1 C-terminal domain is 50% identical to a protein from P . fluorescens SS101 which lacks a putative hrp-box or a T3SS secretion competent N-terminal sequence [68] , suggesting that it may have been recruited as an effector by N-terminal assortment [69] . Several putative proteins present in Pantoea , Serratia , Burkholderia species , as well as Myxobacteria , display ∼50% identity with the HopBH1 C-terminal domain . Remarkably , about 150 amino acids of the HopBH1 C-terminal domain also shares 40% identity with part of the ∼1000 amino acid long P . savastanoi pv . savastanoi NCPPB3335 HrpK . Notably , this hrpK gene ( PSA3335_2516 ) is from a rhizobia-like type III secretion and is different from the hrpK ( Pto ) ( PSA3335_1389 ) of canonical T3SS conserved in plant pathogenic P . syringae [70] . HopBI appears to be confined to Pseudomonas . Neither HopBH1 , nor HopBI1 display similarity to known-effectors . Their virulence functions remain to be determined . Although analysis of type III virulence systems focuses mainly on the characterization and function of T3SS and T3E proteins , several non-T3E genes are co-regulated with the T3SS . They encode hypothetical proteins , transporters , or enzymes likely involved in secondary metabolism ( Figure 5 ) . In contrast to T3E genes , for which functional redundancy is predominant and generation of multiple effector mutants is often required to affect virulence [54] , [57] , [71] , [72] , several single knockout mutants of non-T3E HrpL-regulated genes in PtoDC3000 and Pph1448A displayed reduced virulence on Arabidopsis and beans [18] , [73] . In general , little is known about the non-T3E genes in HrpL regulons , but homology provides reasonable scenarios for several that we identified , and we functionally validated others ( below ) . Among our collection of diverse HrpL-regulated , non-T3E genes , none are present in the HrpL regulon of all six strains tested , and nearly all are distributed in a mosaic pattern among the genomes of available strains ( Figure 5 ) . PSPTO_0370 and orthologs encode a MATE efflux transporter present in an operon with iaaL which is involved in auxin conjugation to IAA-Lys [74] . Porcurated_02977 encodes a putative indole-3-glycerol phosphate synthase . Both potentially alter auxin signaling and could interfere with the balance between immune response and growth and development [75] . Several other putative transporters were identified as HrpL-regulated . PSPTO_2691 encodes a putative membrane protein TerC; PSPTO_0871 a putative macrolide efflux protein; Porcurated_01635 a putative threonine efflux protein; and PSPTO_0838 a putative major facilitator family transporter . Co-regulation of putative transporters with the T3SS suggests that promotion of nutrient acquisition , export of secondary metabolites , or detoxification of plant-encoded antimicrobials are important features of the virulence regulon . PSPTO_0834 , encoding a putative alcohol dehydrogenase , is the first gene of a putative operon comprising five genes ( up to PSPTO_0838 ) . This operon includes genes of unknown function , genes encoding a putative bifunctional deaminase-reductase enzyme and a transporter . The function of this operon remains unknown but at least PSPTO_0834 is required for full virulence of PtoDC3000 on Arabidopsis [18] . The PSPTO_0873-0875 putative operon is widely distributed across Pseudomonas and Erwinia species and also present in Pantoea stewartii pv . stewartii DC283 . In Erwinia and P . stewartii , this operon is physically linked to the T3SS and is HrpL-regulated . PSPTO_0873 is a putative amidinotransferase that makes ornithine and homo-arginine from arginine and lysine . Ornithine or homo-arginine may be then incorporated into a tri- or di-peptide natural product generated by the rest of this operon . Most interestingly , hsvC , hsvB , hsvA from Erwinia amylovora , corresponding to PSPTO_0873-0875 , are required for full virulence on apple shoots [76] . PSPTO_2105 and orthologs encode a protein similar to ApbE from Salmonella typhimurium involved in thiamine synthesis . ApbE was identified through the analysis of several mutants defective in thiamine biosynthesis , and was implicated in iron-sulfur cluster biosynthesis/repair , as well as FAD binding [77]–[79] suggesting a role during oxidative stress [78] . PSPTO_2105 is required for full virulence of PtoDC3000 on Arabidopsis [18] . We extend this finding by showing that the PSPPH_1855 ortholog of PSPTO_2105 is required for full virulence of Pph1448A on French bean ( Figure S9 ) . PSPTO_2130 and orthologs encode LuxR family DNA-binding response regulators that may be involved in regulation of regulons downstream of HrpL . Our deletion mutants of this gene in PtoDC3000 and Pph1448A , or of the entire PtoDC3000 operon , did not alter growth on Arabidopsis or French bean cv . Tendergreen , respectively ( data not shown ) , undermining the probability of a necessary function during plant colonization in our experimental conditions . However this operon is conserved across Pseudomonas , and PFLU_2937 , the ortholog of PSPTO_2129 from P . fluorescence SBW25 , was identified as a plant-induced gene [80] . It therefore remains plausible that this operon is involved in plant association . Porcurated_04644 appears to encode a putative RNA N-methyltransferase , while the hypothetical protein Porcurated_03530 has homology to FliB which , in Salmonella , is responsible for methylation of flagellin [81] . We speculate that both may be involved in modification of conserved molecules known to induce host defense responses [82]–[84] . avrD is widely distributed across bacteria and is involved in the synthesis of syringolides [85] . Syringolides are elicitors of cell death in soybean expressing the Rpg4 disease resistance gene [86] , [87] . The putative function of avrD is discussed below . One of our most striking comparative observations is the relatively small size and diversity of the HrpL regulons of the phylogenetic group II strains PsyB728a and Pja . We observed that most of the non-T3E genes known to be HrpL-regulated in other strains are not present , or lack HrpL-regulation in group II strains , underpinning the conclusion that the limited regulon observed for PsyB728a and Pja can most likely be generalized to all group II strains ( Figure 5 ) . They also contain fewer T3Es than the other clades [3] . The group II strains carry genes for phytotoxins not shared by other P . syringae groups . Expression of these phytotoxins is not regulated by HrpL , and could compensate for missing T3E functions , making a smaller T3E repertoire sufficient to suppress plant defenses [3] . Turnover within the HrpL regulon is known to be influenced by gene gain and loss , mediated by association of genes within the regulon with mobile elements and horizontal gene transfer ( data not shown , Figure 6 A , B ) . However , we also observed that all the group II strains analyzed here have polymorphisms in the hrp-box sequence that correlated with the loss of HrpL-dependent regulation of PSPTO_2105 and orthologs ( likely encoding AbpE ) . Several different polymorphisms within the hrp-box were observed , suggesting independent mutational events ( Figure 4 ) . Additionally , the group II strain orthologs of PSPTO_2130 ( LuxR family ) , carry nucleotide polymorphisms in the consensus hrp-box , and are not HrpL-regulated ( Figure 4 ) . Orthologous genes from Por also display nucleotide variation in this hrp-box , also leading to loss of HrpL-regulation . The substitution patterns of these alterations suggest multiple , independent losses of HrpL-regulation . PSPTO_2130 and its orthologs are part of an operon where the consensus hrp-box is embedded within the first ORF in this operon ( Figure S8 ) and is thus likely to be constrained by the genetic code . Interestingly , PSPTO_2130 and its orthologs have variation in the second half of the hrp box where CCAC is replaced by TCAC . This hrp-box motif , while uncommon , is also found in PSPTO_0370 , PORcurated_01251 ( hopAO1Por ) , and Pjap_00002060 ( hopC1Pja ) , each of which we defined as HrpL-regulated . The promoter erosion we observe could be driven by negative host selection pressure , or weak selection for maintenance of HrpL regulon membership combined with subsequent drift . Similarly , reversion of at least the SNPs could quickly recruit genes back into the HrpL regulon . Because the ORFs have not accumulated stop mutations , these promoter mutations are either relatively recent or there is active maintenance of the ORF sequence , perhaps for expression under different conditions . Horizontal transfer or other types of recombination could explain how 5′ regions diverge and how these regions and associated genes are recruited in to the HrpL regulon . Porcurated_02977 , 01635 , and 04371 , encode an indole-3-glycerol phosphate synthase , a putative threonine efflux transporter and a hypothetical protein , respectively , that are HrpL-regulated . Similar genes are present in Pja and PsyB728a but are not HrpL-regulated ( Figure 5 ) . Putative hrp-boxes can be identified in all three Por genes , but not for the corresponding genes in Pja and PsyB728a . These genes are not syntenic ( data not shown ) . They display high similarity in their coding sequence ( data not shown ) ; however their corresponding 5′ upstream regions are highly divergent . This could be the result of horizontal transfer , though there is no obvious footprint of mobile element DNA , or independent recombination events . Lastly , loss of transcription termination regulation could lead to read-through transcription , and thus provide a mechanism for recruitment of non-T3E genes into the HrpL regulon . This mechanism was first highlighted by the recruitment into the HrpL regulon of the corR gene which was recombined downstream of the hrp-box associated hopAQ1 gene , in PtoDC3000 [25] . We observed that several genes found differentially expressed in our analysis were located downstream of HrpL-regulated T3E genes ( Table S3 ) and could potentially be recruited into the HrpL regulon via loss of transcription termination regulation and subsequent transcriptional read-through . We identified a cluster of HrpL-regulated genes , PSPPH_A0106-A0112 , downstream from avrD that were recruited into a novel HrpL-regulated operon transcribed from the avrD promoter . These genes are flanked in Pla107 and Pph1448A , by transposable elements , suggesting that they could be acquired by horizontal gene transfer ( Figure 6 ) . Deletion mutants of either PSPPH_A0107 or avrD resulted in reduced virulence on French bean . The slightly reduced virulence we observed is in conflict with observations that allelic replacement of avrD by the nptII gene did not result in any growth defect in completive index assays [72] . This discrepancy could be explained by transcription from the nptII promoter in the previous work , or by the use of different growth assays , time points , and bean cultivars . The PSPPH_A0106-A0112 operon is most likely involved in small molecule ( s ) synthesis promoting bacterial growth on host plants . Component ( s ) synthesized by the products of this operon and their effect on plants remain to be determined . However , since syringolides can be made from AvrD-expressing E . coli , and since the PSPPH_A0106-A0112 operon is not present in E . coli , we speculate that that the PSPPH_A0106-A0112 operon is not required for syringolide production . When present , AvrD shares no less than 84% amino acid identity across P . syringae strains . Genes encoding an AvrD-like protein with about 30% identity are widely distributed among bacteria , including Bacillus , Streptomyces and Vibrio . In general , these avrD-like genes are not found as singletons , but instead are linked to genes encoding various enzymes not related to any of the PSPPH_A0112-A0106 genes . In Streptomyces coelicolor A3 ( 2 ) , AvrD is part of an mmy operon responsible for synthesis of methylenomycin [88] . The PSPPH_A0110 to PSPPH_A0107 locus and to some extent the PSPPH_A0106 genes have similarity to genes in operons from Xanthomonas , Acidovorax , Pectobacterium and Ralstonia . Only the Ralstonia solanacearum PSI07 megaplasmid , carries both an avrD-like gene and a PSPPH_A0110-A0106 cluster of genes , but they are not contiguous on this plasmid . PSPPH_A0112 is mainly limited to P . syringae , but shares some homology with HMPREF9336_00100 ( 29% amino acid identity ) found in Segniliparus rugosus ATCC BAA-974 , an opportunistic pathogen associated with mammalian lung disease [89] . HMPREF9336_00100 and an avrD-like gene are linked in Segniliparus rugosus , being separated by only two genes and encoded on the same strand . We additionally observed that this operon has been disrupted by insertion of a transposable element in P . syringae CC1629 , reminiscent of transposon disruptions of T3E genes commonly observed across the P . syringae phylogeny [3] . hrpL is widely distributed , and tightly linked in all hrp/hrc group I T3SS [1] and the non-canonical T3SS found in some P . syringae , as well as the T3SS of P . viridiflava , P . fluorescens , Erwinia , Pantoea stewartii , and Dickeya . It is the key virulence regulator in most if not all of these species . Our work highlights the advantages of integrating next generation transcriptional and genomic data to better understand the role of non-T3E HrpL regulon genes in plant-pathogen interactions . Our approach is readily applied to strains with sequenced genomes and broad phylogenetic sampling [63] to better understand P . syringae virulence mechanisms and their evolution .
For maintenance and transformation , P . syringae were grown in King's B media ( KB ) at 28°C . E . coli DH5α was grown in Luria-Bertani ( LB ) media at 37°C . Antibiotics were used at the following concentrations: 50 µg/ml rifampicin , 25 µg/ml kanamycin , 10 µg/ml tetracycline , and 25 µg/ml gentamycin , according to vector selection . Strains used or analyzed in this study are listed with their abbreviation in Table S8 . Native hrpL from the various P . syringae were PCR amplified using LA-Taq ( TaKaRa ) and oligonucleotides listed in Table S9 containing XbaI and Hind III sites , then cloned into pTOPO-TA ( Invitrogen ) . The pTOPO-TA::hrpL was sequenced , digested with XbaI and HindIII , and cloned into NheI/HindIII-digested pCF340 ( Newman and Fuqua , 1999 ) and designated pBAD::hrpL . Porcurated_02784 , 04644 , 03530 , 01245 , 04371 and their respective upstream region containing the hrp-box were PCR amplified using Pfx ( Invitrogen ) and primers described in Table S9 . Resulting PCR fragments were cloned into pENTR-D-TOPO ( Invitrogen ) and sequenced . Porcurated_04644 was amplified similarly using primers containing attB1/attB2 sites and cloned into the pDONR 207 vector ( Invitrogen ) . All resulting constructs were sub-cloned into either the gateway-compatible pJC532 vector containing the in-frame Δ79avrRpt2 sequence for translocation assays or the pJC531 vector containing an in-frame HA sequence [3] to check for protein expression . All vectors used in this study were transformed into P . syringae strains using tri-parental mating with an E . coli helper strain containing pRK2013 . Pseudomonas strains containing pBAD::hrpLnative or pBAD::EV were grown overnight at 28°C , in KB media supplemented with tetracycline , then sub-cultured in fresh media at OD600 = 0 . 2 , and grown until OD600 = 0 . 4–0 . 5 . Bacteria were washed twice with 10 mM MgCl2 and resuspended in minimal medium [48] ( MM is 50 mMKPO4 pH 5 . 7 , 7 . 6 mM ( NH4 ) 2SO4 , 1 . 7 mM MgCl2 , 1 . 7 mM NaCl ) containing 10 mM mannitol and supplemented with 1% glycerol and 0 . 1% peptone which suppresses hrpL induction . Bacteria were then inoculated in supplemented minimal media at OD600 = 0 . 2 , and incubated shaking for 30 min at 28°C . Expression of hrpL was induced by addition of 200 mM L-arabinose . Aliquots of cell culture were taken 1 , 3 , 5 hours post-induction and treated with RNAprotect reagent ( Qiagen ) . RNA isolation was performed by using the RNeasy minikit ( Qiagen ) . Isolated total RNA was treated twice with TURBO DNase ( Ambion ) . Total RNAs derived from each time point were pooled at a 1 to 3 ratio . 10 µg of pooled total RNA was depleted of 16S and 23S ribosomal RNA using RiboMinus ( Invitrogen ) . cDNA were prepared from ∼1 µg of ribosomal depleted RNA , using random hexamer primers and Superscript II reverse transcriptase ( Invitrogen ) . Second strand cDNA was prepared using DNA polymerase I and Ribonuclease H ( Invitrogen ) . Double stranded cDNA was purified using Qiaquick spin columns ( Qiagen ) and eluted with EB buffer . Double stranded cDNA was sheared using a Covaris Disruptor . Library was prepared according to the manufacturer's protocol ( Illumina ) . Sequencing of the library was performed according the manufacturer's protocol on either Illumina GAII including single-end , 36 cycles or Illumina HiSeq 2000 including single-end , 70 cycles . We analyzed our RNA reads using the GENE-counter pipeline . For the PtoDC3000 , PsyB728A , and Pph1448A datasets , we used the publically available genomes provided by NCBI , along with the transcriptome constructed by NCBI's gene prediction pipeline . For the Por , Pjap and Pla107 dataset , we used in-house assembly for the genome and used JGI's Integrated Microbial Genomes – Expert Review gene prediction pipeline for the transcriptome . All ribosomal RNA genes were excluded from the transcriptome file for all datasets . Transcriptome sequences for each strain were blasted against their corresponding genome and GFF files were constructed from the Blast reports using an in-house script . We processed the RNA reads and aligned the reads using the default parameters of GENE-counter's CASHX read mapping algorithm . Reads mapping to multiple genomic locations were excluded . Annotated genes were included in the analysis only if at least one read in one sample matched that gene which can lead to highly duplicated genes not being considered . The false discovery rate cutoff for determining differential expression was set to 0 . 05 . We made a small modification to GENE-counter's findDGE . pl script that allowed for random seeding during the sample depth normalization process . By repeating the normalization process 300 times we generated B-values to measure and control for normalization effects . The GenBank accession ( http://www . ncbi . nlm . nih . gov/ ) and Gold ID ( http://img . jgi . doe . gov/cgi-bin/w/main . cgi ) of the genomes used in this study are CP000058-CP000060 , Gi04410 , CP000075 , Gi07003 , Gi03478 , and AE016853-AE016855 . RNA-seq data have been deposited in NCBI Gene Expression Omnibus and will be accessible through GEO Series accession number GSE46930 ( http://www . ncbi . nlm . nih . gov/geo/ ) . First , protein sequences of genes found up-regulated in our analysis with B-values≥50% were used to search each genome used in this study with BlastP to identify genes split up in different contigs/scaffolds . Possible duplication was ruled out by comparing the size of the query to the size of the subject sequence ( of complete genomes , principally ) . Putative sequencing errors leading to stop codons and discontinuous ORFs , led to consecutive queries matching the same subject sequence . Only the entry with the most significant q-value was kept . Secondly , genes encoding open reading frames shorter than 60 amino acids were excluded from our data set . Thirdly , loci of genes not previously found HrpL-dependent were assessed for linkage to genes with a hrp-box . As previously described [41] , [90] , we observed potential transcriptional read-through artifacts for which directly HrpL-targeted genes led to apparent up-regulation of adjacent genes . Therefore , genes found differentially expressed adjacent to a HrpL-regulated gene , but on the opposite DNA strand were considered to be putative transcriptional read-through and removed from our analysis . Genes encoded on the same strand as the HrpL-regulated gene were kept . Fourth , genes with a hrp-box embedded within their ORF on either sense or anti-sense strands were not included . Adjacent genes encoded on the same strand as the manually predicted hrp-box were included in the defined HrpL regulon , but genes on the opposite strand of the hrp-box were excluded . All genes removed from the HrpL regulons after manual curation are listed Table S4 . For native gene expression , bacteria were grown for 4 hours in KB media from OD600 = 0 . 2 , washed twice with sterile 10 mM MgCl2 and transferred into MM minimum media containing 10 mM mannitol for PtoDC3000 , PsyB728a , Pla107 , Pja , Por strains or MM minimum media containing 10 mM fructose for Pph1448A strain . Cells were collected after 5 hours of incubation shaking at 28°C and treated with RNAprotect reagent ( Qiagen ) . Total RNA derived from cells grown in MM media or arabinose inducing media ( as above ) was extracted using the RNeasy minikit ( Qiagen ) , DNase treated twice ( Ambion Turbo DNase ) , and cleaned up with Qiagen RNeasy Mini kit . Reverse transcription was performed using SuperScript II ( Invitrogen ) with 2 µg total RNA . Diluted cDNA was analyzed with SYBR green ( Applied Biosystem ) using the Opticon 2 System ( BioRad ) . Primers used are listed in Table S9 . Four week old Col-0 and Col-0 rps2–101c ( rps2 ) plants were hand inoculated with PtoDC300028E [57] carrying Δ79avrRpt2 fusion clones at OD600 = 0 . 1 . Plants were scored for Hypersensitive Response ( HR ) and pictures were taken 24 h after inoculation . Knockout constructs were generated using MTN1907 , a modified version of pLVC-D which allows for SacB counter-selection [3] , [91] . To create Pph1448AΔPSPPH_A0107 , Pph1448AΔPSPPH_A0113 mutants , 5′ and 3′ regions flanking the gene of interest were amplified using Pfx ( Invitrogen ) and combined by overlap extension PCR ( Table S9 ) , then cloned into pENTR-D-TOPO and sequenced . To generate the Pph1448AΔPSPPH_1855 , PsyB728aΔhrpL and PorΔhrpL mutants , nucleotide sequences corresponding to the fused flanking regions of each gene were synthesized including Gateway recombination sites and cloned in the pUC17 vector ( GenScript ) . All five clones were recombined into MTN1907 and transformed into either Pph1448A , PsyB728a or Por by tri-parental mating . After selection on tetracycline plates , merodiploids resulting from homologous recombination were verified by PCR . Two independent merodiploids carrying either a 3′ or 5′ insertion were grown on KB agar containing 5% sucrose to select for the loss of sacB via a second recombination event . Putative clean-deletion mutants were verified by PCR using flanking primers and gene specific primers . Before inoculation , Pph1448A and mutants were grown overnight and sub-cultured from OD600 = 0 . 2 for 4 hours in KB media , then washed twice with 10 mM MgCl2 . Two week old French bean cv . Tendergreen improved ( Livingston Seed Co . ) were dip inoculated with freshly grown bacteria at OD600 = 0 . 001 bacteria in 10 mM MgCl2 and 0 . 04% Silwet L-77 . Four plants were dip inoculated for each strain . Three days and an half after inoculation leaf discs were cored ( 12 to 16 replicates , each 4 cores ) , ground in 10 mM MgCl2 , serially diluted and plated on KB/50 µg/ml rifampicin and bacteria counted . Each set of mutants were tested side by side with the wild type strain at least 3 times . | Pseudomonas syringae are environmentally ubiquitous bacteria of wide phylogenetic distribution , which can cause disease on a broad range of plant species . Pathogenicity requires the master regulator HrpL . HrpL controls the activation of virulence factor genes , including those encoding the type III secretion system which facilitates translocation of bacterial proteins into host cells . Here we overlaid transcriptome profiling of genes onto their phylogenetic distribution by characterizing the HrpL regulon across six diverse strains of P . syringae . We identified novel putative virulence factors , discovered two novel effector families , and functionally characterized an operon most likely involved in secondary metabolism that we show is required for virulence . We demonstrated that the size and composition of the HrpL regulon varies among strains , and explored how genes are recruited into , or lost from , the virulence regulon . Overall , our work widens the understanding of P . syringae pathogenicity and presents an experimental paradigm extensible to other pathogenic bacterial species . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2014 | Variable Suites of Non-effector Genes Are Co-regulated in the Type III Secretion Virulence Regulon across the Pseudomonas syringae Phylogeny |
The discovery of the life-threatening zoonotic infection Plasmodium knowlesi has added to the challenges of prompt and accurate malaria diagnosis and surveillance . In this study from Aceh Province , Indonesia , a malaria elimination setting where P . knowlesi endemicity was not previously known , we report the laboratory investigation and difficulties encountered when using molecular detection methods for quality assurance of microscopically identified clinical cases . From 2014 to 2015 , 20 ( 49% ) P . falciparum , 16 ( 39% ) P . vivax , 3 ( 7% ) P . malariae , and 2 ( 5% ) indeterminate species were identified by microscopy from four sentinel health facilities . At a provincial-level reference laboratory , loop-mediated isothermal amplification ( LAMP ) , a field-friendly molecular method , was performed and confirmed Plasmodium in all samples though further species-identification was limited by the unavailability of non-falciparum species-specific testing with the platform used . At a national reference laboratory , several molecular methods including nested PCR ( nPCR ) targeting the 18 small sub-unit ( 18S ) ribosomal RNA , nPCR targeting the cytochrome-b ( cytb ) gene , a P . knowlesi-specific nPCR , and finally sequencing , were necessary to ultimately classify the samples as: 19 ( 46% ) P . knowlesi , 8 ( 20% ) P . falciparum , 14 ( 34% ) P . vivax . Microscopy was unable to identify or mis-classified up to 56% of confirmed cases , including all cases of P . knowlesi . With the nPCR methods targeting the four human-only species , P . knowlesi was missed ( 18S rRNA method ) or showed cross-reactivity for P . vivax ( cytb method ) . To facilitate diagnosis and management of potentially fatal P . knowlesi infection and surveillance for elimination of human-only malaria in Indonesia and other affected settings , new detection methods are needed for testing at the point-of-care and in local reference laboratories .
Plasmodium knowlesi is a newly emergent zoonotic human malaria species previously thought to only infect macaques . Since the first report of a human case from Peninsular Malaysia in 1965 [1] and the large cluster of human knowlesi malaria in Sarawak in 2004 [2] , endemic cases have been reported from other Asian countries including Brunei , Cambodia , India , Malaysia , Myanmar , Philippines , Singapore , Thailand , Vietnam , Indonesian Borneo [3–5] , and more recently Sumatra Island [6 , 7] . The identification of P . knowlesi infection is important for clinical and public health reasons . Infection in humans is most often uncomplicated , but 6–9% of symptomatic patients develop severe malaria and 0 . 3–1 . 8% of cases die [8–10] . Fatal outcomes have been associated with misdiagnosis of parasite species by microscopy , resulting in delays in appropriate management [11 , 12] . From a public health perspective , malaria control programs aim to decrease morbidity and mortality from all Plasmodium species affecting humans . As P . knowlesi infection is associated with a number of different risk factors than infections caused by other Plasmodium species [6 , 13] ( e . g . forest-related exposures ) , it may require different interventions . For subnational and national areas aiming to achieve and maintain malaria elimination , or the interruption of local transmission of human-only species , as is the goal in Indonesia , accurate species identification is critical . In most of Asia , microscopy is the standard for malaria diagnosis and surveillance . However microscopy has recognized limitations in diagnostic accuracy and species identification [14] . For P . knowlesi specifically , different asexual blood stages can resemble P . falciparum and P . malariae , and in routine practice it is misidentified as all human-only species [15] . Therefore , a variety of PCR methods have been utilized to distinguish P . knowlesi from other Plasmodium species [16 , 17] . With its simpler requirements and faster turnaround time , loop mediated isothermal amplification ( LAMP ) , another nucleic acid-based detection method , may be a more practical alternative in resource-limited field settings [18–20] . However , the relative benefits and limitations of LAMP and the various other PCR methods are not clear , particularly for field settings . To support malaria elimination efforts in Aceh Province , Indonesia , a pre-elimination area with known endemicity of P . vivax and P . falciparum , we introduced the use of molecular detection for quality assurance of microscopy-identified cases from health facilities by establishing LAMP testing at the provincial level reference laboratory . As previously reported , the finding of indeterminate species triggered further molecular testing that led to the first reported finding of P . knowlesi in Indonesia outside of Borneo [6] . Epidemiological investigation revealed that P . knowlesi infection was associated with forest exposures , particularly overnight stays due to work [6] . In this study , we present the laboratory details of this real-world investigation whereby the use of serial molecular detection methods including LAMP , two nPCR methods , P . knowlesi-specific nPCR , and sequencing led to the identification and confirmation of P . knowlesi infection . Challenges encountered in this experience have relevance to malaria diagnosis and surveillance in other settings where P . knowlesi may be present and can inform research and development of improved P . knowlesi detection methods .
The study was conducted in Aceh Besar District , Aceh Province , Sumatra island , Indonesia , a low-transmission setting that aims to eliminate malaria by 2020 . The 2013 incidence of malaria was 0 . 4/1000 , and 68 ( 39% ) of cases were reported as P . vivax , 71 ( 41% ) as P . falciparum , and the remaining 34 unspecified or mixed P . falciparum/P . vivax [6] . The sentinel sites included five primary health centers that reported 78% of all cases reported in Aceh Besar in 2013 . During the study period June 2014 to December 2015 , 41 patients were diagnosed with microscopy-confirmed malaria and recruited for enrolment . This number of cases was a convenience sample from an umbrella study where health facility-identified cases triggered active case finding in villages [6] . After written consent was obtained and prior to treatment , venous blood was collected and partly used to prepare dried blood spots ( DBS ) using Whatman 3MM paper . DBS along with remaining whole blood were initially stored at 4°C , transferred to -20°C within a week of collection , and then stored at -80°C . Antimalarial treatment was based on microscopy results and according to Indonesian government’s national policy . Ethical approval for the study was obtained from the National Institute of Health , Research and Development of the Indonesian Ministry of Health ( number LB . 02 . 01/5 . 2/KE . 111/2014 and LB . 02 . 01/5 . 2/KE . 211/2015 ) and IRB Committee of the University of California , San Francisco . Written informed consent was obtained from all adults or a parent or guardian for participants less than 18 years of age . For quality assurance of microscopy performed at health centers , blood smears were re-read by certified microscopists at the provincial laboratory according to national guidelines . For further quality assurance at the provincial-level , LAMP was selected due to its field-friendly platform . Initial extraction of DNA and LAMP testing were performed at the provincial laboratory . DNA was extracted from DBS using the Saponin/Chelex method [21] . Pan-LAMP testing followed by Pf-LAMP specific testing for Pan-LAMP positive samples was also performed using the commercially available Loopamp MALARIA Pan/Pf detection kit in accordance to manufacturer’s instructions ( EIKEN Chemical , Co . , Ltd . , Japan ) . Species identification for non-falciparum species was not available with this LAMP platform , but this was not anticipated to be a problem because Aceh was considered to be endemic for only P . falciparum and P . vivax malaria before the study was launched [22] . As such , Pan-LAMP positive , Pf-LAMP negative samples were expected to be P . vivax . Further molecular testing was performed at the Malaria Pathogenesis laboratory at the Eijkman Institute in Jakarta , using chelex-extracted DNA from a second DBS . Genus-specific PCR targeting the mitochondrial cytb gene followed by AluI enzyme digestion for species identification of the four main human species was used initially , as previously described [23] . After a report of indeterminate species and suspicion of P . knowlesi by a field microscopist , as well as limited data on the performance of the cytb nPCR method for detection of P . knowlesi , additional methods were employed including nPCR testing targeting the 18S rRNA gene for the four human-only species [24] , and P . knowlesi-specific nPCR [16] for all samples . For a proportion of samples testing positive by P . knowlesi specific nPCR , DNA was extracted from whole blood using the QIAamp DNA Mini kit ( Qiagen , CA ) and Sanger targeted genome sequencing [25] was performed ( Eijkman Institute Sequencing Facility ) . To prevent DNA contamination , all extractions were performed in rooms separate from where amplification was conducted . Extracted DNA was stored at -20°C . Results from microscopy and each molecular method were compared to a gold standard established through serial molecular testing: P . falciparum and P . vivax classification were based on species-specific positivity by both cytb and 18S rRNA nPCR , and P . knowlesi classification was based on genus-specific PCR positivity by both cytb and 18S rRNA nPCR and P . knowlesi-specific nPCR positivity . With regards to diagnostic performance for species identification , we were not able to calculate sensitivity , specificity , or negative predictive value ( NPV ) due to having not included a representative sample of microscopy-negative infections . However , positive predictive values ( PPV ) were calculated .
From June 2014 to December 2015 , 41 malaria cases were included in the study analysis . Forty-two were initially identified from the sentinel health facilities by microscopy and confirmed by cross-checking at the provincial laboratory , but one case ( P . vivax by microscopy ) was excluded as the DBS had insufficient blood for subsequent molecular testing . The 41 cases included: 20 P . falciparum ( 49% ) , 16 P . vivax ( 39% ) , 3 P . malariae ( 7% ) , and 2 with indeterminate morphology ( 5% ) ( Table 1 ) . Parasite density ranged from 66 to 355 , 400 parasite/μL blood . The median and range of parasite density ( in brackets ) for microscopy-diagnosed P . falciparum , P . vivax and P . malariae were 5 , 447 ( 66 to 54 , 970 ) , 32 , 157 ( 703 to 355 , 400 ) and 3 , 842 ( 1 , 760 to 7 , 133 ) . The parasite densities of the indeterminate samples were 803 and 1 , 473 , respectively . Microphotography of the indeterminate samples showed resemblance to other species ( Fig 1 ) . Genus-specific Pan-LAMP testing at the provincial laboratory was positive in all 41 isolates ( examples in Fig 2 ) , and 8 tested positive by Pf-LAMP testing ( Table 1 ) . By cytb PCR genus-specific testing and using the AluI restriction digest reaction for species identification , 8 ( 19 . 5% ) were classified as P . falciparum , 33 ( 80 . 5% ) as P . vivax . By 18S rRNA nPCR , there were 8 P . falciparum ( 19 . 5% ) , 14 P . vivax ( 34 . 1% ) , and 19 ( 46 . 3% ) did not amplify . P . knowlesi-specific nPCR was positive in 19/41 ( 46 . 3% ) , of which 11 underwent sequencing and showed 100% identity to a published P . knowlesi 18S rRNA gene sequence ( P . knowlesi strain H1 chromosome 3 , GenBank accession number AM910985 ) . Microscopy was unable to classify or mis-classified 23 of 41 ( 56% ) malaria cases confirmed by the gold standard of serial molecular testing ( Table 1 ) . These included all 19 P . knowlesi cases , of which 17 were mis-classified as P . falciparum ( n = 8 ) , P . vivax ( n = 6 ) , or P . malariae ( n = 3 ) , and 2 were unable to be classified . There were also 4 P . vivax cases that were mis-classified as P . falciparum by microscopy . Sixty percent ( 12/20 ) of cases identified by microscopy as P . falciparum were either P . vivax or P . knowlesi; 37 . 5% ( 6/16 ) of cases identified by microscopy as P . vivax were P . knowlesi . All P . malariae and indeterminate species by microscopy were P . knowlesi . Genus-specific testing by LAMP identified all infections , though species identification was limited by the unavailability of non-falciparum species-specific testing with the platform used . Pf-LAMP testing mis-classified one P . knowlesi mono-infection as P . falciparum but otherwise correctly identified all the P . falciparum cases . Of cases classified as P . vivax by cytb PCR , 58% ( 19/33 ) were later confirmed as P . knowlesi and showed a similar banding pattern to P . vivax ( Table 1 and Fig 3A ) . Using 18S rRNA species-specific nPCR for the four main human species , P . falciparum and P . vivax were correctly identified but all P . knowlesi infections were missed ( Fig 3B ) . There was no cross-reactivity with P . vivax using P . knowlesi-specific nPCR ( Fig 3C ) . The positive predictive values ( PPV ) for species identification by different diagnostic methods using the gold standard of serial molecular testing are shown in Table 2 . PPV was low for P . falciparum , P . vivax , and P . malariae identification by microscopy and for P . vivax identification by cytb nPCR . Where samples were available , PPV was high for all other methods .
To support malaria diagnosis and surveillance in Aceh Province , a low transmission setting in Indonesia that is aiming for malaria elimination , we utilized molecular testing for quality assurance of microscopy-confirmed cases from health facilities . As previously published , this work resulted in the first report of P . knowlesi in Indonesia outside Borneo , and an epidemiological investigation revealed that forest exposures are a key risk factor for this zoonotic infection [6] . In this study , we report the details and difficulties of species identification using microscopy at the point of care and a variety of molecular methods at reference laboratories . Microscopy mis-classified P . knowlesi cases as P . malariae or P . falciparum , as commonly reported elsewhere , but also as P . vivax , which has been less commonly reported [15] . The PPVs for the identification of other species ( Pf , Pv , and Pm ) were also poor . At the provincial reference laboratory , LAMP , a field-friendly molecular method , was useful in confirming all Plasmodium infections , though further species identification was limited by the unavailability of non-falciparum species-specific testing with the platform used . Use of less field-friendly nPCR methods at a national reference laboratory to identify P . knowlesi infection was not straightforward . All P . knowlesi cases did not amplify with a standard nPCR method ( 18S rRNA ) targeting the four human-only species . With the cytb method , there was cross-reactivity with P . vivax for all P . knowlesi cases . We highlight the difficulties of P . knowlesi diagnosis at the point-of-care and reference laboratory levels in a setting where endemicity was not previously known and bring attention to an emerging challenge for malaria elimination . The recent discovery and emergence of P . knowlesi , a fifth human species previously thought to only infect macaques , has created an additional challenge for species identification . Microscopy is difficult because the morphology at different stages resembles other malaria species [26] . The diagnostic sensitivity and specificity of available immunochromatographic rapid diagnostic tests ( RDTs ) for P . knowlesi detection is poor , leaving no other useful point-of-care diagnostic test [27–29] . Despite some global knowledge on the potential geographical distribution and extent of transmission of P . knowlesi [4] , this information may lack resolution at local levels , and health workers and microscopists on the front-lines may have limited knowledge and/or a low index of suspicion for P . knowlesi . In our study , the investigation into P . knowlesi was prompted by the observation by an astute microscopist of unusual morphology in two malaria cases , as well as the known local presence of pig-tailed and long-tailed macaques and Anopheles leucosphyrus , a known vector on Sumatra island [30] . For quality control in reference laboratories , none of the nucleic acid-based methods for both genus and species-specific identification were found to be suitable . With LAMP , a molecular detection method that has been promoted for use in resource-limited settings due to the rapid turnaround time and simple methods , genus-specific testing was reliable , as has been reported from Malaysia [20] . However a P . knowlesi-specific commercial kit was not available for use in our study , and evaluations of other P . knowlesi-specific LAMP assays have reported cross-reactivity with P . vivax [18] . The P . knowlesi-specific PCR method utilized in this study did not cross-react with P . vivax infections , with excellent specificity as observed previously [16] . The nPCR methods used have problems with missed infections and/or species mis-classification . With commonly used 18S rRNA nPCR targeting the four human-only species , a commonly used reference standard , P . knowlesi either does not amplify ( as occurred in this study ) or is mis-classified as P . vivax due to high sequence homology at the target sequences [31 , 32] . With the cytb nPCR method that we used , our finding of cross-reactivity between P . knowlesi and P . vivax has not been previously reported , but can also be explained by high sequence homology at the target mitochondrial sequences . Others have reported P . knowlesi amplification using P . vivax-specific PCR targeting the mitochondrial gene cox1 [33] . Other more sensitive and specific molecular methods for P . knowlesi detection in mixed species settings have recently been developed [7 , 34 , 35] and could be considered for future surveillance in our study setting . The challenge of accurate P . knowlesi detection is of both clinical and public health significance . In Malaysia , where the clinical disease has been well studied , P . knowlesi is associated with at least as high a risk of severe disease compared with P . falciparum [36] and in early series , a high proportion had fatal outcomes [8 , 37] . Following a number of interventions in Sabah state , case-fatality rates have fallen 6-fold [9] . These have included improved and now routine statewide molecular surveillance , more recent laboratory microscopy reporting of “P . malariae” as “P . knowlesi” , and enhanced implementation of standardized referral and clinical protocols , including first-line use of artemisinin-based combination therapy and early intravenous artesunate [9 , 36] . Progression to severe disease is due not only to missed diagnoses , but also its ability to cause severe malaria at relatively low parasite densities [36] . Mis-classification of P . knowlesi as P . vivax , as occurred at the point of care in our study , also results in unnecessary treatment with primaquine , an antimalarial not indicated for P . knowlesi , but necessary for radical cure of the latent liver stages with P . vivax . In our study , we did not experience any severe adverse events from the unnecessary use of primaquine , but use in subjects with underlying severe glucose-6-phosphate dehydrogenase deficiency is known to be associated with life-threatening hemolysis . While only recently recognized in areas of Aceh and North Sumatra , there has been little molecular surveillance of P . knowlesi distribution and incidence elsewhere in Indonesia , particularly across Kalimantan , Sulawesi and other regions of Sumatra , where modelling predicts a high risk of human infection [38] . From a public health perspective , accurate identification of P . knowlesi is critical to the design and implementation of effective malaria interventions . In a related study in Aceh Province and also in Malaysia , adult males with forest-related and agricultural occupational exposure are at significantly higher risk of being infected with P . knowlesi [6 , 13] . Interventions would therefore need to be targeted to this population . As well as continued promotion of conventional malaria prevention activities to reduce peridomestic transmission [13] , other interventions would need to be targeted to P . knowlesi-transmitting mosquitos , the interface between humans and macaques , and to individual risk factors for infection identified in different settings . Further investigation into the epidemiology and transmission of P . knowlesi in Aceh Besar is needed .
Limitations of microscopy to identify P . knowlesi are well established . Our challenges using LAMP and PCR for species identification in a setting with previously unknown P . knowlesi endemicity add to a growing literature on the limitations of molecular methods as well . For settings approaching malaria elimination and/or where epidemiological conditions are predicted to support P . knowlesi transmission to humans , quality assurance of malaria diagnosis and species identification is essential , but at present , practical and accurate methods are not available for local and peripheral reference laboratories . Development , evaluation and implementation of improved P . knowlesi detection methods for use at both the point-of-care and in local reference laboratories are needed . | In Southeast Asia , Plasmodium knowlesi , a malaria parasite of macaques , was recently discovered to infect humans . This emerging disease is important because it has potential for causing severe disease and death , and it is a threat to malaria elimination efforts in the region . In this report from Aceh Province , Indonesia , where P . knowlesi was only recently discovered , the authors report on the laboratory challenges of distinguishing this species from other human species . Using several different molecular methods , they investigated 41 malaria cases which by microscopy , were initially reported as: P . falciparum ( 49% ) , P . vivax ( 39% ) , P . malariae ( 7% ) , and indeterminate ( 5% ) . Only after using a P . knowlesi-specific nPCR method and sequencing , did they find that nearly half were P . knowlesi . Consistent with a sparse literature , a field-friendly molecular method ( genus-specific LAMP ) reliably detected P . knowlesi , while use of a more standard reference laboratory molecular method ( 18S rRNA nPCR targeting the four human-only species ) missed the infections . Also another reference laboratory molecular method ( cytb nPCR ) mis-classified P . knowlesi infections as P . vivax due to cross-reactivity . To address the emerging threat of P . knowlesi , new detection methods are needed for point-of-care and reference testing . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion",
"Conclusions"
]
| [
"parasite",
"groups",
"medicine",
"and",
"health",
"sciences",
"plasmodium",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"animals",
"parasitology",
"mammals",
"indonesia",
"primates",
"apicomplexa",
"protozoans",
"molecular",
"biology",
"techniques",
"cellular",
"structures",
"and",
"organelles",
"old",
"world",
"monkeys",
"research",
"facilities",
"research",
"and",
"analysis",
"methods",
"monkeys",
"malarial",
"parasites",
"artificial",
"gene",
"amplification",
"and",
"extension",
"molecular",
"biology",
"research",
"laboratories",
"ribosomes",
"macaque",
"people",
"and",
"places",
"biochemistry",
"rna",
"eukaryota",
"ribosomal",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"asia",
"polymerase",
"chain",
"reaction",
"oceania",
"biology",
"and",
"life",
"sciences",
"government",
"laboratories",
"malaria",
"non-coding",
"rna",
"amniotes",
"organisms"
]
| 2018 | Laboratory challenges of Plasmodium species identification in Aceh Province, Indonesia, a malaria elimination setting with newly discovered P. knowlesi |
Duplications at 15q11 . 2-q13 . 3 overlapping the Prader-Willi/Angelman syndrome ( PWS/AS ) region have been associated with developmental delay ( DD ) , autism spectrum disorder ( ASD ) and schizophrenia ( SZ ) . Due to presence of imprinted genes within the region , the parental origin of these duplications may be key to the pathogenicity . Duplications of maternal origin are associated with disease , whereas the pathogenicity of paternal ones is unclear . To clarify the role of maternal and paternal duplications , we conducted the largest and most detailed study to date of parental origin of 15q11 . 2-q13 . 3 interstitial duplications in DD , ASD and SZ cohorts . We show , for the first time , that paternal duplications lead to an increased risk of developing DD/ASD/multiple congenital anomalies ( MCA ) , but do not appear to increase risk for SZ . The importance of the epigenetic status of 15q11 . 2-q13 . 3 duplications was further underlined by analysis of a number of families , in which the duplication was paternally derived in the mother , who was unaffected , whereas her offspring , who inherited a maternally derived duplication , suffered from psychotic illness . Interestingly , the most consistent clinical characteristics of SZ patients with 15q11 . 2-q13 . 3 duplications were learning or developmental problems , found in 76% of carriers . Despite their lower pathogenicity , paternal duplications are less frequent in the general population with a general population prevalence of 0 . 0033% compared to 0 . 0069% for maternal duplications . This may be due to lower fecundity of male carriers and differential survival of embryos , something echoed in the findings that both types of duplications are de novo in just over 50% of cases . Isodicentric chromosome 15 ( idic15 ) or interstitial triplications were not observed in SZ patients or in controls . Overall , this study refines the distinct roles of maternal and paternal interstitial duplications at 15q11 . 2-q13 . 3 , underlining the critical importance of maternally expressed imprinted genes in the contribution of Copy Number Variants ( CNVs ) at this interval to the incidence of psychotic illness . This work will have tangible benefits for patients with 15q11 . 2-q13 . 3 duplications by aiding genetic counseling .
Recurrent duplications of ~4Mb at 15q11 . 2-q13 . 3 , overlapping the Prader-Willi syndrome /Angelman syndrome ( PWS/AS ) region between breakpoints 2 and 3 or 1 and 3 ( BP2-BP3 or BP1-BP3 ) are recognised risk factors for developmental delay ( DD ) and autism spectrum disorders ( ASD ) [1–4] . More recently , these duplications were implicated as risk factors for schizophrenia ( SZ ) [5–8] , however data are limited . The 15q11 . 2-q13 . 3 region contains a cluster of imprinted genes , which are expressed from one parental allele only as a consequence of germline epigenetic events ( Fig 1 ) . Within this cluster there are several paternally expressed genes , including SNRPN , MKRN3 , MAGEL2 and NECDIN , and two maternally expressed genes , namely ATP10A , and UBE3A . The genes within the 15q11 . 2-q13 . 3 interval are mostly canonical imprinted genes , in that expression is robustly monoallelic , although the imprinting status of ATP10A appears to be polymorphic and influenced by gender [9] . As a consequence of the presence of both maternally and paternally expressed imprinted genes , CNVs at this interval may be expected to have different phenotypes depending on their parent of origin . Indeed , most studies that have tested the parental origin of 15q11 . 2-q13 . 3 duplications show that those of maternal origin are usually responsible for the disease phenotypes . However , the penetrance of 15q11 . 2-q13 . 3 duplications has not yet been estimated . Moreover , whilst rare duplications of paternal origin have generally been regarded as benign [1 , 10] , this has not been studied systematically . The assumption that duplications of paternal origin are benign comes from small numbers of observations of healthy mothers who carry duplications of paternal origin ( and have transmitted them , making them maternal in the affected offspring ) , as well from the fact that they are much rarer in cohorts of DD/ASD children [1 , 10] . However , this pattern can also be explained by a lower penetrance of the paternal duplications and a lower prevalence in the general population . Only very large studies on patients and healthy controls can provide an accurate estimate of their role . The observations of paternal duplications occurring de novo , their extreme rarity in controls and their large size and high gene content , made us suspect that they are under selection pressure , like all other similar CNVs for which we have calculated the selection pressure [11] and therefore are likely to be pathogenic , although with a lower penetrance . The aim of this study was to conduct the largest and most detailed assessment of 15q11 . 2-q13 . 3 interstitial duplications to date . By examining large cohorts of DD , ASD and SZ , along with large numbers of controls , we were able to estimate the prevalence , penetrance , and selection coefficients of 15q11 . 2-q13 . 3 interstitial duplications of maternal and paternal origin for SZ and for other neurodevelopmental disorders , and to identify clinical features common to SZ carriers of these duplications . We clearly implicate 15q11 . 2-q13 . 3 interstitial duplications of paternal origin in the aetiology of DD , but do not find them at increased rates in SZ , which is significantly associated only with duplications of maternal origin . These data clarify the contribution of imprinted genes within the PWS/AS interval to psychopathology , and have important , tangible benefits for patients with 15q11 . 2-q13 . 3 duplications by aiding genetic counseling .
The prevalence rates of 15q11 . 2-q13 . 3 interstitial duplications in SZ , other neurodevelopmental disorders , and controls are shown in Table 1 . Among 28 , 138 SZ probands , there were 25 individuals with 15q11 . 2-q13 . 3 interstitial duplications , of whom 24 were of maternal and only one of paternal origin . Prevalence estimates in SZ were therefore 0 . 085% ( 95%CI = 0 . 057–0 . 13% ) for maternal and 0 . 0036% ( 95%CI = 0 . 00064–0 . 02% ) for paternal duplications . Among 51 , 001 probands with DD/ASD/MCA from two large studies on referrals to clinical genetics clinics with DD/ASD/MCA [18 , 21] , two from ASD cohorts and the current study [5 , 17] , 53 ( 0 . 1% ) had an interstitial 15q11 . 2-q13 . 3 duplication . Isodicentric chr15 ( idic15 ) and interstitial triplications reported in these 51 , 001 probands ( N = 6 ) were excluded from analysis , while no triplications were observed in SZ or control subjects . Only a small proportion of the previously published duplications had been reported for parental origin . In order to arrive at a more precise estimate of the ratio between maternal and paternal interstitial duplications in DD/ASD/MCA cohorts , we analysed an additional 13 individuals from the BBGRE database ( 20 , 260 individuals analysed for the current study ) and one new ASD case from Iceland , and included 37 DD/ASD/MCA subjects from three studies that estimated the parental origin of duplications in such patients , but reported no population prevalence data [10 , 19 , 20] ( Table 1 ) . This gave us a total of 60 DD/ASD/MCA subjects with established parental origin: 50 ( 83 . 3% ) maternal and 10 ( 16 . 7% ) paternal . Using these proportions , we extrapolated the prevalence rates of maternal and paternal duplications for the 53 DD/ASD/MCA systematically ascertained carriers at 0 . 087% ( 95%CI = 0 . 065–0 . 12% ) for maternal and 0 . 017% ( 95%CI = 0 . 009–0 . 033% ) for paternal duplications , respectively ( Table 1 ) . Among 149 , 780 controls , there were four individuals with maternal and four with paternal duplication origin , giving identical rates of 0 . 0027% for both parental types ( 95%CI = 0 . 001–0 . 0069% ) . Analysis of bipolar disorder ( BD ) datasets was not among the aims of this study . Just for the record , our review of 8 , 084 BD probands [22] found no 15q11 . 2-q13 . 3 duplication carriers . Although we reported one such case in our original publication [5] , the available data suggests that these duplications do not play any significant role in BD . Table 2 shows the estimates for the prevalence of 15q11 . 2-q13 . 3 duplications in different population groups , and the estimated penetrance for SZ and other neurodevelopmental disorders . We estimated the general population frequency of 15q11 . 2-q13 . 3 duplications of maternal origin to be about two times higher than those of paternal origin ( 0 . 0069% vs . 0 . 0033% ) . The penetrance of maternal duplications for DD/ASD/MCA is very high ( 50 . 5% ) and is about 2 . 5 times higher than those of paternal origin ( 20 . 7% ) . Maternal duplications also have high penetrance for SZ ( 12 . 3% ) but in contrast , paternal duplications appear not to increase risk for SZ ( penetrance of 1 . 1% ) , although we cannot completely exclude their role in SZ , as this estimate is based on a single observation in a SZ patient and therefore has large 95%CIs . We describe 10 relatives of probands that carried 15q11 . 2-q13 . 3 duplications ( Fig 2 and S1 Table ) . Five of those were affected with psychosis or DD/ASD and all had 15q11 . 2-q13 . 3 duplications of maternal origin . Of the 5 unaffected relatives , only two had duplications of maternal origin ( one of these , 50320–1 , had depressive disorder ) . The three remaining unaffected individuals with paternal duplications had no reported neuropsychiatric phenotypes ( but had not been specifically assessed ) . Six of the relatives were transmitting mothers and two were transmitting fathers . Five of the six mothers had offspring affected with SZ , while the two transmitting fathers had one healthy and one Attention Deficit Hyperactivity Disorder ( ADHD ) offspring ( Fig 2 ) . The remaining affected individuals in Fig 2 , all marked with no fill ( white ) , were not tested , as we had no DNA from them , but they are compatible with having maternal duplications . We do not have sufficient data on offspring of 15q11 . 2-q13 . 3 duplication carriers to make direct estimates of reproductive fitness and therefore present results for selection coefficients based on the ratio between de novo and inherited duplications: de novo/ ( de novo+inherited ) , following our previous work [11] ( Table 3 ) . The estimated selection coefficients appear similar: 0 . 55 for maternal and 0 . 58 for paternal duplications ( but small numbers of paternal duplications preclude accurate estimates ) . Using the formula discussed in the Methods ( μ = qs ) , we estimated the mutation rates as follows: maternal duplications , 0 . 000069×0 . 55 ≈ 1 per 27 , 000 newborns and paternal duplications , 0 . 000033×0 . 58 ≈ 1 per 50 , 000 newborns . Here we use population rates , rather than allele frequencies ( q ) , and mutation rates per newborn , rather than per gamete ( μ ) , as the formula suggests , simplifying the presentation . We point out that the estimates for both s and μ are less reliable compared to those for other CNVs [11] , and that we refer to mutation rates in newborns , rather than in germ cells , which differ by orders of magnitude ( Discussion ) . Our collaboration compiled a series of 29 duplication carriers with SZ or schizoaffective disorder ( SZA ) , including the affected relatives ( S1 and S2 Tables ) , of which only a small proportion have been reported before [5 , 7 , 8] . The 20 SZ/SZA cases with data on age at onset had a relatively early mean age at onset of 18 . 1 ( SD = 6 . 9 ) years and in five of those the illness started during childhood ( <13 years ) . Developmental and/or cognitive data were available for 21 cases and 76% of them had recorded learning/developmental problems . These ranged from mild ID ( n = 3 ) to borderline ID ( n = 8 ) with the rest having unspecified or nonverbal learning difficulties . The median IQ score was 75 ( range 62–89 ) among the 11 cases with formal IQ tests . Specific psychiatric symptoms included catatonia ( n = 6 ) , disorganized behaviour ( n = 5 ) and prominent antisocial traits ( n = 5 ) . Epilepsy was reported in only one case . Among the 54 cases included in this study ( S1 and S2 Tables ) , there were three carriers where ADHD was listed as a phenotype . One of these was of paternal origin , from a study that tested 727 children with ADHD [23] . The only case with a paternal duplication among the 12 cases in the study of Aypar et al . [10] also had ADHD . Still , these numbers are clearly too small to conclude that paternal duplications have a specific role in ADHD . It should be noted that several of the population controls who carried duplications ( all from Iceland ) were not specifically subjected to formal neuropsychiatric assessment ( S1 Table ) . However , they had not been registered as psychiatric patients and there was no information to indicate that they had developmental delay . Interestingly , two of the controls ( control 2 and control 4 ) recently participated in a neuropsychiatric test battery project conducted in Iceland conducted by the Icelandic authors of the current paper . This revealed that Control 2 ( paternal origin duplication ) has an IQ of 100 and Global Assessment of Functioning ( GAF ) score of 81 and also is registered as having dyslexia , while Control 4 ( maternal origin duplication ) has an IQ of 84 and GAF score of 90 . A third “unaffected” control ( Control 3 ) had Alzheimer’s disease at the age of 64 . These observations indicate that even population controls who do not suffer with our target diagnoses , could have some subtle cognitive phenotypes , highlighting the variable penetrance of this CNV . Seventeen individuals analysed for this study ( probands with SZ/SZA , their affected relatives and one control ) , had extended clinical/physical data available ( S2 Table ) . Congenital anomalies were rare: one with cleft palate and one with cardiomegaly . Dysmorphic features were more common , recorded in 10 carriers , and included micro- , macro- or dolichocephaly , high palate , facial asymmetry and others . Six cases ( 35% ) had urological conditions including urethral stricture , polyuria , urge incontinence , nephrectomy and recurrent urinary tract infections . Endocrinology problems included hypocalcemia , hypercholesterolemia , diabetes mellitus and hypothyroidism , however we cannot state whether they are more common than in any population of SZ patients , or had been caused by medication .
Maternal duplications of the PWS/AS critical region at chromosome 15q11 . 2-q13 . 3 are known to be pathogenic , causing DD and intellectual disability [24] , and are also among the most common single genetic risk factors for ASD [1 , 25] . More recently they were also implicated as risk factors for SZ [5–8] , although with a much lower estimated penetrance than for DD/ASD [26] and based on very few observations . In contrast , paternal duplications have not been considered to be pathogenic [1 , 10] . Here we conducted the largest and most detailed assessment of interstitial duplications at 15q11 . 2-q13 . 3 to date . We were able to estimate the prevalence , penetrance , and selection coefficients of 15q11 . 2-q13 . 3 duplications of maternal and paternal origin for SZ and other neurodevelopmental disorders , and to identify clinical features common to SZ carriers of these duplications . We clearly implicate 15q11 . 2-q13 . 3 duplications of paternal origin in the aetiology of DD , but not for SZ , where only maternal duplications increased risk . Our data confirm that maternal interstitial duplications have a high penetrance ( 62 . 4% ) for any neurodevelopmental disorder . Much of this was accounted for by the DD/ASD/MCA group , and only 12 . 3% by SZ ( Table 2 ) . This CNV is found in about 1:1176 patients suffering with SZ ( 95%CI = 1:769–1:1754 ) . In contrast , paternal duplications do not appear to increase risk for SZ , ( penetrance of 1 . 1% ) , suggesting that only maternal duplications have a specific effect on psychosis ( although a small role for paternal duplications cannot be confidently excluded due to the wide confidence intervals ) . The importance of epigenetic status of duplications at this interval was further underlined by analysis of a number of families ( Fig 2 ) . Duplications in two unaffected mothers had a DNA-methylation pattern indicative of being paternally derived , whereas their offspring , who possessed a maternally derived duplication , suffered from psychotic illness . Although they appear to have no role in SZ , we show for the first time that paternal duplications are pathogenic , increasing the risk for DD/ASD/MCA with a penetrance of 20 . 7% ( 95%CI = 4 . 4–100 ) . One reason why paternal duplications have been regarded as non-pathogenic in the past is their rare occurrence in patients . Here we demonstrate that they are also rare in the general population as a whole , with prevalence rates of 0 . 0033% for paternal and 0 . 0069% for maternal duplications ( Table 2 ) . Their pathogenicity is supported by the strong selection pressure operating against them ( s = 0 . 58 , Table 3 ) . We tested if some of these results could be due to an overestimation of the population prevalence of DD/ASD/MCA or SZ , as those referred for testing or taking part in SZ studies might constitute a more severely affected sub-group . However , even if we use lower population rates of 0 . 5% for SZ and 2% for the combined group of ASD/DD/MCA , our penetrance estimates remain very high ( S3 Table ) . The penetrance might even be higher , if we include subtle cognitive phenotypes , as described above for those population controls who took part in formal neurocognitive testing . The question arises as to why paternal duplications are rarer than maternal ones in the general population , despite their lower pathogenicity . One explanation is a lower mutation rate in males . Indeed we estimate ( with great approximation ) that 1:27 , 000 and 1:50 , 000 newborns have a mutation that arose on maternal and paternal chromosomes , respectively . This refers to the rate in newborns with de novo mutations , while the mutation rate in sperm is much higher , observed in about 1:400 sperm cells [27] , indicating a strong negative selection against such embryos before birth . Differences in the mutation rates according to the parental origin have been shown for other CNVs as well [28] , so this may be one explanation of the observed difference . The ( possibly ) lower rate of de novo mutations of paternal origin among newborns cannot , on their own , explain their extreme rarity in the population . Firstly , paternal duplications should be less efficiently eliminated from the population by negative selection pressure , due to their lower penetrance for neurodevelopmental disorders . Secondly , some maternal duplications will change to paternal when transmitted from male carriers . We now suggest one further explanation for their rarity: male patients with SZ and other neurodevelopmental disorders have lower fecundity . Indeed , men suffering with SZ have only half the number of offspring compared to women with SZ [29] . This effect was demonstrated for another high-penetrance CNV , the 22q11 . 2 deletion , where male carriers had on average three times fewer offspring than female carriers with the same deletion ( 0 . 3 vs . 0 . 9 per carrier ) [30] . This parental bias could reduce substantially the number of inherited paternal 15q11-q13 duplications , compared to maternal ones . We did not observe any trend for SZ patients to have a particular size of duplication , identifying individuals with all four possible combinations of breakpoints between BP1 and BP4 ( Fig 1 ) . Within the region common to all these duplications ( BP2-BP3 ) , the most likely candidate gene casing the neurodevelopmental phenotypes is UBE3A ( Fig 1 ) . It is only expressed from the maternal allele in neurons [31] and so maternal , but not paternal , duplications spanning this gene would lead to an over-dosage of expression in the brain . ATP10A is also maternally expressed , but the polymorphic nature of the imprinting of this gene [9] makes the canonically imprinted UBE3A the main causal candidate from an epigenetic perspective . Moreover , UBE3A also has a strong neural pedigree as it encodes E6-AP ubiquitin ligase , which is important in the degradation of proteins such as p53 and the ubiquitins [32] , and has been shown to influence the glutamatergic system via its action on the synaptic protein Arc [33] . The role of this gene is supported by a recent clinical case with a micro-duplication encompassing only UBE3A , pointing to this gene as being key to the neuropsychiatric phenotypes [34] . The most consistent phenotypic characteristic of SZ carriers of these duplications is their cognitive deficit . Features of DD or intellectual deficit were recorded in 76% of carriers of maternal duplications on whom data was available . The median IQ score was 75 ( range 62–89 ) among those who had a test ( although this IQ result might be biased towards lower scores , as patients with more obvious learning problems are more likely to be referred for such testing ) . Congenital anomalies were fairly rare but mild dysmorphic features were observed in 10/17 cases with available clinical data . The interest in the clinical presentation of psychosis among carriers of 15q11 . 2-q13 . 3 duplications of maternal origin started following reports that individuals with Prader-Willi syndrome who had maternal uniparental disomy ( two maternally inherited copies and no paternal copy ) might present with cycloid psychoses ( acute polymorphic clinical pictures with prominent affective or motor symptoms ) [35 , 36] . In our original paper [5] we noted that one of our cases had SZA and one had prominent affective symptoms . Here we present more cases with extended clinical descriptions . The rate of SZA disorder was 20% ( 4 out of 20 ) for cases with more detailed clinical records . SZ/SZA patients tended to have an early age at onset ( mean of 18 . 1 years , SD = 6 . 9 ) . Catatonia was recorded in six patients and disorganised or aggressive behaviour in eight patients . Antisocial traits were noted in all five cases from Canada . Although no single clinical picture emerges , it appears that psychotic patients with maternal 15q11 . 2-q13 . 3 duplications are more likely to present with disorganised , aggressive , antisocial and/or catatonic features . We did not find cases with cycloid psychoses , suggesting this may be restricted to Prader-Willi syndrome patients with maternal uniparental disomy as a consequence of the combined effect of both the increased dosage of maternally expressed genes , and the loss of paternally expressed genes . We acknowledge certain limitations in our study . SNP arrays or array CGH can accurately determine extent and copy number of unbalanced chromosome regions ( S2 Fig shows a selection of duplications and triplications tested with Illumina SNP arrays ) , but give no information on the structural arrangement and position of the material . This can only be established by examining chromosome preparations , in conjunction with FISH probes . As we only had DNA material , this was not possible for this study . However , the presence of three copies is generally indicative of an interstitial duplication , whilst four copies ( triplication ) suggests the presence of a supernumerary idic15 . We excluded triplications from our study . Some rare idic15 cases can also represent only duplications and be indistinguishable from interstitial duplications by array testing . Therefore , while the vast majority of duplications in this study are likely to be interstitial , a small number could be supernumerary chromosomes , but still representing only duplications of the genetic material . As the main factor for pathogenicity is likely to be the copy number of the genes in the region , we feel that this limitation does not affect our conclusions . Interstitial triplications/idic15 are not the subject of our work , but we can report that there were no triplications in SZ subjects . This could be due to higher pathogenicity , leading to early onset neurodevelopmental phenotypes . Another limitation of our study concerns the role of paternal duplications in SZ , were we cannot completely exclude a small role , due to the wide confidence intervals for the prevalence and penetrance estimates . Finally , as discussed above , several unaffected controls had not had formal testing and although we can be confident that they do not suffer with severe neurodevelopmental disorders , we cannot exclude subtle phenotypes , and indeed have detected certain problems among the tested controls . This is not surprising , as a large study from Iceland has already shown that “healthy”carriers of pathogenic CNVs have lower cognitive performance [39] . In conclusion , our study clarifies the distinct roles of maternal and paternal interstitial duplications at 15q11 . 2-q13 . 3 in neuropsychiatric disorders , underlining the importance of maternally expressed imprinted genes in this interval to the incidence of psychotic illness . We also show that paternal duplications are pathogenic , increasing risk for DD/ASD/MCA with a penetrance of 20 . 7% . Defining the parent-of-origin of duplications at 15q11 . 2-q13 . 3 , which does not require parental DNA , may allow the refinement of genetic counselling and/or therapeutic intervention for individuals carrying these CNVs .
The CLOZUK study has UK National Research Ethics approval ( Ethics Committee WALES REC 2 , Study ID: 10/WSE02/15 ) . The CLOZUK samples were collected anonymously from across the UK ( thus without express consent ) , consistent with the UK Human Tissue Act and with the approval of the above ethics committee . We collated the available clinical and molecular data from large ( >400 cases ) , systematically ascertained CNV studies of SZ from the literature , or known to us via our collaborations . We similarly collected data on other neurodevelopmental disorders such as DD/ASD and multiple congenital anomalies ( MCA ) , focusing on two large studies based on referrals to genetic clinics and on three studies that specifically determined the parental origin of these duplications and reported on the presence of triplications/idic15 ( Table 1 ) . We used the reported information on parental origin of the 15q11 . 2-q13 . 3 duplications or , where possible , obtained DNA samples to establish this ( details in S1 Text ) . Most teams that we approached responded to our requests , but some had no access to patient DNA and could not take part . The control cohorts included 149 , 780 individuals ( Table 1 ) . The largest sample was from Iceland ( ~115 , 000 genotyped individuals at the time writing ) , a population-based cohort containing related individuals and those affected with medical and/or neuropsychiatric conditions . An ideal control cohort would only have one individual per family to avoid biases from over-sampling within population lineages . However , the Icelandic sample is designed to achieve total population coverage; therefore it is impractical to single out unrelated individuals . The rate of 15q11 . 2-q13 . 3 duplications in this population was assumed to be unbiased without excluding relatives of individuals with or without 15q11 . 2-q13 . 3 duplications . In fact , this population approaches the ideal general population ascertainment , that should give the best estimates for the various analyses performed here , therefore we decided to retain the two pairs of relatives who were carriers in this population . Penetrance was estimated according to the formula originally proposed by Vassos et al . [37] , updated for the joint effect of DD/ASD/MCA and for SZ , as we proposed earlier [26] . Briefly , we first estimated the rate of the duplications in the general population . We assumed that the control ( healthy ) population comprises 95% of the population , as explained in our previous work [11] . The rest of the population is made up of approximately 1% SZ and 4% DD/ASD/MCA cases . From this we estimated the proportion of 15q11 . 2-q13 . 3 duplication carriers in the general population that develop SZ or DD/ASD/MCA ( the penetrance ) . Although many duplication carriers might have both SZ and DD/ASD/MCA , we made the simplifying assumption that they would usually be ascertained only once . As the DD/ASD/MCA patients referred for genetic testing and the SZ patients recruited for studies might represent a more severely affected sub-group , we repeated the estimates after reducing by half these population rates ( down to 0 . 5% for SZ and 2% for DD/ASD/MCA ) and show these results in the S1 Text , S3 Table . Estimating 95% confidence intervals for the frequencies in each population followed the method we used in our previous paper [26] . Briefly , we first estimated the binomial CIs for the frequencies of CNVs in each population , including controls and the general population , using the Wilson score interval . Upper and lower 95% bounds for penetrance were estimated from the upper or lower bounds of CNV frequencies in patients and the lower or upper bounds of the frequencies in the general population . According to the mutation-selection balance theory , pathogenic mutations in the general population are found at low frequencies , where the addition of de novo mutations is balanced by the selection pressure against them ( q = μ/s ) , where q is the allele frequency in the general population , μ is the mutation frequency and s is the selection coefficient [11 , 26] . In order to estimate the selection coefficients ( s ) of 15q11 . 2-q13 . 3 duplications of maternal and paternal origin , we used the methods outlined in our previous work [11 , 26] as follows: The selection coefficient of a CNV can be approximated as the proportion of de novo CNVs out of the total number of CNVs ( de novo and inherited ) that are observed in an unbiased sample of CNV carriers . This is because the number of CNVs filtered out by natural selection is approximately equal to those introduced in the population as de novo mutations , after making the simplifying assumption that the frequency of the CNV does not change from generation to generation . We then estimated the mutation rate of the duplications , using the above formula . We should point out that the actual differences between maternal and paternal mutation rates are prone to error , e . g . due to the flipping of maternal and paternal duplications according to the gender of transmitting parents and possible differences in fecundity between affected male and female carriers ( Discussion ) . We established the parental origin of 15q11 . 2-q13 . 3 duplications for carriers from studies where it was previously unknown or not reported ( Table 1 and S1 Table ) . These were the studies/datasets by Gawlick , Alexic , the BBGRE database [18] ( https://bbgre . brc . iop . kcl . ac . uk/ ) , and new carriers in the Icelandic population and three published datasets [12 , 16 , 23] . Parental origin was established by one of several methods . For DNA available to the Cardiff laboratory , we used a methylation-sensitive high-resolution melt curve analysis , as described previously [8 , 19 , 38] , ( S1 Text and S1 Fig ) . The same method was used for the samples from Canada [7] . In the Icelandic sample parental origin was either determined with methylation-sensitive Southern analysis [5] or by long-range haplotype analysis ( S1 Text ) . Microsatellite analysis was used for the BBGRE dataset [18] . | The genetic interval 15q11 . 2-q13 . 3 on human chromosome 15 contains several so-called “imprinted genes” which are subject to epigenetic marking leading to activity from only one parental copy . This is in contrast to non-imprinted genes , whose activity is independent of their parent-of-origin . Deletions affecting the 15q11 . 2-q13 . 3 interval cause Prader-Willi and Angelman syndromes ( PWS/AS ) , depending on whether the deletions are paternally or maternally derived respectively . Duplications at the PWS/AS interval region may also lead to neurodevelopmental disorders , including developmental delay ( DD ) , autism spectrum disorder ( ASD ) and schizophrenia ( SZ ) . Due to presence of imprinted genes within the region , the parental origin of these duplications may be key to the pathogenicity . We show , for the first time , that paternal duplications lead to an increased risk of developing DD/ASD/multiple congenital anomalies ( MCA ) but , unlike maternal duplication , do not appear to increase risk for SZ . This study refines the distinct roles of maternal and paternal duplications at 15q11 . 2-q13 . 3 , underlining the critical importance of maternally active imprinted genes in the contribution to the incidence of psychotic illness . This work will have tangible benefits for patients with 15q11 . 2-q13 . 3 duplications by aiding genetic counseling . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"neuropsychiatric",
"disorders",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"adhd",
"prader-willi",
"syndrome",
"angelman",
"syndrome",
"social",
"sciences",
"developmental",
"psychology",
"neuroscience",
"disorders",
"of",
"imprinting",
"autism",
"spectrum",
"disorder",
"developmental",
"neuroscience",
"neurodevelopmental",
"disorders",
"gene",
"expression",
"schizophrenia",
"pathogenesis",
"mental",
"health",
"and",
"psychiatry",
"clinical",
"genetics",
"psychology",
"neurology",
"genetics",
"biology",
"and",
"life",
"sciences"
]
| 2016 | Parental Origin of Interstitial Duplications at 15q11.2-q13.3 in Schizophrenia and Neurodevelopmental Disorders |
Class switch recombination ( CSR ) requires activation-induced cytidine deaminase ( AID ) to trigger DNA double strand breaks ( DSBs ) at the immunoglobulin heavy chain ( IGH ) in B cells . Joining of AID-dependent DSBs within IGH facilitate CSR and effective humoral immunity , but ligation to DSBs in non-IGH chromosomes leads to chromosomal translocations . Thus , the mechanism by which AID-dependent DSBs are repaired requires careful examination . The random activity of AID in IGH leads to a spectrum of DSB structures . In this report , we investigated how DSB structure impacts end-joining leading to CSR and chromosomal translocations in human B cells , for which models of CSR are inefficient and not readily available . Using CRISPR/Cas9 to model AID-dependent DSBs in IGH and non-IGH genes , we found that DSBs with 5’ and 3’ overhangs led to increased processing during end-joining compared to blunt DSBs . We observed that 5’ overhangs were removed and 3’ overhangs were filled in at recombination junctions , suggesting that different subsets of enzymes are required for repair based on DSB polarity . Surprisingly , while Cas9-mediated switching preferentially utilized NHEJ regardless of DSB structure , A-EJ strongly preferred repairing blunt DSBs leading to translocations in the absence of NHEJ . We found that DSB polarity influenced frequency of Cas9-mediated switching and translocations more than overhang length . Lastly , recombination junctions from staggered DSBs exhibited templated insertions , suggesting iterative resection and filling in during repair . Our results demonstrate that DSB structure biases repair towards NHEJ or A-EJ to complete recombination leading to CSR and translocations , thus helping to elucidate the mechanism of genome rearrangements in human B cells .
DNA recombination at the immunoglobulin heavy chain ( IGH ) locus is required for class switch recombination ( CSR ) , the process that changes the class of immunoglobulin expressed by B cells , e . g . , from IgM to IgG or IgA , etc . CSR increases diversity of immunoglobulin isotypes , which is necessary for comprehensive humoral immunity . Recombination at the IGH locus is initiated by activation-induced cytidine deaminase ( AID ) [1] , which deaminates deoxycytidine to deoxyuridine mostly within WRC ( W = A/T , R = A/G ) motifs in IGH switch regions . These switch regions comprise 3–4 kilobases of non-coding DNA enriched for WRC motifs and are positioned upstream of each of the constant regions that encode immunoglobulin isotypes . Removal of deoxyuridines by the base-excision and mismatch repair pathways results in a DNA single strand break , or nick [2] . AID-dependent nicks on opposite DNA strands can melt into a DNA double strand break ( DSB ) . Ligation of DSBs in donor and acceptor switch regions places the acceptor constant region downstream of the rearranged VDJ gene segment . Expression of the recombined VDJ-acceptor constant region sequence gives rise to an immunoglobulin of a new isotype . Most AID-dependent DSBs at the IGH locus are either repaired faithfully with no effect on B cell receptor expression or in a productive manner leading to CSR . However , AID-dependent DSBs in the IGH locus can sometimes ligate to DSBs in other chromosomes , leading to chromosomal translocations that are hallmarks of B cell lymphomas [3–6] . Thus , characterizing the molecular mechanism of how AID-dependent DSBs are repaired within or between chromosomes will help explain the basis of antibody diversification and lymphomagenesis . DSB repair during CSR primarily relies on canonical non-homologous end-joining ( NHEJ ) [7] and , to a lesser extent , alternative end-joining ( A-EJ ) [8 , 9] . NHEJ and A-EJ are largely distinguished by the extent of mutagenicity associated with repair . Broadly speaking , NHEJ involves the ligation of blunt or nearly blunt DSB ends with little to no insertions/deletions ( indels ) . In contrast , A-EJ is characterized by exonuclease-driven deletions from the DSB end , known as resection , to expose short tracts of homologous nucleotides between non-homologous sequences , known as microhomology , which are used to join DSB ends [10] . Evidence of repair by A-EJ during CSR and other instances of DSB repair in mouse B cells has been best characterized when NHEJ is defective or suboptimal [9 , 11] . CSR is an attractive model for studying DSB repair and identifying NHEJ factors because it is a physiological , programmed process that is easily studied in mouse models . Primary mouse B cells undergo CSR ex vivo in response to lipopolysaccharide and cytokines while the CH12 mouse B cell line robustly undergoes CSR in vitro upon cytokine stimulation [12] . However , while overall CSR frequency is affected by mutation of putative NHEJ factors , the nature of the switch regions are not conducive to assessing the detailed role of these same factors . This is because AID can potentially act on thousands of deoxycytidines on both DNA strands in the switch regions , resulting in a mix of DSB structures , such as blunt ends or protruding 5’ or 3’ overhangs of varying length . Different factors or end-joining pathways might be needed to rejoin different types of ends . Furthermore , the high density of WRC motifs in the switch regions together with the seemingly random activity of AID make it difficult to determine the precise points where nicks and rejoining occurred . We have previously described an experimental system to work around these uncertainties . We used CRISPR/Cas9 and combinations of single guide RNAs ( sgRNAs ) to make predetermined DSBs similar to those that are expected from AID-mediated nicks to characterize DSB formation and repair in the CH12 mouse B cell line [13] . However , much less is understood about joining AID-dependent DSBs in human B cells , mainly due to limited experimental material . The CL-01 human B cell line can undergo CSR in vitro upon cytokine stimulation but at frequencies far below that of CH12 mouse B cells [14] . B cell subsets from human tonsils can also express AID and undergo CSR ex vivo upon cytokine stimulation [15] , but these uncloned cells are difficult to experimentally manipulate and are not suitable for detailed biochemical studies . In this study , we have induced class switching and chromosomal translocations in the BJAB human B cell line by creating predetermined DSB intermediates using CRISPR/Cas9 . These DSBs emulate DSB structures expected from physiological AID deamination , allowing us to assess how DSBs of known structure are ultimately repaired . Thus , we are able to analyze the mechanism of DSB repair leading to genome rearrangements in human B cells with the same precision as in mouse B cells . By comparing how different DSB structures are repaired by human B cells leading to switching or translocations , we found that DSB structure can bias repair of distal DSBs towards NHEJ and A-EJ in human B cells .
To characterize end-joining of distal , non-homologous loci in human B cells , we used the BJAB human B cell line . BJAB cells are Burkitt’s lymphoma-derived cells that express surface IgM but do not harbor the IGH-MYC translocation typical of Burkitt’s lymphoma [16] . These properties allowed us to target the native IGH locus in BJAB cells with Cas9 and induce de novo genetic rearrangements within chromosomes ( i . e . class switching from IgM to IgA ) and between chromosomes ( i . e . IGH-BCL6 chromosomal translocations ) . Throughout this study , we generated DSBs with varied end structures at AID target loci using Cas9 and single guide RNAs ( sgRNAs ) ( Fig 1A ) . We first tested whether we could induce Cas9-mediated switching in BJAB cells with blunt DSBs , as we had previously done in mouse B cell lines [13] . To do so , we transiently transfected BJAB cells with two plasmids encoding Cas9 and sgRNAs targeting either the region upstream of switch region μ ( S’μ ) or α ( S’α ) to create DSBs ( Fig 1A and 1B , S1 Table ) . We chose to target these regions in order to avoid the repetitive sequences present in the switch regions . Recombination between DSBs in S’μ and S’α replaces constant region μ ( Cμ ) with constant region α ( Cα ) downstream of the rearranged VDJ gene segment , allowing for surface IgA expression ( Fig 1B ) . Cells were harvested 5 days post-transfection for flow cytometric detection of surface IgA ( Fig 1C , S1 Fig ) . While a single blunt DSB in S’μ was not able to induce Cas9-mediated switching , blunt DSBs in both S’μ and S’α were able to induce Cas9-mediated switching in BJAB cells ( Fig 1C ) . Therefore , CRISPR/Cas9 is a suitable method of inducing DSBs at the native IGH locus leading to class switching . The majority of AID-dependent DSBs are staggered formed by nicks on opposite DNA strands [17] . Therefore , we next measured whether staggered DSBs in human B cells could also trigger Cas9-mediated switching . To produce staggered DSBs in BJAB cells , we used the Cas9 nickases , whereby a D10A or N863A substitution inactivate the Cas9 RuvC or HNH nuclease domains , respectively [18] . Expression of the Cas9 D10A nickase with a pair of sgRNAs produces a staggered DSB with a 5’ single-stranded DNA overhang ( hereafter termed a 5’DSB ) ( Fig 1A ) . Applying the same strategy with the same sgRNAs in combination with the Cas9 N863A nickase gives rise to a 3’DSB ( Fig 1A ) . Therefore , the position and overhang length of a particular DSB can be predefined using particular sgRNA combinations ( S2 Table ) . By transfecting BJAB cells with four plasmids encoding either the Cas9 D10A or N863A nickases and each of four sgRNAs , we produced 5’ and 3’ overhangs of 38 nucleotides in S’μ and 63 nucleotides in S’α ( Fig 1B ) . 5’DSBs with a 38 nt overhang in S’μ and a 63 nt overhang in S’α were able to induce Cas9-mediated switching at a frequency slightly lower than that of blunt DSBs , whereas 3’DSBs with identical overhang lengths were significantly poorer substrates ( Fig 1C , S1 Fig ) . These results suggest that 5’DSBs are better substrates for Cas9-mediated switching than 3’DSBs in human B cells ( see Discussion for alternative possibilities ) . To assess whether these staggered DSB overhang lengths resembled those triggered by AID in vivo , we determined how far apart AID-mediated nicks are typically spaced during CSR . To do so , we analyzed AID-mediated mutations at Sμ from Ung-/-Msh2-/- mice [2] . Mice that are doubly deficient in UNG and MSH2 are unable to excise AID-mediated uridines or repair dU:dG mismatches , respectively . Replication across these mismatches leads to mutations , >99% of which are transition mutations [2] . Thus , the distance between dC→dT and dG→dA transition mutations serves as a proxy for the distance between deaminated dCs on opposite strands . We found that the distance between dC→dT on opposite strands at Sμ from Ung-/-Msh2-/- mice ranged from 0 to 360 nt , with an average of 38 nt and a median of 27 nt ( S2 Fig , S3 Table ) . This result suggests that the lengths of staggered DSB overhangs we designed in the IGH locus ( 38 nt in S'μ and 63 nt in S'α ) are within the expected physiological range of AID-mediated staggered DSBs . It is important to note , however , that our analysis might underestimate the true distance between AID-mediated nicks because the transition mutations included in our analysis were unlikely to simultaneously arise or persist within a defined time-frame . At the same time , this analysis of AID-dependent mutations might overestimate the true distance between AID-mediate nicks as it was conducted 5’ of switch region μ , which contains less WRC motifs than the proper switch region μ . Therefore , although we cannot conclude with absolute certainty that the distance between transition mutations in the switch regions exactly represents dC deamination density , it seems that the overhang lengths we chose in this study are at least within what is caused by AID in mouse B cells . We next analyzed S’μ-S’α junctions from blunt , 5’ , and 3’DSBs to examine the footprint of end-joining pathways used for repair . To this end , we sequenced S’μ-S’α junctions from transfected cells and quantified the amount of resection and microhomology usage at S’μ-S’α junctions to discern repair by NHEJ and A-EJ . We found that blunt DSBs led to S’μ-S’α junctions with low amounts of resection ( Fig 1D , S1 Appendix ) and microhomology usage ( Fig 1E ) , consistent with previous findings that blunt DSBs are primarily substrates for NHEJ [19] . In contrast , 5’DSBs led to S’μ-S’α junctions with more resection ( Fig 1D ) and microhomology usage ( Fig 1E ) than blunt or 3’DSBs . Interestingly , we observed that the median amount of 5’DSB resection coincided with the sum length of S’μ and S’α overhangs , while the median amount of 3’DSB resection fell below this sum length ( Fig 1D , dotted line ) . This pattern was also evident by visualizing individual resection of S’μ versus S’α overhangs from each unique junction ( S3 Fig ) . This result suggests that 5’ overhangs tend to be removed prior to end-joining , while 3’ overhangs tend to be filled in . To determine whether the S’μ-S’α end-joining patterns we observed in BJAB cells were a general property of human B cells , we induced Cas9-mediated switching in Ramos cells , another human B cell line . Blunt and 5’DSBs led to a greater frequency of Cas9-mediated switching to IgA in Ramos cells than 3’DSBs ( Fig 1F ) . Like we observed in BJAB cells , 5’DSBs led to S’μ-S’α junctions in Ramos cells with increased resection compared to those from blunt or 3’DSBs ( Fig 1G ) . In Ramos cells , like BJAB cells , 5’DSBs tended to be resected to the sum total overhang length in S’μ and S’α , while resection of 3’DSBs fell below this sum total ( Fig 1G , dotted line ) , supporting the model that 5’DSBs are resected while 3’DSBs are filled in to complete end-joining . Interestingly , we observed that 5’DSBs led to increased microhomology usage at S’μ-S’α junctions compared to those from blunt DSBs but not 3’DSBs ( Fig 1H ) . Together , our results suggest that in human B cells , 5’ and 3’DSBs lead to more processing during end-joining than blunt DSBs to complete Cas9-mediated switching . Our results suggest that 5’DSBs lead to increased resection and increased microhomology typically associated with A-EJ , while 3’DSBs may be processed by a DNA polymerase prior to end-joining . Previous reports have suggested that sequence context can influence Cas9 repair outcomes [20 , 21] . One study found that sgRNAs that target the same sequence at multiple sites within the genome lead to similar Cas9-mediated indel profiles [20] . Another recent study found that a thymine in the fourth position after the protospacer adjacent motif ( PAM ) was more predictive of single nucleotide indels than a guanine in the same position [21] , suggesting that sequence might play a role in dictating end-joining characteristics . The mechanism behind this potential bias , however , remains unclear . To determine whether the S’μ-S’α end-joining patterns we observed were due to the nature of the sequence targeted by Cas9 rather than DSB polarity , we designed sgRNAs targeting a slightly different region upstream of S’α , approximately 150 nt downstream of the original sgRNA pair ( S4 Fig ) . Using this new pair of sgRNAs targeting S’α ( S2 Table ) , we generated blunt , 5’ , and 3’DSBs at S’μ and S’α to induce Cas9-mediated switching . Despite the different sequence targeted in S’α , we observed similar end-joining patterns in BJAB cells based on DSB structure . That is , 5’DSBs led to S’μ-S’α junctions with increased resection ( S4 Fig ) and microhomology ( S4 Fig ) than 3’DSBs . Again , we observed that the median amount of resection of 5’DSBs approached the sum total overhang length , while 3’DSBs tended to be resected below this amount ( S4 Fig ) . These results support the notion that DSB polarity triggers different modes of end processing . Although other studies have identified a role for sequence context in influencing DSB repair outcomes , we did not find evidence that simply changing sequence context significantly affected resection or microhomology at Cas9-mediated S’μ-S’α junctions . Rather , the polarity of a DSB within a given sequence had a more pronounced influence on end-joining properties . AID activity can potentially lead to chromosomal translocations between IGH and so-called “off-target” genes . To determine whether DSB structure affects inter-chromosomal joining in human B cells , we designed sgRNAs targeting S’μ on chromosome 14 and BCL6 on chromosome 3 ( Fig 2A ) . Recombination between Cas9-induced DSBs at these sites results in the IGH-BCL6 chromosomal translocation , hereafter termed “der3” , a marker of diffuse large B cell lymphoma and follicular lymphoma [22 , 23] . Cas9-mediated der3 was detected and quantified using a nested PCR assay . Briefly , approximately 45 , 000 genome equivalents ( 150 ng of human genomic DNA ) were used for round 1 amplification of der3 by conventional PCR followed by round 2 amplification by quantitative PCR ( qPCR ) . The amount of der3 in a given sample was normalized to the amount of a control genomic region on chromosome 13 amplified from total genomic DNA using a similar nested PCR assay . By transiently transfecting plasmids encoding the relevant Cas9 enzymes and sgRNA combinations ( S2 Table ) , we generated Cas9-induced der3 in BJAB cells with blunt , 5’ , and 3’DSBs ( S1 Fig ) . As we observed in Cas9-mediated switching , blunt and 5’DSBs were better substrates for Cas9-mediated der3 compared to 3’DSBs ( Fig 2B , S1 Fig ) . Even though 3’DSBs led to the lowest frequency of Cas9-mediated translocations , we were able to generate enough translocations from 3’DSBs for S’μ-BCL6 junction analysis with BJAB cells , which was previously not the case using CH12 mouse B cells [13] . To analyze the relative contributions of NHEJ and A-EJ to der3 formation , we sequenced S’μ-BCL6 junctions from BJAB cells . We found that the pattern of resection and microhomology usage at S’μ-BCL6 junctions was similar to what we observed at S’μ-S’α junctions . 5’DSBs led to S’μ-BCL6 junctions with more resection ( Fig 2C ) and microhomology usage ( Fig 2D ) than blunt DSBs or 3’DSBs . We found that the median amount of resection for 5’DSBs was equivalent to the sum total overhang length in S’μ and BCL6 , while the median amount of resection of 3’DSBs fell below the total overhang length ( Fig 2D , dotted line , S3 Fig ) . This finding mirrors our observation at S’μ-S’α junctions that 5’ overhangs are likely removed prior to repair while 3’DSBs are filled in ( Fig 1D , dotted line ) . Taken together , our data suggest that 5’ and 3’DSBs lead to more end processing than blunt DSBs leading to chromosomal translocations , with 5’DSBs leading to chromosomal translocation junctions exhibiting increased resection and microhomology usage . CSR is thought to be largely driven by non-homologous end-joining ( NHEJ ) in human B cells . Patients with mutations or deficiencies in the NHEJ factors DNA ligase 4 ( LIG4 ) and XRCC4-like factor ( XLF ) exhibit defects in CSR , as evidenced by low serum titers of IgG and IgA [24 , 25] . To confirm that Cas9-mediated switching recapitulated the requirement for NHEJ , we generated three LIG4-deficient BJAB clones using CRISPR/Cas9 gene editing ( S5 Fig ) . We found that LIG4 deficiency reduced Cas9-mediated switching to IgA via blunt , 5’ , and 3’DSBs ( Fig 3A–3C ) . This result suggests that Cas9-mediated switching , like AID-mediated CSR , primarily relies on NHEJ . We and others previously demonstrated that NHEJ deficiency leads to increased chromosomal translocations in mouse cells , suggesting that either NHEJ inhibits or A-EJ drives translocation in mouse cells [13 , 26–28] . Specifically , we showed that DNA ligase IV-deficiency led to a significant increase in Cas9-induced translocations from blunt and 5’DSBs in CH12 mouse B cells [13] . In contrast , it has been suggested that chromosomal translocations in human cells are NHEJ-dependent [29] . Ghezraoui et al illustrated that chromosomal translocations were significantly reduced in XRCC4- and LIG4-deficient human cells using 5’DSBs generated by zinc-finger nucleases , TALENs , and Cas9 D10A . The reason why A-EJ is more important for chromosomal translocations than CSR in mouse B cells remains unknown . To test whether our model system recapitulates this finding , we measured Cas9-mediated der3 in LIG4-deficient BJAB cells . We found that LIG4-/- BJAB cells had reduced Cas9-mediated der3 using either 5’ or 3’DSBs ( Fig 3E and 3F ) . We also confirmed that resection and microhomology at S’μ-BCL6 junctions from 5’DSBs were both increased in the absence of LIG4 , as expected of repair by A-EJ ( Fig 3G and 3H ) . Interestingly , we observed that Cas9-mediated der3 frequency was dramatically increased in LIG4-/- BJAB cells using blunt DSBs ( Fig 3D ) . The significance of this finding is considered in the Discussion . Taken together , our results suggest that the roles of NHEJ and A-EJ in rejoining chromosomal translocations in human B cells is dependent on DSB structure . These findings also support the notion that there are important differences in the molecular mechanism underlying chromosomal translocations between mouse and human B cells . DSB intermediates leading to CSR presumably arise from AID-mediated nicks on opposite DNA strands that simultaneously persist . Since the switch regions are WRC motif-rich and AID can theoretically trigger nicks at any deoxycytidine within WRC motifs , the length of staggered DSB overhangs likely ranges greatly from one recombination event to the next . We previously demonstrated that nicks spaced 248 nt apart could induce Cas9-mediated switching in mouse B cells if oriented to produce a 5’DSB , but not a 3’DSB [13] . It is unknown whether these distal nicks require exonucleolytic processing to form a DSB . We therefore tested whether increasing the distance between Cas9-mediated nicks , thereby producing 5’ and 3’DSBs with longer overhangs , affects the frequency and nature of Cas9-mediated switching . To address this , we used two pairs of sgRNAs targeting S’α to create 5’ and 3’DSBs with either 63 nt ( short ) or 121 nt ( long ) overhangs ( Fig 4A ) . The overhang length in S’μ was kept constant at 38 nt . We did not observe changes in Cas9-mediated switching frequency when S’α overhangs were short or long ( Fig 4B ) . Furthermore , 5’DSBs were better substrates for Cas9-mediated switching than 3’DSBs whether S’α overhangs were short or long ( Fig 4B ) . We next asked whether short or long overhangs affected resection and microhomology at S’μ-S’α junctions . We found that longer 5’ overhangs in S’α did not affect total resection at S’μ-S’α junctions , although longer 3’ overhangs in S’α did lead to a subtle increase in total resection ( Fig 4C , S3 Fig ) . Importantly , we still observed that 5’DSBs were resected more than 3’DSBs , supporting the notion that 5’DSBs are resected while 3’DSBs are filled in , even when overhang length is increased . Longer overhangs in S’α did not stimulate increased resection of S’μ ( Fig 4D ) , suggesting that ligation of staggered DSBs does not require symmetrical resection . We did not observe significant changes in microhomology at S’μ-S’α junctions whether 5’ or 3’DSB overhangs were short or long ( Fig 4E ) . Altogether , these findings suggest that the polarity of the DSB influences Cas9-mediated switching frequency more than overhang length . WRC motifs are abundant in the switch regions but less dense in “off-target” genes [30] . Accordingly , the frequency of AID-dependent mutations is higher in immunoglobulin loci than “off-target” genes [31] . Therefore , one would predict that the density of AID-mediated nicks in “off-target” genes to be lower , giving rise to staggered DSBs with longer overhangs . To test whether longer 5’ or 3’ overhangs in BCL6 affect the frequency and nature of Cas9-mediated der3 , we used two pairs of sgRNAs in BCL6 to generate 5’ and 3’DSBs with either a 57 or 169 nt overhang while keeping overhang length in S’μ constant ( Fig 5A ) . Like we observed in Cas9-mediated switching , increasing the overhang length in BCL6 did not significantly affect Cas9-mediated der3 frequency ( Fig 5B ) . 5’DSBs remained better substrates for Cas9-mediated der3 than 3’DSBs whether BCL6 overhangs were short or long ( Fig 5B ) , supporting the idea that DSB polarity influences suitability for recombination more than overhang length . 5’DSBs with longer overhangs in BCL6 were resected to the new sum total of S’μ and BCL6 overhangs , while 3’DSBs with longer overhangs in BCL6 were filled in ( Fig 5C , dotted versus solid line , S3 Fig ) . Increasing overhang length of staggered DSBs in BCL6 had no significant effect on resection of S’μ ( Fig 5D ) , supporting our hypothesis that resection of staggered DSBs in non-homologous partners is not typically symmetrical . As with Cas9-induced switching , we found no effect of overhang length on microhomology usage at Cas9-mediated der3 junctions ( Fig 5E ) . Together , these data suggest that DSB polarity plays a more significant role than overhang length in dictating the frequency and nature of genome rearrangements in human B cells . Our sequencing analysis of recombination events revealed inserted nucleotides at a subset of S’μ-S’α and S’μ-BCL6 junctions . Although the frequency of insertions at S’μ-S’α and S’μ-BCL6 junctions was similar regardless of DSB structure or polarity ( Fig 6A and 6C ) , staggered DSBs led to junctions with longer insertions than blunt DSBs ( Fig 6B and 6D ) . That is , insertions at most junctions from blunt DSBs were only a few nucleotides long , whereas insertions at junctions from 5’ and 3’ DSBs were almost as long as the overhang itself ( Fig 6B and 6D , dotted and solid lines ) . Indeed , these insertions were mostly composed of duplications of the S’μ , S’α , or BCL6 overhangs themselves ( Fig 6E and 6F ) . The presence of these “templated insertions” suggest that staggered DSBs undergo cycles of resection and filling in until repair is achieved . Our data suggest that the presence of a single-stranded DNA overhang may stimulate iterative repair more than a blunt DSB end .
In this study , we used CRISPR/Cas9 to generate site-specific , predefined DSB structures to model DNA lesions arising downstream of AID activity in human B cells . We found that DSB structure strongly influenced the mode of end processing associated with DNA repair leading to genome rearrangements like CSR or chromosomal translocations . The end processing profiles we observed at switch or translocation junctions were often , but not always , in alignment with generally accepted definitions of NHEJ and A-EJ . For example , we observed that 5’DSBs typically led to switch and translocation junctions exhibiting increased resection and increased microhomology , two characteristics of repair by A-EJ , compared to blunt or 3’DSBs . A previous study found that decreasing the density of nicks in the switch regions by knocking down AID , presumably leading to more staggered DSBs rather than blunt DSBs , increased microhomology usage at switch junctions [32] . We expand on this observation by showing that staggered DSBs specifically with 5’ overhangs skew towards increased microhomology usage at switch and translocation junctions . In addition , we previously found that 5’DSBs were preferentially repaired by A-EJ in CH12 mouse B cells [13] , suggesting that the influence of DSB polarity on end-joining pathway choice is likely a general property of mammalian cells . However , we also found that 3’DSBs undergo more end processing than blunt DSBs , although not in a manner characterized by resection . Specifically , we observed that while 5’ overhangs tended to be resected during end-joining , 3’ overhangs were filled in during end-joining . This result suggests that DSB polarity not only influences repair by NHEJ or A-EJ , but also how DSBs are processed and repaired by specific subsets of enzymes . It is not currently known which enzymes resect 5’DSBs or fill in 3’DSBs during end-joining . One recent study found that mutation of the 5’ to 3’ exonuclease CtIP , but not Mre11 or Exo1 , increased the fraction of precisely ligated DSBs with no resection , suggesting that CtIP is required for resection of DSBs [33] . A second study showed that either CtIP knockdown or chemical inhibition of Mre11 endo- and exonuclease activity led to a reduction in chromatin-bound RPA , a marker of single-stranded DNA , in G2 phase after ionizing radiation-induced DNA damage [34] . This result suggests that several candidate 5’ to 3’ exonucleases can resect DSBs in vivo , but could potentially act redundantly . Alternatively , 5’ overhangs could be removed by a flap endonuclease such as FEN1 [35] . We acknowledge that the manner ( s ) by which nucleotides were resected in our study , whether through exonucleolytic degradation or endonucleolytic degradation , differ mechanistically . Our model allows us only to quantify the number of nucleotides lost from the native sequence in a given switch or translocation junction , not the mechanism by which those nucleotides were lost . However , we can reasonably conclude that any resection beyond the sum total overhang length was likely achieved through exonucleolytic degradation . Regardless of how 5’ overhangs are removed , our data suggest that removal of the 5’ overhang might be required for end-joining , perhaps in order to expose microhomology to facilitate ligation . In contrast , filling in of 3’ overhangs during end-joining is poorly understood . One known example of physiological 3’DSBs in end-joining occurs during V ( D ) J recombination , whereby the opening of DNA hairpins at coding ends by the endonuclease Artemis results in 3’ overhangs [36] . These 3’ overhangs can be filled in , presumably by a DNA polymerase , giving rise to P nucleotides and junctional diversity within the B cell receptor repertoire . However , the mechanism for filling in 3’ overhangs leading to P-nucleotides has not yet been characterized . Identifying the enzymes that process 5’ and 3’DSBs leading to end-joining will help elucidate how the spectrum of AID-dependent DSBs are recognized , processed , and ligated to complete CSR or translocations . Our data support the model that different factors and end-joining pathways are indeed required to repair different DSB structures . One subset of enzymes that is known to recognize and repair AID-dependent DSBs comprise the NHEJ pathway . In this report , we demonstrate that NHEJ is indeed required for Cas9-mediated switching regardless of DSB polarity ( Fig 3A–3C ) . This finding supports the notion that NHEJ is the primary end-joining pathway leading to CSR in human B cells as it is in mouse B cells . Interestingly , we found that the role of NHEJ in chromosomal translocations was contingent on DSB polarity . Our data demonstrate that NHEJ deficiency led to a defect in Cas9-induced der3 using staggered DSBs , but a dramatic increase in der3 using blunt DSBs ( Fig 3D–3F ) . This finding is similar to a previous study showing that chromosomal translocations were reduced in NHEJ-deficient human cells using endonucleases that generated staggered DSBs , but were unchanged using blunt DSBs [29] . We had also found that ligase IV deficiency in CH12 mouse B cells led to a more pronounced increase in translocations from blunt DSBs than 5’DSBs [13] . Furthermore , we observed in this study that translocation junctions from blunt DSBs exhibited more microhomology and resection than switch junctions from blunt DSBs ( Figs 1D , 1E , 2C and 2D ) . Taken together , our results suggest that in the absence of NHEJ , A-EJ prefers joining blunt DSBs over staggered DSBs leading to chromosomal translocations in human B cells . We propose two explanations for this observation . First , perhaps the act of converting a blunt DSB into a staggered DSB represents a key signaling step during A-EJ . Blunt DSBs can indeed be repaired by A-EJ; indeed , we observe that in wild-type cells , blunt DSBs occasionally lead to recombination junctions with elevated resection and microhomology usage ( Figs 1D , 1E , 2D and 2E ) . Thus , blunt DSBs generated by Cas9 are sometimes resected into staggered DSBs when they are finally repaired . A-EJ in NHEJ-deficient cells might magnify this processing , which could stimulate chromosomal translocations . Still , it is unclear why this would not also apply to CSR , evoking the larger question of why A-EJ drives chromosomal translocations but not CSR in mouse B cells . Second , A-EJ is typically suppressed or overshadowed by NHEJ , so the preference for A-EJ to repair blunt DSBs leading to chromosomal translocations may be due to a lack of competition from NHEJ . Several studies have demonstrated that NHEJ suppresses A-EJ during CSR , as evidenced by an overall increase in microhomology at switch junctions from when NHEJ is depleted in mouse B cells [7 , 9] . This may be because binding of NHEJ factors to DSB ends inhibits resection , a step shared between homologous recombination and A-EJ , or because NHEJ simply occurs faster than A-EJ [37] . We observed that the majority of switch and translocations junctions from blunt DSBs in this study exhibited little to no microhomology , suggesting repair predominantly by NHEJ ( Figs 1E and 2E ) . In NHEJ-deficient cells , however , A-EJ is free to engage blunt DSBs that are typically repaired by NHEJ without competition for DSB ends , inhibition of resection , or time constraints . These atypical circumstances could lead to increased irregular end-joining such as chromosomal translocations . In this report , we observed that 5’DSBs were better substrates for Cas9-mediated switching and translocations than 3’DSBs ( Figs 1C and 2B ) . This pattern was independent of overhang length ( Figs 4B and 5B ) . However , the reason behind this bias is unknown . One possible explanation is that 5’ to 3’ resection is an important step in end-joining as it is for homologous recombination . For example , Cas9-induced 5’DSBs stimulate higher rates of homologous recombination than 3’DSBs using the DR-GFP homologous recombination reporter assay [38 , 39] . Another study found that Cas9-induced 5’DSBs led to more gene conversion , a pathway that relies on homologous recombination , than 3’DSBs [40] . In support of this hypothesis , we observed in this study that 5’ overhangs tended to be resected prior to end-joining while 3’ overhangs were filled in . Perhaps the lack of a 5’ overhang to resect makes 3’DSBs poor substrates for end-joining . Another possible explanation for the increased level of Cas9-mediated switching and der3 using 5’DSBs versus 3’DSBs could be due to different enzymatic activity of the Cas9 D10A and N863A nickases . That is , Cas9 D10A could simply lead to more DSBs than Cas9 N863A . Some studies have demonstrated that currently available RuvC-null Cas9 nickases ( e . g . D10A ) generate more indels than HNH-null Cas9 nickases ( e . g . N863A or H840A ) [41 , 42] . However , indel formation is likely an inaccurate measure of cleavage efficiency , as we and others demonstrate that DSB polarity itself can influence repair by NHEJ or A-EJ which may in turn lead to different frequencies of indel formation [13 , 40] . Nonetheless , Cas9 D10A and N863A nickases have been shown to have similar nicking efficiency in vitro [40] . An ideal study would use the same nuclease to produce 5’ and 3’DSBs , which in theory could be achieved using the Cas9 D10A nickase and orienting sgRNAs with PAMs facing outwards or inwards , respectively . In practice , however , paired sgRNAs with inward-facing PAMs are incompatible with Cas9 cleavage [40 , 41] . While we believe that the Cas9/Cas9 nickase system is the best currently available method to model AID-dependent DSB intermediates for CSR and chromosomal translocations , we acknowledge that Cas9 does not perfectly mimic AID in all aspects . For example , AID is a deaminase while Cas9 is a nuclease . Modeling AID-dependent DSBs with Cas9 therefore overlooks two intermediate enzymatic steps—deamination of dC by AID , and excision of dUs by UNG to create an abasic site , followed by the creation of a nick by APE—between AID deamination and nick formation . Future studies could use an enzymatically dead Cas9 fused to cytidine deaminase rather than the Cas9 nickase to more accurately model AID function [43] . Another limitation of our model system is that we targeted Cas9 cleavage 5’ of the switch regions rather than the repetitive switch regions themselves to avoid introducing multiple DSBs with one sgRNA . Although the loci 5’ of the switch regions are physiological sites of AID deamination [44] , our model does not take into account the formation of secondary structures in the repetitive switch regions that could be conducive to deoxycytidine deamination and DSBs [45] . Cas9 may also imperfectly model AID during CSR and chromosomal translocations in its residence time on DNA after cleavage and its influence on DNA repair processes . In conclusion , we found that DSB structure strongly influences repair by NHEJ and A-EJ in human B cells . Our data suggest that the spectrum of AID-dependent DSBs is not uniformly recognized and repaired by human B cells . These findings expand our knowledge of DSB repair in B cells leading to humoral immunity or lymphomagenesis and may eventually inform therapeutic interventions designed to promote desired DSB repair outcomes .
BJAB cells were cultured in RPMI 1640 medium with L-Glutamine supplemented with 10% fetal bovine serum , 5% NCTC 109 , 50 μM β-mercaptoethanol , and penicillin/streptomycin . Ramos cells were cultured in IMDM supplemented with 10% fetal bovine serum and penicillin/streptomycin . All cells were cultured at 37°C and 5% CO2 . sgRNAs were chosen based on favorable on- and off-target scores calculated by Benchling [46] . sgRNAs were cloned into plasmids encoding Cas9 ( Addgene: #42330 ) , Cas9D10A ( Addgene: #42335 ) , and Cas9N863A . Cas9N863A plasmid was constructed by site-directed mutagenesis of Addgene #42330 . One million BJAB cells in standard electroporation buffer were transfected with 4 μg of each relevant plasmid using a BioRad Gene Pulser using the following parameters: 350 V , 975 μF , ∞ Ω , and 4 mm gap . BJAB cells were incubated on ice for 5 minutes and then transferred into complete media . One million Ramos cells in IMDM were transfected using the following parameters: 250 V , 950 μF , ∞ Ω , and 4 mm gap . Ramos cells were incubated at room temperature for 5 minutes and then transferred into complete media . For chromosomal translocations , cells were harvested 5 days post-transfection and lysed to isolate genomic DNA by standard phenol-chloroform extraction and ethanol precipitation . For Cas9-induced switching , cells were harvested 5 days ( BJAB ) or 3 days ( Ramos ) post-transfection for genomic DNA isolation by proteinase K digestion and flow cytometry analysis of IgA expression ( see below ) . One million BJAB cells in standard electroporation buffer were transfected with 4 μg of sgRNAs LIG4 BRCTd G1 and G2 targeting DNA ligase 4 ( S1 Table ) . Cells were plated at 0 . 5 cells/well in a 96-well tissue culture plate 3 days post-transfection to obtain individual clones . Clones were screened for mutations by the mismatch cleavage assay as previously described [13] . Knockout clones were confirmed by sequencing . Der3 translocations were detected by nested PCR . Round 1 was amplified from 8 separate reactions using primers S’μ F . 1 and BCL6 R . 1 ( S1 Table ) and 150 ng of phenol chloroform-isolated genomic DNA by conventional PCR . Amplicons from the 8 reactions were then pooled and diluted ten-fold to use as template for Round 2 . Round 2 was amplified by qPCR using S’μ F . 2 and BCL6 R . 2 primers ( S1 Table ) and SsoAdvanced Universal SYBR Green Supermix ( BioRad ) . A control genomic region on chromosome 13 ( S1 Table ) was amplified from 10 ng of genomic DNA template by nested PCR to obtain reference Ct values , using round 1 primers c13 F . 1 and R . 1 followed by round 2 primers c13 F . 2 and R . 2 ( S1 Table ) . Der3 frequencies were calculated using the ΔΔCt method , using S’μ_1 + BCL6_1 WT Cas9 ( Fig 2B , second column ) or S’μ_1+2 + BCL6_1+2 Cas9 D10A ( Fig 5B , second column ) as control samples . Cells were harvested 5 days post-transfection , washed , and incubated with anti-IgA antibodies ( Southern Biotech , cat . 2050–09 ) for 30 min at 4°C in the dark . Cells were washed and analyzed for IgA expression on a BD LSR II at the Faculty of Medicine Flow Cytometry Facility ( University of Toronto , Toronto , Ontario , Canada ) . Der3 junctions for sequencing analysis were amplified by nested conventional PCR from genomic DNA using primers S’μ F . 1 and BCL6 R . 1 ( round 1 ) and S’μ F . 2 and BCL6 R . 2 ( round 2 ) ( S1 Table ) . S’μ-S’α junctions were amplified from genomic DNA by nested PCR using primers S’μ F . 1 and S’α R . 1 ( round 1 ) and S’μ F . 2 and S’α R . 2 ( round 2 ) ( S1 Table ) . Junction PCR products were purified using standard spin columns , cloned into a TA-cloning vector ( pGEM-T Easy , Promega ) , and transformed into chemically competent bacteria . Plasmids containing proper inserts were purified and Sanger sequenced at The Center for Applied Genomics ( The Hospital for Sick Children , Toronto , Ontario , Canada ) . Alignments of S’μ-BCL6 and S’μ-S’α junctions to reference sequences were performed using Basic Local Alignment Search Tool ( BLAST ) . The following reference sequences were used for analysis and are available from the National Center for Biotechnology Information ( NCBI ) : IGH ( NCBI Accession NG_001019 ) and BCL6 ( NCBI Accession NG_007149 . 1 ) . Resection is defined as the number of nucleotides in the native sequence spanning from 3 nt upstream of the PAM of the most distal sgRNA to the point where the junction changes alignment from one reference sequence to another , including any microhomologous nucleotides , that are not in the recombined sequence . Insertions are defined as the number of nucleotides in the recombined sequence that are not in the native sequence , starting from the end of alignment between the recombined and native sequences . All junctions included in sequencing data in this report are unique to genomic DNA harvested from a single transfection , and are reported in S1 Appendix . All statistical comparisons were made using the Mann-Whitney test . Stars denote p-values: ns>0 . 05 , *≤0 . 05 , **≤0 . 01 , ***≤0 . 001 , ****≤0 . 0001 . | The production of different classes of antibodies/immunoglobulins ( IgM , IgG , etc . ) is essential for protection against diverse pathogens and effective immunity . This cellular process is triggered by the enzyme activation-induced cytidine deaminase ( AID ) . AID mutates DNA predominantly in antibody genes , generating different types of DNA breaks . Repair of DNA breaks initiated by AID leads to the production of different antibody classes . Erroneous repair of this damage can also lead to chromosomal translocations , a hallmark of lymphomas and other cancers . In this study , we used CRISPR/Cas9 technology to model the different types of DNA breaks physiologically produced by AID . We found that the specific structure of these DNA breaks strongly influenced how they were repaired . That is , different types of DNA breaks inform different modes of rejoining . Our findings show that not all types of DNA breaks are treated equally by genome maintenance machinery in the cell . These observations provide insight into the molecular mechanisms behind antibody-dependent immunity and lymphomagenesis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immunology",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"sequence",
"motif",
"analysis",
"dna",
"mammalian",
"genomics",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"white",
"blood",
"cells",
"genomics",
"sequence",
"alignment",
"animal",
"cells",
"chromosome",
"biology",
"chromosomal",
"aberrations",
"bioinformatics",
"animal",
"genomics",
"genetic",
"loci",
"surgical",
"resection",
"antibody-producing",
"cells",
"biochemistry",
"chromosomal",
"translocations",
"cell",
"biology",
"b",
"cells",
"nucleic",
"acids",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"dna",
"repair",
"non-homologous",
"end",
"joining"
]
| 2019 | DSB structure impacts DNA recombination leading to class switching and chromosomal translocations in human B cells |
Mali has become increasingly interested in the evaluation of transmission of both Wuchereria bancrofti and Onchocerca volvulus as prevalences of both infections move toward their respective elimination targets . The SD Bioline Onchocerciasis/LF IgG4 Rapid Test was used in 2 evaluation units ( EU ) to assess its performance as an integrated surveillance tool for elimination of lymphatic filariasis ( LF ) and onchocerciasis . A cross sectional survey with SD Bioline Onchocerciasis/LF IgG4 Rapid Test was piggy-backed onto a transmission assessment survey ( TAS ) ( using the immunochromatographic card test ( ICT ) Binax Filariasis Now test for filarial adult circulating antigen ( CFA ) detection ) for LF in Mali among 6–7 year old children in 2016 as part of the TAS in two EUs namely Kadiolo-Kolondieba in the region of Sikasso and Bafoulabe -Kita-Oussoubidiagna-Yelimane in the region of Kayes . In the EU of Kadiolo- Kolondieba , of the 1 , 625 children tested , the overall prevalence of W . bancrofti CFA was 0 . 62% ( 10/1 , 625 ) [CI = 0 . 31–1 . 09]; while that of IgG4 to Wb123 was 0 . 19% ( 3/1 , 600 ) [CI = 0 . 04–0 . 50] . The number of positives tested with the two tests were statistically comparable ( p = 0 . 09 ) . In the EU of Bafoulabe-Kita-Oussoubidiagna-Yelimane , an overall prevalence of W . bancrofti CFA was 0% ( 0/1 , 700 ) and that of Wb123 IgG4 antibody was 0 . 06% ( 1/1 , 700 ) , with no statistically significant difference between the two rates ( p = 0 . 99 ) . In the EU of Kadiolo- Kolondieba , the prevalence of Ov16-specific IgG4 was 0 . 19% ( 3/1 , 600 ) [CI = 0 . 04–0 . 50] . All 3 positives were in the previously O . volvulus-hyperendemic district of Kolondieba . In the EU of Bafoulabe-Kita-Oussoubidiagna-Yelimane , an overall prevalence of Ov16-specific IgG4 was 0 . 18% ( 3/1 , 700 ) [CI = 0 . 04–0 . 47] . These 3 Ov16 IgG4 positives were from previously O . volvulus-mesoendemic district of Kita . The SD Bioline Onchocerciasis/LF IgG4 Rapid test appears to be a good tool for integrated exposure measures of LF and onchocerciasis in co-endemic areas .
Approaches to onchocerciasis control and lymphatic filariasis ( LF ) elimination have proceeded along parallel but independent courses in Mali . For LF , annual mass drug administration ( MDA ) with albendazole and ivermectin began in 2004 and continued for 5 to 9 consecutive MDA campaigns . Onchocerciasis was originally highly endemic in five regions of Mali ( Kayes , Koulikoro , Sikasso , Segou and Mopti [1]; the eastern most regions ( Koulikoro rive droite , Sikasso , Ségou and Mopti ) were included in the original onchocerciasis control programme ( OCP ) based on vector control with larvicides [2] . The western parts of the endemic regions ( Kayes and Koulikoro rive gauche ) were part of the western extension of OCP in which ivermectin administration was used in MDA compaigns and later community directed therapeutic intervention ( CDTi ) under the umbrella of African Programme for Onchocerciasis Control ( APOC ) using community drug distributors ( CDDs ) . LF and onchocerciasis are overlapping in many districts of Mali [1] . Both the LF and onchocerciasis control programs were implemented based on precontrol mapping data for each of diseases [1] . For meso- and hyper-endemic onchocerciasis , the use of skin snips and eye examination for mapping demonstrated that 5 of the 8 administrative regions of Mali were endemic for O . volvulus [1 , 3]; whereas all of the 8 administrative units ( encompassing 75 health districts ) were endemic for LF based on a 2004 survey using the immunochromatographic test ( ICT ) known commercially as the Binax Filariasis Now test ( Alere , Portland , ME ) where circulating filarial antigen ( CFA ) prevalences were greater than 1% [1 , 4] . In 2005 , the results of longitudinal studies from 3 formerly O . volvulus hyperendemic foci in Mali and Senegal provided evidence that onchocerciasis elimination could be achieved based only on mass drug adminstration of ivermectin [5] . The evaluations used skin snips for the detection of microfilaridermia and blackfly dissection [5 , 6] . At the same time , when LF was mapped and reported to be endemic throughout the country , all the districts in Mali were treated using MDA ( ivermectin with albendazole ) for LF . LF transmission assessment surveys ( TAS ) as recommended by World Health Organization ( WHO ) were initiated in 2012 and are currently being performed in 22 evaluation units ( EU ) . An EU includes one or several endemic districts based on geographic location , treatment coverage and population size [7] . The Binax Filariasis Now ICT cards and more recently the Filariasis Test Strip ( FTS ) [8] have been used for LF mapping and for the TAS , but these tests may have their limitations because of their slow kinetics of disappearance and their potential cross-reaction in cases of Loa loa infection , a filarial parasite that is absent in Mali [9] . For onchocericasis , post-treatment surveillance based on positivity in children with Ov16-based immunoassays is the current gold standard , but the challenge remains in the definition of prevalence cutoffs using the various forms of the Ov16 ELISA [10] or the SD Bioline’s Ov16-containing RDTs [11] . Initially , O . volvulus infection mapping in Mali was conducted using skin snip and eye examination that left many hypo-endemic areas excluded from the various control programs and from further consideration for CDTi . As a consequence of “redistricting” in 2016 , the number of onchocerciasis-endemic districts increased from 17 to 34 . Among these 34 , only 2 have stopped CDTi . Hence , re-mapping is needed in many potentially O . volvulus -endemic areas . Mass drug adminstration of ivermectin is still ongoing in 20 districts , and 12 are under surveillance . These 12 districts ( under surveillance for onchocerciasis ) had previously been part of the OCP vector control program; however , they have received albendazole and ivermectin for LF for at least 5 MDA rounds . Beyond these 34 districts , there may be a need for re-mapping potentially O . volvulus-hypo-endemic regions if elimination goals are to be achieved in the near future [12] . In this study , the SD Bioline Onchocerciasis/LF IgG4 Rapid Test was used concurrently with the ICT in two LF evaluation units ( EU ) in Mali consisting of districts with different endemicity levels for LF and onchocerciasis prior to MDA/CDTi to assess its performance as an integrated surveillance tool for LF and onchocerciasis elimination .
As this work was conducted as part of the National Neglected Tropical Diseases control program activity , it was deemed exempt from IRB approval by the ethical committee of the University of Science , Techniques and Technologies of Bamako . However , a protocol related to the evalution of SD Bioline Onchocerciasis/LF IgG4 Rapid Test in Mali was approved by the ethical committee ( 2017/199/CE/FMPOS ) . Participation in this study was entirely voluntary . The study was clearly explained to the community leaders and health authorities and their permission obtained before any activities were undertaken . Oral consent was obtained from the legal guardians of all the 6- to 7-year old children due to low literacy level in communities . In the Sikasso region , the study was conducted in the EU of the district of Kadiolo and Kolondièba . In this EU , the district of Kadiolo had been endemic for W . bancrofti ( 24 . 4% pre-control ) but hypoendemic for O . volvulus ( 20% pre-control ) , while the district of Kolondieba had been co-endemic for these two parasites ( 37% and 60% pre-control prevalences for LF and onchocerciasis respectively ) [1 , 13] ( Table 1 ) . In Kayes region , the entire EU of the district of Bafoulabe , Kita , Oussoubidiagna and Yelimane was known to be endemic for LF ( with 9 . 6% pre-control prevalence in each district ) , but O . volvulus was found to be endemic in certain districts with pre-control prevalences as follows: Kita ( mesoendemic-40% ) , Bafoulabe ( mesoendemic—42% ) , Oussoubidiagna ( hyperendemic-60% ) , Yelimane ( hypoendemic—33% ) ( Table 1 ) . In all these districts , 2016 was the last year of ivermectin and albendazole distribution for LF although mass drug adminstration of ivermectin continues for onchocerciasis [1 , 13] . The present study was piggy-backed onto TAS surveys ( using the Binax Filariasis Now test for filarial adult circulating antigen detection ) for LF across 2 EUs in Mali to demonstrate the utility of the SD Bioline Onchocerciasis/LF IgG4 Rapid Test for integrated assessment of W . bancrofti and O . volvulus transmission [14] . The sample size builder ( SSB ) was used to automate the calculations for determining appropriate survey strategy and sample size calculations based on TAS sampling strategy . The design of the surveys is flexible in order to best fit the local situation and depends upon factors such as the primary school enrolment rate , the populations size , the number of schools or enumeration areas , and the cost of different survey methods [7] . For this current study community based survey was conducted due to low school enrollment rate . The number of villages and the number of households included in the study were determined by the SSB [7] . Children of 6 and 7 year old within the randomly selected households made the sample . The minimum size required for the TAS were 1 , 556 and 1 , 692 children 6–7 year old ( according to the SSB ) respectively for EUs of Kadiolo-Kolondieba and Bafoulabe -Kita-Oussoubidiagna-Yelimane . The estimated total population of 6 to 7 year old children were 37 , 620 and 71 , 152 respectively in the EU of Kadiolo- Kolondieba and Bafoulabe-Kita-Oussoubidiagna-Yelimane ( Table 2 ) . A multistage sampling technique was used for the sampling . In each EU , the villages and the backup villages were randomly selected using the SSB tool . The backup villages were chosen in addition to 30 villages that made the cluster for an EU . In the one case where the selected village was inaccessible or refused to participate in the survey , a back up village was used to replace it . At the village level , the list of the households was made available , and two tables of random numbers generated from computer using the SSB were also used to select randomly the households to be included in the study . In each selected household , all the children 6–7 years of age were included in the study . A household was defined as a group of persons living in the same house and sharing the same food . Children aged 6-and 7-year-old were tested using Binax Filariasis Now test and retested ~ 6 months later with the SD Bioline Onchocerciasis/LF IgG4 Rapid Test in the EU of Kadiolo–Kolondieba at the point of care ( POC ) ( Table 2 ) . In the EU of Bafoulabe -Kita-Oussoubidiagna-Yelimane , dried blood spots ( DBS ) were collected and cryopreserved at the time of the TAS as the SD Bioline Onchocerciasis/LF IgG4 Rapid Test were not available . All tests for CFA were performed at the POC using the Binax Filariasis Now test ( Alere , Portland , ME ) , using the manufacturer’s instructions . In the EU of Kadiolo–Kolondieba , the SD Bioline Onchocerciasis/LF IgG4 Rapid Test was used at the POC again exactly as recommended by the manufacturer ( SD Diagnostics , Korea ) . For the dried blood spots used for antibody assessments in the EU of Bafoulabe-Kita-Oussoubidiagna-Yelimane , the blood was collected on filter paper spots ( TropBio , Townsville , Australia ) and cryopreserved in liquid nitrogen dry shippers in the field before being transported to the Filariasis Research Unit laboratory in Bamako the capital city of Mali . The DBS were then stored at– 80°C prior to elution as described in Kamgno et al , [15] and tested using the SD Bioline Onchocerciasis/LF IgG4 Rapid Test . The SD Bioline Onchocerciasis/LF IgG4 Rapid Test measures simultaneously the presence of IgG4 antibodies to Wb123 and Ov16 . All data analyses were performed using SPSS Version 24 ( Statistical package for Social Sciences ) and used the 5% level of significance . The Fisher's exact test was used when appropriate and the Clopper-Pearson 95% confidence interval around the prevalences were used for statistical comparisons . The geographic coordinates were measured for each village visited during the survey using mobile phones and thereafter used to map the ICT positive individuals as well as those positive for IgG4 antibodies against Wb123 and Ov16 .
In the EU of Kadiolo- Kolondieba , an overall prevalence of W . bancrofti infection based on CFA was found to be 0 . 62% [95% CI = 0 . 31–1 . 09] . When assessed at the district level , the CFA prevalence was 0 . 69% [95% CI = 0 . 21–1 . 65] in Kadiolo and 0 . 57% [95% CI = 0 . 23–1 . 19] in Kolondieba . In the same EU , the overall prevalence of Wb123 IgG4 was 0 . 19% [95% CI = 0 . 04–0 . 50]; when assessed per district , the prevalence of IgG4 to Wb123 was 0 . 20% [95% CI = 0 . 0–0 . 97] in Kadiolo and 0 . 18% [95% CI = 0 . 03–0 . 60] in Kolondieba ( Table 4 ) . The prevalences obtained using either of the two tests were statistically comparable ( p = 0 . 99 ) . In the EU of Bafoulabe-Kita-Oussoubidiagna-Yelimane , an overall prevalence of W . bancrofti infection based on CFA in chidren was 0% ( 0/1 , 700 ) . For the Wb123 IgG4 antibody , only 1/1 , 700 ( 0 . 06% [95% CI = 0 . 0–0 . 28] ) was found . This single positive was reported in the district of Bafoulabe; thus the local prevalence of Wb123 IgG4 was 0 . 21% [95% CI = 0 . 01–1 . 06] ( Table 4 ) . There were no differences between the CFA positive cases and those positive for Wb123 ( p = 0 . 99 ) . In the EU of Kadiolo–Kolondieba , 3/1 , 600 children were positive for Ov16-specific IgG4 with an overall antibody prevalence of 0 . 19% [95% CI = 0 . 04–0 . 50] being found . All 3 positives were in the district of Kolondieba , leading to a local prevalence of 0 . 27% [95% CI = 0 . 06–0 . 74] in this previously O . volvulus-hyperendemic district ( Table 5 ) . It should be noted that 0/502 children in Kadiolo ( an O . volvulus-hypoendemic region ) were IgG4 positive to Ov16 , but the upper 95% confidence level around this 0% prevalence was 0 . 59% . There was no statistical difference between the number positive ( 0/502 ) in Kadiolo , an O . volvulus-hypoendemic district and Kolondieba a previously O . volvulus-hyperendemic district ( 3/1 , 095 ) ( p = 0 . 55 ) . In the EU of Bafoulabe-Kita-Oussoubidiagna-Yelimane , an overall prevalence of Ov16-specific IgG4 of 0 . 18% [95% CI = 0 . 04–0 . 47] was recorded . These 3 individual Ov16 IgG4 positives were in the district of Kita among the 713 children tested , giving a prevalence of 0 . 42% for this district [95% CI = 0 . 10–1 . 14] ( Table 5 ) . Again it should be noted that in the O . volvulus-hypoendemic district ( Yelimane ) , although there were no Ov16 positive children , the upper limit of the 95% CI exceeded 1% , suggesting that cut off point of antibody positivity for children should be raised above the 0 . 1% threshold . The comparison of positive cases of Ov16 between the previous O . volvulus-hypoendemic district ( Yelimane ) and the previous O . volvulus-mesoendemic district with 3 positives cases was similar ( p = 0 . 99 ) . CFA positive children were found primarily in the EU of Kadiolo-Kolondieba with 10 CFA positives detected using Binax Filariasis Now test in five different villages ( Fig 1 ) . No CFA carriers were found in the 31 villages of the Bafoulabe-Kita-Oussoubidiagna-Yelimane EU , but one child was found positive for Wb123 IgG4 in the district of Bafoulabe ( Fig 1 ) . In the EU of Kadiolo Kolondieba , 3 positive Wb123 IgG4 children were observed in three different villages . In the same EU , 3 Ov16 IgG4 positive children were found in 3 different villages ( Fig 1 ) . In the EU of Bafoulabe-Kita-Oussoubidiagna-Yelimane , 3 Ov16 IgG4 positive cases were detected in a single village of Kita district previously mesoendemic at the pre-control ( Fig 1 ) . Overall , single Ov16 or Wb123 IgG4-positive children were spread across 5 different villages of the 30 villages screened in Kadiolo–Kolondieba EU ( Fig 1 ) .
Mali has become increasingly interested in the evaluation of transmission of both O . volvulus and W . bancrofti as prevalences of both infections move toward their respective elimination targets . This reduction is due to the fact that the national lymphatic filariasis elimination program , since its creation , has distributed the required number of mass treatment campaigns with at least 5 consecutive rounds of albendazole and ivermectin treatment; thus TAS is ongoing in the different EUs [1] . With regard to onchocerciasis , after periods of vector control , and ivermectin mass drug administration lasting more than 24 years in previously hyper- and meso-endemic foci , the program needs to fulfill the criteria for the cessation of ivermectin distribution [1] . Additionally , elimination goals likely were reached in most endemic foci by 2006 [6] . In 2016 , as part of the TAS for LF using the Binax Filariasis Now test by the national LF elimination program in two EUs , namely Kadiolo-Kolondieba in the region of Sikasso and Bafoulabe-Kita-Oussoubidiagna-Yelimane in the region of Kayes , the SD Bioline Onchocerciasis/LF IgG4 Rapid Test was used to assess its performance as an integrated surveillance tool [14] to inform each of the two national elimination programs involved in LF and onchocericiasis elimination respectively . For the assessment of W . bancrofti transmission , in the EU of Kadiolo- Kolondieba , W . bancrofti circulating filarial antigen ( CFA ) prevalence was 0 . 62% [95% CI = 0 . 31–1 . 09] while the seroprevalence ( antibodies ) targeting the same parasite was 0 . 19% [95% CI = 0 . 04–0 . 50] . In the EU of Bafoulabe-Kita-Oussoubidiagna-Yelimane , the CFA prevalence was 0% [95% CI = 0 . 0–0 . 17] while the antibody seroprevalence was 0 . 06% [95% CI = 0 . 0–0 . 28] . The prevalences of LF infection measured using these two tests were comparable and with all the tests , the upper bounds did not reach the cutoff point of 2% ( Table 4 ) . Recent studies suggested that the Binax Filariasis Now test may overestimate the W . bancrofti infection prevalence [16 , 17] . Given that the specificity of the tests for CFA may be lower than previously thought in Loa-endemic areas [18] , given that antibodies to Wb123 appear earlier than CFA in longitudinally followed children [19] and given the excellent superimposition of W . bancrofti prevalence results using either antibody ( Wb123 ) or antigen ( CFA ) testing , we suggest that the measurement of IgG4 to Wb123 might be preferable in decisions to stop MDA or in TAS following cessation of MDA as it allows the measurement of recent exposure . The challenge remains , however , to determine an acceptable prevalence threshold that would guarantee interruption of W . bancrofti transmission particularly since vector monitoring is not part of most LF elimination programs . In the Gambia , the recent use of the Wb123 IgG4 test suggests that this serological tool could be used to decide to stop MDA [20] . Therefore , maybe the SD Bioline Onchocerciasis/LF IgG4 Rapid Test could be utilized for transmission assessment and as a decision tool for MDA cessation , should modelling studies help to corroborate these empiric findings . Despite the application of the LF strategy for sampling of O . volvulus ( not the recommended target age group less than 10 year old and sample size of 3 , 000 children ) assessment , the level of exposure to O . volvulus , as determined by the detection of the IgG4 antibodies to Ov16 , the current prevalence estimates overlaid quite nicely with previous epidemiological profiles of the different study districts ( Tables 1 , 3 and 5 ) . All Ov16 antibody positive subjects were reported in the previously known onchocerciasis meso- and/or hyperendemic districts [6] . Our results show that O . volvulus transmission level was above the current elimination threshold ( 0 . 1% ) among the 6–7 year old , though the utility of this 0 . 1% target as a threshold is currently under debate [11 , 21] . The results of this study suggest that many of the hypoendemic areas are not in need of ivermectin distribution [22] . However , the sample size used in our study was calculated for LF transmission assessment and not for the evaluation of O . volvulus transmission ( typically 3 , 000 children per transmission focus ) [21] . Within the context of onchocerciasis elimination , new data are needed to re-categorize O . volvulus transmission potentials in subregions of countries such as Mali where many areas have undergone more than 24 years of CDTi with no or few epidemiological assessments [1] . Moreover , entomologic ( blackfly ) data are needed to confirm our findings that onchocerciasis transmission has been interrupted in these previously endemic district before to decide about CDTi cessation as suggested in a previous study [11] . The serological profiles reported here are similar to what was observed in southern Mexico where a decision to interrupt CDTi was taken [23] . Overall , the SD Bioline Onchocerciasis/LF IgG4 Rapid Test Wb123/Ov16 test , has 84% sensitivity and 98–99% specificity [24] . As no antibody-based test is likely to have specificities of >99% ( meaning 1% false positive rates ) the current threshold of elimination of onchocerciasis based on Ov16-specific antibodies ( by whatever test available ) must be increased and also results must be underscored by transmission assessment in pooled blackfly populations [25] . Based on the present study using the SD Bioline Onchocerciasis/LF IgG4 Rapid Test in previously LF- and onchocerciasis co-endemic regions in Mali , this POC test appears to be a good tool for integrated measurement of exposure to W . bancrofti and O . volvulus in children , with the potential to be used as a surrogate for interruption of transmission in a given focus as well as in post-MDA impact mesures in local populations in countries such as Mali . These data suggest that there is likely little or no transmission of either O . volvulus or W . bancrofti . Moreover , given the upper limits of the CI for Ov16 IgG4 in an O . volvulus-hypo-endemic region ( and the fact that it does not differ from the hyper- and meso-endemic regions ) a 0 . 1% threshold based on Ov-16-based IgG4 immunoassays may certainly be too conservative . Obviously , additional data underlying the utility of this POC test will be necessary if the elimination goals for both onchocerciasis and LF are to be achieved . | After years of mass drug administration for lymphatic filariasis and onchocerciasis , many co-endemic countries need to demonstrate the interruption of transmission of these parasitic infections . This interruption of transmission is the key to stopping mass drug administration . One of the major challenges facing national programs is a reliable point of care diagnostic tool capable of detecting recent infection in children . To this end , the SD Bioline Onchocerciasis / LF IgG4 Rapid Test was used to evaluate the transmission interruption of lymphatic filariasis and onchocerciasis in previously endemic regions of Mali . This rapid diagnostic tool demonstrated good performance in 6 of districts in 2 “so-called’ evaluation units . There was very good overlap of the serological results obtained by the rapid diagnostic test ( RDT ) and previous endemicity profile of the different study districts for the two infections . We would argue that this RDT could be a single integrated approach for assessing transmission in areas where lymphatic filariasis and onchocerciasis are/were co-endemic . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"mali",
"onchocerca",
"volvulus",
"immune",
"physiology",
"helminths",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"parasitic",
"diseases",
"animals",
"onchocerca",
"filariasis",
"pharmaceutics",
"drug",
"administration",
"neglected",
"tropical",
"diseases",
"onchocerciasis",
"antibodies",
"lymphatic",
"filariasis",
"africa",
"immune",
"system",
"proteins",
"wuchereria",
"bancrofti",
"proteins",
"wuchereria",
"people",
"and",
"places",
"biochemistry",
"helminth",
"infections",
"eukaryota",
"physiology",
"nematoda",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy",
"organisms"
]
| 2019 | Integrated seroprevalence-based assessment of Wuchereria bancrofti and Onchocerca volvulus in two lymphatic filariasis evaluation units of Mali with the SD Bioline Onchocerciasis/LF IgG4 Rapid Test |
Neutrophil extracellular traps ( NETs ) comprise an ejected lattice of chromatin enmeshed with granular and nuclear proteins that are capable of capturing and killing microbial invaders . Although widely employed to combat infection , the antimicrobial mechanism of NETs remains enigmatic . Efforts to elucidate the bactericidal component of NETs have focused on the role of NET-bound proteins including histones , calprotectin and cathepsin G protease; however , exogenous and microbial derived deoxyribonuclease ( DNase ) remains the most potent inhibitor of NET function . DNA possesses a rapid bactericidal activity due to its ability to sequester surface bound cations , disrupt membrane integrity and lyse bacterial cells . Here we demonstrate that direct contact and the phosphodiester backbone are required for the cation chelating , antimicrobial property of DNA . By treating NETs with excess cations or phosphatase enzyme , the antimicrobial activity of NETs is neutralized , but NET structure , including the localization and function of NET-bound proteins , is maintained . Using intravital microscopy , we visualized NET-like structures in the skin of a mouse during infection with Pseudomonas aeruginosa . Relative to other bacteria , P . aeruginosa is a weak inducer of NETosis and is more resistant to NETs . During NET exposure , we demonstrate that P . aeruginosa responds by inducing the expression of surface modifications to defend against DNA-induced membrane destabilization and NET-mediated killing . Further , we show induction of this bacterial response to NETs is largely due to the bacterial detection of DNA . Therefore , we conclude that the DNA backbone contributes both to the antibacterial nature of NETs and as a signal perceived by microbes to elicit host-resistance strategies .
Neutrophils are central mediators of the innate immune defense system and perform their role by killing invading microbes through phagocytosis , degranulation , and the release of neutrophil extracellular traps ( NETs ) [1 , 2] . The scaffold of NETs is composed of genomic DNA , which is enmeshed with antimicrobial proteins normally found in the nucleus , granules , or cytoplasm of neutrophils [1 , 2] . Although largely characterized ex vivo using purified human neutrophils and the chemical inducer phorbol-12-myristate 13-acetate ( PMA ) , the process of NETosis resulting in the generation of NETs has been observed in vivo during infection where these structures function to trap bacteria , fungi , protozoa and viruses [1–5] . The mechanism by which NETs kill microbial invaders remains controversial [5–7] . Given the detection of known antimicrobial proteins that decorate the genomic lattice structure [1 , 2 , 8 , 9] , current models describing the antimicrobial function of NETs focus on the role of NET-bound proteins . However , most of these proteins are present in low abundance and evidence of their antimicrobial function while bound to the NET structure is limited to a few proteins [1 , 8–11] . NET-bound calprotectin is a zinc-chelating protein with antimicrobial activity against Candida and Klebsiella that can be neutralized with excess zinc [3 , 9 , 10] . Histones , the most abundant NET-bound proteins [10] , possess direct membrane-acting antibacterial activity [12 , 13] and were shown to contribute to killing of Staphylococcus and Shigella [1] . Cathepsin G , a granular serine protease , is required for the clearance of Neisseria by NETs [8] . To demonstrate the antibacterial contribution of the latter two NET-bound proteins , antibodies raised to histones or cathepsin G were shown to limit the bactericidal capacity of NETs towards these pathogens [1 , 8] . Immunocompromised individuals , including those with Cystic Fibrosis ( CF ) , are particularly susceptible to P . aeruginosa infection , which is a major cause of morbidity and mortality in these patients . Chronic P . aeruginosa infection of the CF lung leads to an intense inflammatory immune response , resulting in the recruitment of large numbers of neutrophils to the site of infection [14 , 15] . CF sputum is highly enriched in neutrophil-derived DNA , including that of NET origin , indicating that neutrophils deploy NETs in an effort to combat infection of the lung [16 , 17] . However , persistence of P . aeruginosa in the midst of sustained neutrophil presence and NETosis suggests that the pathogen is capable of evading this host immune response [18] . An important virulence strategy adopted by successful microbial pathogens is the tolerance of NETs . In a number of cases described so far , microbial invaders accomplish this goal by either avoiding or disarming neutrophil extracellular traps . Modification of the bacterial capsule and surface-localized lipoteichoic acid reduces the trapping of Streptococcus pneumonia in NETs [19] . The secretion of extracellular nucleases by Staphylococcus aureus , Streptococcus pneumonia , group A Streptococcus and Vibrio cholerae highlight a shared virulence strategy that functions to degrade the DNA-backbone of NETs , enabling evasion or liberation of the bacteria from entrapment [2 , 20–22] . Although microbial or exogenous DNase is proposed to dissolve NET structures to avoid capture , it has not been considered that the DNA backbone itself may be antimicrobial . We have previously demonstrated that extracellular DNA is an efficient chelator of divalent metal cations [23] . This cation chelation has pleiotropic effects on P . aeruginosa depending on the concentration of extracellular DNA . At subinhibitory levels , sequestration of cations by DNA leads to the expression of genes controlled by the two component systems PhoPQ and PmrAB that sense Mg2+ limitation [23–25] . However , at higher concentrations , extracellular DNA causes dramatic disruptions to the bacterial envelope integrity , which leads to lysis and rapid cell death [23] . It is predicted that the phosphodiester backbone is required for cation sequestration of metal cations on the bacterial surface and the membrane-destabilizing antimicrobial activity of extracellular DNA [23] . Therefore , we hypothesize that the DNA backbone of NETs contributes to their antibacterial activity . Here we show that neutralizing the membrane destabilizing activity of extracellular DNA by quenching the capacity of the phosphodiester backbone to chelate cations protected bacteria from NETs . P . aeruginosa is capable of detecting the DNA lattice of NETs , and in response , upregulates genes required to modify the bacterial outer membrane surface to tolerate the toxic effects of DNA-mediated NET killing .
P . aeruginosa has been shown to induce the formation of neutrophil extracellular traps in purified human neutrophils [18] and NETs have been observed in CF sputum [16 , 17] , where P . aeruginosa is a predominant pathogen . PMA-induced NETs contained known neutrophil proteins embedded in the DNA lattice , including myeloperoxidase ( MPO ) and histones [2] ( Fig . 1A ) . We also show that Gfp-tagged P . aeruginosa is efficiently trapped and aggregated in NETs ( Fig . 1A ) . However , in vivo NETosis during P . aeruginosa infection has not been reported . Therefore intravital confocal microscopy was used to determine whether P . aeruginosa elicited NETosis in the mouse skin infection model [26] . Infection with P . aeruginosa led to the production of large NET-like structures that stained with the DNA-binding dye Sytox green and entrapped ChFP-labeled P . aeruginosa ( Fig . 1B and C ) . In addition to the presence of NETs , neutrophils remained chemotactic and phagocytosed bacteria , suggesting that multiple neutrophil clearance mechanisms are employed in vivo to combat P . aeruginosa ( Fig . 1D and S1 Movie ) . Given that P . aeruginosa induces the production of NETs in vitro and in vivo ( Fig . 1 ) , we sought to compare the relative abilities of P . aeruginosa , S . aureus , E . coli and the chemical inducer PMA to elicit NETosis . In the presence of purified human neutrophils , PMA , E . coli and S . aureus induced significantly more NET formation , relative to P . aeruginosa within 1 hour of coincubation ( Fig . 2A ) . However , at later time points ( 3h ) , P . aeruginosa elicited the formation of similar amounts of NET structures . Furthermore , quantification of the number of NETs and NET-area using the skin infection model confirmed that P . aeruginosa weakly induces NETosis relative to S . aureus in vivo ( Fig . 2B and C ) . One possible explanation for the reduced NET formation in the presence of P . aeruginosa is the production of a microbial secreted DNase that degrades NETs more efficiently than other organisms . Therefore , we measured the DNase activity in overnight culture supernatants , as well as in supernatants isolated from coincubating bacteria with PMA-treated neutrophils . We demonstrate that there was little , if any , DNase activity produced by P . aeruginosa , S . aureus or E . coli under these conditions ( S1 Fig . ) . Since cations are a requirement for DNase activity , we were able to restore DNase activity in S . aureus supernatants after the addition of excess cations ( S1 Fig . ) . However , under the cation-free conditions used to quantitate antibacterial NET function , even in the presence of PMA-stimulated neutrophils , DNase activity was not detected and thus P . aeruginosa appears to limit NETosis through an uncharacterized mechanism . In order to characterize the bactericidal capacity of NETs , neutrophils were treated with PMA to stimulate maximal NETosis and cytochalasin D to block phagocytosis , thus restricting bacterial killing to extracellular NET function [1 , 8 , 21 , 22] . Importantly , the addition of cytochalasin D had no effect on NETosis induced in PMA-treated neutrophils ( S2 Fig . ) . We used the conventional method of direct bacterial counts to enumerate the number of bacteria before and after challenge with PMA-induced NETs . Direct counts of NET-exposed bacteria revealed that P . aeruginosa was most tolerant to NET killing , whereas S . aureus and E . coli were significantly more sensitive ( Fig . 3A ) . The addition of deoxyribonuclease ( DNase ) restored bacterial survival of the NET-sensitive organisms E . coli and S . aureus , confirming that killing was mediated by extracellular NET function ( Fig . 3A ) . Further , the kinetics of bacterial killing by PMA-generated NETs was determined by measuring the loss of luminescence from chromosomally-tagged luminescent P . aeruginosa strain , PAO1::p16Slux [23] , and plasmid-borne luminescent E . coli / pσ70-lux [27] . This approach confirmed that P . aeruginosa was more tolerant to NET killing than E . coli , where luminescence rapidly decreased upon neutrophil challenge ( Fig . 3B ) . The addition of exogenous DNase I and the production of secreted DNases by bacteria are the most effective means to disable NET killing [1 , 2 , 20 , 21] . It is thought that DNase treatment dissolves NET structures , thereby releasing and diluting the antimicrobial proteins bound to NETs . We have previously shown that extracellular DNA has a potent antibacterial activity , as purified salmon DNA ( 2% w/v ) causes several log orders of bacterial killing within minutes and breaks the integrity of both the inner and outer membranes , leading to lysis [23] . DNA is a very efficient cation chelator and the antimicrobial activity of DNA can be blocked with addition of excess divalent metal cations [23] . It is predicted that the cation chelation is mediated by the phosphodiester backbone . To confirm the mechanism by which extracellular DNA kills bacteria , we monitored the loss of P . aeruginosa viability in the presence of dilute extracellular DNA ( 0 . 125% w/v; Fig . 4A ) . Bacterial survival was restored if DNA was pretreated with DNase or excess 5mM Mg2+ , which degrades DNA or saturates its cation chelating ability , respectively , thus neutralizing the antibacterial activity ( Fig . 4A ) . Pretreatment of extracellular DNA with calf intestinal alkaline phosphatase ( PTase ) , which cleaves 5’-phosphates , also blocked the observed antibacterial activity ( Fig . 4A ) . The addition of decreasing amounts of DNase , PTase or Mg2+ , resulted in marked , dose-dependent , decreases in bacterial survival when challenged with 0 . 15% DNA ( Fig . 4B ) . The DNA-mediated damage to the outer membrane leads to the formation of ChFP-enriched outer membrane vesicle-like structures ( OMVs; Fig . 4C ) [23] . However , incubation of P . aeruginosa with extracellular DNA pretreated with DNase , PTase and Mg2+ greatly diminished the number of ChFP-labeled OMVs relative to control conditions ( Fig . 4D ) , confirming that the bactericidal mechanism of extracellular DNA is through disruption of the bacterial membrane . Extracellular DNA-mediated damage to membrane integrity was confirmed by using flow cytometry ( Fig . 4E ) . Treatment of P . aeruginosa with DNA resulted in a new population of cells that were dual positive for SYTO9 and PI , indicative of membrane damage and increased PI uptake , and possibly dead cells . The increased staining of DNA-exposed bacteria by membrane-impermeable propidium iodide ( PI ) was not observed in DNase and Mg2+ pretreatments and was reduced in PTase pretreated DNA samples ( Fig . 4E and F ) . To ensure cation chelation was responsible for the observed bacterial membrane destabilization , we assessed whether the known cation chelator EDTA could cause membrane disruption . Like DNA , EDTA caused major outer membrane disruptions and the release of OMVs , as well as a dramatic increase in PI-staining of EDTA-treated cells monitored by flow cytometry ( S3 Fig . ) [28] . To determine whether the antibacterial capacity of DNA requires direct bacterial contact or can be mediated through passive cation sequestration , we exposed P . aeruginosa to high concentrations of DNA spatially separated by an ion-permeable barrier . Blocking direct interaction between P . aeruginosa and DNA resulted in bacterial survival after a prolonged exposure , compared to the rapid antibacterial activity of DNA in direct contact with the bacteria ( S4 Fig . ) . Together , these results demonstrate that the antibacterial activity of extracellular DNA requires direct contact , and the phosphate backbone for cation chelation , leading to membrane disruption and bacterial cell death . The bactericidal activity of neutrophil extracellular traps is attributed to direct contact and exposure of bacteria to the antimicrobial proteins embedded in the DNA scaffold of NETs [1 , 2] . Given the antimicrobial activity of DNA , we propose that the DNA backbone of the NET itself is antibacterial . Therefore , if DNA contributes to bacterial killing , treatments that quench the cation chelation potential of the DNA backbone will block bactericidal activity of NETs . To address this possibility , PMA-activated neutrophils were treated with the addition of DNase , PTase or excess Mg2+ and bacterial viability was monitored . The DNA-targeted treatments completely protected P . aeruginosa and E . coli from killing by neutrophil extracellular traps ( Fig . 5A ) . To confirm these results , we monitored the luminescence of P . aeruginosa PAO1::p16Slux co-incubated with PMA-activated neutrophils and observed that the antibacterial effects of NETs were neutralized by treatment with exogenous DNase , Mg2+ cations and PTase ( Fig . 5B ) . To determine whether the restored bacterial viability in the presence of NETs was due to preventing damage to the bacterial envelope , we performed flow cytometry to assess membrane integrity . Increased PI staining of NET-exposed bacteria is an indicator of membrane damage , which was completely blocked by addition of DNase ( Fig . 5C ) . The addition of excess Mg2+ or treatment with exogenous PTase also limited membrane damage ( Fig . 5C ) . Importantly , both Mg2+ and PTase treatments neutralized the antimicrobial activity of NETs ( Fig . 5A and B ) did not disrupt overall NET architecture , as MPO and histones were still present within the treated NET structures ( Fig . 6 ) . To assess the function of a NET-bound protein , we measured elastase activity in PMA-treated neutrophils and showed no difference in elastase activity when NETs were treated with exogenous PTase or Mg2+ ( S5 Fig . ) . NET structures remain intact and contain functional proteins ( elastase ) after treatment with PTase or excess Mg2+ , but are no longer antibacterial for E . coli and P . aeruginosa ( Fig . 5 ) . Together these results suggest an antibacterial mechanism wherein the DNA backbone of NETs target and destabilize the bacterial membrane and promotes cell death . Subinhibitory concentrations of extracellular DNA sequester Mg2+ and trigger the expression of multiple surface modifications that are known to protect the bacterial outer membrane from antimicrobial peptide ( AP ) damage and killing [23–25] . The arn operon ( PA3552-PA3559 ) is required for the covalent addition of aminoarabinose to the phosphates of lipid A and the spermidine synthesis genes ( PA4773-PA4774; speDE homologs ) are required for production of the polycation spermidine on the outer surface [23 , 24] . Both modifications substitute for divalent metal cations , mask the negative charges of the outer surface and thus contribute to AP resistance [23 , 24 , 29–31] . Given that these modifications stabilize the bacterial envelope , we sought to determine whether these surface modification pathways provided a more general mechanism to resist bacterial membrane damage . We noted that P . aeruginosa strains with mutations in the arn or spermidine biosynthetic pathways were significantly less capable of surviving exposure to DNA ( Fig . 7A ) . Given the role of these pathways for tolerance of exogenous DNA , we then investigated whether the DNA component of NETs induced expression of the arn or spermidine synthesis genes in P . aeruginosa . Expression of both pathways was strongly induced 2–6 fold following co-incubation with NETs produced by PMA-activated neutrophils ( Fig . 7B ) . To confirm that DNA was the component of NETs that led to induction of the bacterial gene expression response , the addition of excess Mg2+ cations and enzymatic treatment with DNase ( Fig . 7C ) and PTase ( S6 Fig . ) all blocked the induction of the arn and spermidine operons response to NETs . While these treatments specifically neutralize DNA , we also considered the possibility that NET-bound antimicrobial proteins including histones or LL-37 may elicit these protective responses . We have previously shown that sub-MIC concentrations of antimicrobial peptides induce both outer surface modifications [29] . Therefore , to assess the relative capacity of each NET component to act as a bacterial signal , we compared the ability of purified histones , the well-characterized APs polymyxin B and colistin , and DNA to induce the expression of the spermidine synthesis pathway . Although all NET components induced expression of the PA4773-PA4774 spermidine synthesis pathway ( S7 Fig . ) , DNA was the most potent inducer of this bacterial response ( S7 Fig . ) . The upregulation of protective outer membrane modifications ( aminoarabinose-modified LPS and surface spermidine production ) by NETs , and by the individual NET components of DNA and histones , suggests that these modifications are required to defend the membrane against assault from multiple innate immune components that target the bacterial membrane . Both modifications result in stable substitutions for divalent metal cations in the outer membrane , and protect P . aeruginosa from antimicrobial peptides [23 , 24 , 29–31] and DNA killing ( Fig . 7A ) . P . aeruginosa mutants in the arn and spermidine synthesis genes also exhibited increased susceptibility to the disruptive effects of NETs ( Fig . 7D ) . We next compared the susceptibilities of P . aeruginosa , E . coli and S . aureus to histone and DNA killing . Surprisingly , P . aeruginosa was the most susceptible to DNA killing , while S . aureus was the most tolerant , the opposite pattern of NET susceptibility ( Figs . 3A and 8A ) . The bactericidal capacity of purified histones was modest , where after 2 hours exposure P . aeruginosa was the more histone tolerant ( Fig . 8B ) . Taken together , these results highlight the ability of P . aeruginosa to quickly respond and defend against the DNA and histone mediated-antibacterial effects of NETs by stabilizing the outer membrane .
Given the observations that exogenous or secreted microbial DNases protect bacteria against NET killing and that extracellular DNA has rapid , membrane-damaging antibacterial activity , we sought to test the hypothesis that the DNA backbone of NETs contributes to their bactericidal function . Extracellular DNA possessed contact-mediated antibacterial activity that could be neutralized by enzymatic and cationic treatments that degrade or quench the capacity of DNA to chelate cations ( Fig . 4 and S3 Fig . ) . NETs exposed to the same treatments that target the DNA scaffold were unable to cause bacterial membrane damage or to cause bacterial killing of P . aeruginosa and E . coli ( Fig . 5 ) . Therefore , we propose a novel bactericidal mechanism of NETs whereby the removal of surface-stabilizing cations by the DNA phosphodiester backbone results in bacterial lysis ( Fig . 4B and C ) . To address the controversy surrounding the bacterial killing ability of NETs [5–7] , we used multiple viability assays to measure NET killing and membrane damage , which included direct bacterial counts , luminescence viability assays and flow cytometry of PI-stained cells [32] . Taken together , these data support the general notion that NETs are directly antimicrobial . Deciphering the specific antimicrobial mechanisms of NETs has been limited to a few candidate proteins [1 , 8–11] . Most studies have focused on the NET-bound proteins as the antimicrobial components , given their important role during phagocytosis and degranulation . Although granular , cytoplasmic and nuclear proteins derived from neutrophils can be detected in NETs by immunofluorescence , the abundance of most NET-bound proteins is low ( <1–6% ) , relative to histones , which comprise 65% of the total protein content [10] . Although classically characterized as chromatin structural proteins , NET-bound histones possess antimicrobial activity that can be neutralized through the addition of anti-histone antibodies [1 , 12 , 13] . However , other neutrophil proteins exhibit altered or reduced enzymatic activity when enmeshed in the NET backbone raising questions as to their antimicrobial capacity . For example , S . aureus tolerates myeloperoxidase in NETs unless supplemented with the addition of exogenous H2O2 [11] . Neutrophil elastase activity increases in DNase treated sputum from Cystic Fibrosis patients , suggesting that DNA may inhibit elastase activity , thus limiting the role of elastase as a potential NET-bound factor [17] . We observed that both the antibacterial activity of NETs and the ability to induce protective bacterial responses were blocked by treatments that target extracellular DNA , suggesting that the NET scaffold is not simply a passive structure ( Figs . 5 and 7 ) [23–25 , 31] . The most potent inducing triggers of the P . aeruginosa surface modifications are purified eDNA , followed by APs and purified histones ( S7 Fig . ) . Since being widely induced by these components , it is not surprising that the aminoarabinose-modified LPS and spermidine synthesis pathways protect the outer membrane from DNA , NETs ( Fig . 7 ) and antimicrobial proteins [23 , 24 , 29–31] . The low potency of the tested histones may be explained by the fact that our assays were performed with a mixture of full-length histones , which had modest antibacterial activity ( Fig . 8 ) . Recent evidence highlights that histones are proteolytically processed by proteases such as elastase during the process of nucleus decondensation , prior to NET release [33] It is therefore likely that potent bactericidal histone-derived peptides are present in NETs as an important antibacterial component of NETs . P . aeruginosa appears to mount a multifunctional , outer membrane defense strategy to combat multiple antimicrobial components enmeshed in NETs . Consistent with this model , we noticed unexpected susceptibility patterns that also suggest that NET killing may be the result of the combinatorial effect of DNA and NET-bound proteins . The observation that S . aureus is susceptible to NET killing but tolerant to DNA ( Figs . 3 and 8 ) , suggests that NET killing of S . aureus is likely dependent on other anti-staphylococcal proteins enmeshed in the DNA lattice . The DNA susceptibility phenotype of P . aeruginosa may explain the potency of DNA as the strongest inducer of the protective outer surface modifications that contribute to the observed NET tolerance . It is intriguing to speculate that the NET-bound , antimicrobial proteins act in concert with the antibacterial activity of DNA to provide broad-spectrum protection against a wide range of microbial pathogens . Modification of the bacterial cell surface and the production of secreted DNases are virulence strategies utilized by microbial pathogens to evade NET killing [19–22] . Here we report that the spermidine and the arn surface modification pathways are required to tolerate the antibacterial action of both DNA and NETs ( Fig . 7 ) . The covalent addition of aminoarabinose to the lipid A component of LPS masks the negative charges of core LPS phosphates , and the polycationic nature of spermidine ( +3 charge ) substitutes for surface divalent metal cations , and may also bind and neutralize DNA . In addition , spermidine possesses an antioxidant activity that protects bacterial membrane lipids from oxidative damage [24] and therefore may protect P . aeruginosa from NET-induced oxidative damage [11] . Combined , these results suggest that the spermidine and arn surface modifications possess multiple protective roles that may contribute to resisting a broad range of antimicrobial components present within NETs . We propose that bacterial surface-bound , divalent metal cations are displaced by direct contact with extracellular DNA , and that DNA-induced surface modifications prevent outer membrane disruption and bacterial killing by NETs . Therefore , the antibacterial mechanism of cation chelation exerted by DNA is distinct from that of other previously characterized antimicrobial cation chelating proteins such as calprotectin . We have previously demonstrated that DNA chelates diverse metal cations ( Mg2+ , Ca2+ , Zn2+ , Mn2+ ) [23] while calprotectin chelates zinc and manganese [34] . Additionally , the antimicrobial function of calprotectin is contact-independent whereas the bactericidal function of DNA requires contact ( S4 Fig . ) . Further , sequestration of zinc and manganese by calprotectin does not target the microbial membrane but rather sequesters cation cofactors required by bacterial enzymes such as superoxide dismutase , which protects bacteria from superoxide [34] . In summary , we have identified that the DNA backbone is a bona fide antibacterial component of neutrophil extracellular traps . The DNA scaffold structure also acts as a warning signal perceived by P . aeruginosa . Overall , these results support a model where the membrane-destabilizing activity of the DNA scaffold contributes to the bactericidal capacity of NETs , while the cation chelating activity acts as a signal perceived by P . aeruginosa that leads to upregulation of protective surface modifications . These results highlight a dynamic bacterial-host interaction between an opportunistic pathogen that causes chronic infections in the lungs of individuals with Cystic Fibrosis , an infection site known to be rich in neutrophil DNA and neutrophil extracellular traps [16 , 17] . This ability to sense and defend against NETosis may help explain the long-term persistence of P . aeruginosa in CF lung infections .
All strains and plasmids used in this study are shown in S1 Table . Bacterial cultures were routinely grown at 37°C in LB or BM2 defined minimal media with 0 . 5 mM MgSO4 , unless otherwise stated . S . aureus was grown overnight in BHI media . When necessary , the following antibiotics were used: 50 µg/mL tetracycline for P . aeruginosa mini-Tn 5-lux mutants , and 50 µg/mL kanamycin for E . coli DH5α/ pσ70-lux . Mid-log cultures were used for co-incubation experiments with neutrophils or extracellular DNA . Neutrophils were isolated from healthy donors as previously described [35] . Whole blood was collected and mixed 5:1 in acid citrate dextrose , followed by removal of red blood cells using dextran sedimentation and hypotonic lysis with KCl . After all red blood cells were lysed , the cell pellet was subjected to Ficol-Histopaque density centrifugation . The subsequent pellet was resuspended in 2 mL of HBSS ( Hank’s balanced salt solution , no cations; Invitrogen 14175–095 ) . The viable cell concentration was determined using a haemocytometer and Trypan blue staining . Glass cover slips were HAS-coated and placed in 6-well tissue culture plates . Neutrophils were added at 2 . 0×106 cells/mL per well , adhered ( 30 min , 5% CO2 , 37°C ) and treated with cytochalasin D ( 10 µg/well ) and PMA ( 25 nM ) to activate NETosis [35] . Mid-log bacterial cultures were diluted in HBSS ( no cations ) ( 5 . 0x107 CFU/mL ) for an MOI of 25:1 , centrifuged to the neutrophils , and coincubated for 1–4 hours ( 5% CO2 , 37°C ) . Cells on cover slips were fixed with 4% paraformaldehyde , washed with 250 µL of 10% FBS ( Invitrogen ) in PBS and stained with either DNA dyes and/or various primary and secondary antibodies ( described below ) . For NET visualization with antibodies , the primary anti-human MPO antibody ( DakoCytomation- A0398 ) was diluted into 10% FBS in PBS ( 1/500 ) . 30 µL was added to adhered neutrophils , incubated ( 30 min , 37°C ) and washed twice with sterile PBS . 40 µL of the secondary anti-rabbit Cy 5 antibody ( Jackson ImmunoResearch 60354 ) ( 1/500 dilution ) was added . After 15 min incubation in the dark , cover slips were washed twice with PBS , and prepared with mounting media . Anti-DNA and anti-histone antibodies where obtained from Dr . Marvin Fritzler . Either the anti-DNA ( 1:10 ) or anti-histone ( 1:500 ) antibodies [36 , 37] were added as described above . The anti-human secondary antibodies with Alexa Flour 647 ( Invitrogen A21445 , 1/500 ) were added to cover slips and mounted as described above . Images of human NETs were acquired using the Leica DMI 4000B inverted microscope equipped with ORCA R2 digital camera and Metamorph software for image acquisition using the 63X or 100X objectives . The following excitation and emission filters were used for blue fluorescence ( Ex 390/40; Em 455/50 ) , red fluorescence ( Ex 555/25; Em 605/52 ) , far red fluorescence ( Ex 645/30; Em 705/72 ) and green fluorescence ( Ex 490/20; Em 525/36 ) . Images were formatted and analyzed using the Imaris 7 . 0 . 0 imaging software . All images shown are representative of at least three experiments . Mice were anaesthetized ( 10 mg/kg xylazine hydrochloride and 200 mg/kg ketamine hydrochloride ) and body temperature was maintained using a rectal probe and heating pad . The mice were pretreated with intradermal MIP-2 ( 0 . 2µg/injection ) diluted in sterile normal saline 30 minutes prior to imaging . The right jugular vein was cannulated to administer additional anesthetic and fluorescent dyes . The microcirculation of the dorsal skin was prepared for microscopy as previously described [26] . Briefly , after shaving the mouse’s back , a midline dorsal incision was made extending from the tail region up to the level of the occiput . The skin was separated from the underlying tissue , remaining attached laterally to ensure the blood supply remained intact . The area of skin was then extended over a viewing pedestal and secured along the edges using 5 . 0 sutures . The loose connective tissue lying on top of the dermal microvasculature was carefully removed by dissection under an operating microscope . The exposed dermal microvasculature was immersed in isotonic saline and covered with a coverslip held in place with vacuum grease . Alexa Fluor 649 conjugated anti-mouse GR-1 antibody ( 10µl per mouse i . v . ; eBioscience ) was used visualization of neutrophils . To visualize NETs in vivo the membrane impermeable dyes SYTOX-green or SYTOX-orange were administered ( diluted 1:1000 with sterile saline , 100µl per mouse i . v . ) . MIP-2 injection ( 0 . 2µg/injection ) was initiated 30 min prior to Pseudomonas aeruginosa or S . aureus administration . Following baseline visualization , all bacteria were directly administered into the field of view using a tuberculin needle ( 1×108 CFU/100µl of sterile saline , i . d . ) . Spinning disk confocal intravital microscopy was performed using an Olympus BX51WI ( Olympus , Center Valley , PA ) upright microscope equipped with a 20×/0 . 95 XLUM Plan Fl water immersion objective . The microscope was equipped with a confocal light path ( WaveFx , Quorum , Guelph , ON ) based on a modified Yokogawa CSU-10 head ( Yokogawa Electric Corporation , Tokyo , Japan ) . Laser excitation at 488 , 561 and 649nm ( Cobalt , Stockholm , Sweden ) , was used in rapid succession and fluorescence in green , red and blue channels was visualized with the appropriate long pass filters ( Semrock , Rochester , NY ) . Exposure time for all wavelengths was between 500 and 600ms . Sensitivity settings were maintained at the same level for all experiments . A 512×512 pixels back-thinned EMCCD camera ( C9100–13 , Hamamatsu , Bridgewater , NJ ) was used for fluorescence detection . Volocity Acquisition software ( Improvision Inc . , Lexington , MA ) was used to drive the confocal microscope . Images captured using the spinning disk were processed and analyzed in Volocity 6 . 0 . 1 . NET area and NET number were quantified using the Volocity software . NET area was determined using Volocity imaging software . Briefly , in each field of view ( FOV ) the threshold of the corresponding fluorescent channel in which NET structures were stained was set to eliminate the background staining of the skin . The area and number of NET positive structures in the FOV was calculated and counted via the Volocity 6 . 0 . 1 software . Structures that showed no characteristic NET-like shape and resembled the staining of a nucleus of an obvious dead cell in the FOV were excluded from the quantification manually . NET image analysis was performed in at least two infected or uninfected animals and from 5 fields of view . NETs were quantitated by measuring the amount of extracellular DNA that stains with the cell impermeant dye Styox Green [35] . Cell culture media ( CCM ) consisted of 48 . 5 mL of RPMI 1640 ( Invitrogen ) , 0 . 5 mL of 1 . 0 M HEPES and 1 . 0 mL of human serum albumin ( HAS; Innovative Research ) . Neutrophils were diluted into CCM ( 2 . 0×105 cell/well ) and added to an HSA-coated , 96-well black , clear-bottom plate ( Thermo Scientific ) . As a positive control , PMA ( 25 nM/well ) was added to activate the neutrophils [35] . For bacterial activation of NETs , mid-log bacterial cultures were diluted in CCM ( 2 . 0×106 CFU/well ) for a multiplicity of infection ( MOI ) of 10:1 ( bacteria to neutrophils ) . DNase ( 430 kU/well; VWR 31149 ) was added to degrade extracellular DNA in NETs . Bacteria were gently centrifuged onto the adhered neutrophils ( 800xg , 10 min ) . Sytox green ( 2 . 5 µM; Invitrogen ) was added to each well and green fluorescence ( Ex 490/8; Em 535/25 ) was measured with Perkin Elmer 1420 Multilabel Counter Victor3 between 1 and 4 hours . All values shown are the mean from at least three individual replicates and each experiment was performed at least 3 times . We noticed variation in the background level of NET staining between individual donors , but there was a reproducible and robust 2 . 5 to 10-fold increase in NETosis after PMA treatment of neutrophils from various donors ( S2 Fig . ) . Quantification of Bacterial Viability and Gene Expression Using Plate Counts and Luminescence NET killing was examined using direct plate counting methods where a reduction in cell number indicated bacterial killing . All NET killing experiments were performed in HBSS solution lacking divalent cations . Isolated human neutrophils were mixed with mid-log bacteria in HBSS ( no cations ) in black , clear-bottom 96-well plates with of 2 . 0 × 107 CFU bacteria and 2 . 0 × 106 neutrophils ( MOI 10:1 ) . After a 1–4 hr incubation , 50 µL of DNase I solution ( 430 kU/mL ) was added to every well , mixed , and incubated for 30 min at 37°C , in order to release bacteria trapped in NETs for accurate plate counts . 15 µL of suspension was serially diluted ( 1/10 ) in 0 . 9% NaCl solution in a sterile 96-well plate and 5 µL from each well was stamped onto LB agar plates to obtain bacterial plate count data for time zero ( T0 ) and after 4h ( T4 ) . CFU/ml values from T4 and T0 time points were used to calculate the percentage survival by subtracting the T4—T0 plate counts and dividing the ‘bacteria and neutrophil’ conditions by the ‘bacteria alone’ conditions and multiplying by 100 . For lux viability and gene expression assays [23] , bacteria were centrifuged onto the adhered neutrophils and placed in the Victor3 plate reader for luminescence ( CPS ) measurements every 20 minutes for 3–4 hours . All values shown are the mean from at least six individual replicates and each experiment was performed at least 3 times . SYTO9 stains the DNA in all cells and propidium iodide ( PI ) stains the DNA in dead cells and cells with damaged membranes [28 , 32] . The sample of bacteria and neutrophils ( ~200 µL ) was placed in 5 mL polystyrene round-bottom sample tubes and stained with SYTO9 and propidium iodide at final concentrations of 0 . 02 mM and 0 . 2 mM , respectively . The tubes were centrifuged at 300x gravity and incubated ( RT , 15 min ) . Bacterial cells were analyzed using the BD LSRII flow cytometer ( BD Bioscience , San Jose , USA ) equipped with a blue laser ( 488nm ) and a green laser ( 532nm ) . Unstained , mid-log bacterial cells were used to gate the forward scatter ( FSC ) and size scatter ( SSC ) parameters . For green and red fluorescence profiles , SYTO9 was excited by blue ( 488nm ) laser with emission filters 525/50BP and 505LP and PI was excited using the green ( 532 nm ) laser with emission filters 610/20BP and 600LP . All detectors were set to the logarithmic amplification with the following voltages , 500 , 240 , 596 , and 489 and threshold was set at 200 for both FSC and SSC . For each sample , 50 000 events were acquired using the BD FACSDiva software 6 . 1 . 3 . The Hierarchical gating strategy was used to determine double positive population of bacterial cells ( stained with both SYTO9 and PI ) where gate P1 is the total population of FSC and SSC gated events , as determined from bacteria alone control and then applied to all other samples . P2 is the population of events stained by SYTO9 and P3 is the population stained with both SYTO9 and PI . Neutrophils and mid-log bacteria controls do not contribute any autofluorescence or PI-stained events when stained with either or both of the SYTO9/PI dyes . Values displayed in each density plot represent the percentage of 50 000 cells ( N value ) in each quadrant gate and each experiment was performed with at least 5 times . During the coincubation experiments of bacteria and PMA-activated neutrophils , exogenous Mg2+ was added at a final concentration of 5 mM MgSO4 . For the enzyme treatments , deoxyribonuclease ( DNase I , VWR ) was added at a final concentration of 430 kU/well and calf intestinal alkaline phosphatase ( PTase , Invitrogen ) at a final concentration of 16 . 6 U/well . Maximum enzyme amounts were added to bacterial-neutrophil mixtures that had no effect on bacterial viability and without the addition of enzyme buffers . The killing experiments were incubated for up to 4 hours in the 5% CO2 incubator at 37°C . All % survival values shown are the mean from at least three individual replicates and each experiment was performed at least 3 times . P . aeruginosa was grown to mid-log in LB medium ( OD600 = 0 . 2–0 . 4 ) , washed and resuspended in 10 mM Tris buffer ( pH 7 . 4; 1 . 0 × 107 CFU/well ) . Cells were incubated with fish sperm DNA ( 0 . 125% , w/v; USB ) or with DNA that had been pretreated with exogenous DNase I ( 150 kU/well ) , PTase ( 50 U/well ) or 5 mM MgSO4 . DNA was pretreated for up to 3 hrs at 37°C in order to neutralize the antimicrobial activity . To determine if DNA killing required direct cell contact , 2% w/v fish sperm DNA ( USB ) was resuspended in 10 mM Tris pH 7 . 4 was placed in sealed dialysis membranes ( MW cutoff 3500 Da ) and allowed to dialyze into 10 mM Tris pH 7 . 4 for 4 hours , exchanging the buffer every hour . Cells from mid-log P . aeruginosa PAO1 cultures ( 1 × 107 CFU ) were washed into 10 mM Tris pH 7 . 4 and coincubated directly with 1% or 0 . 125% dialyzed DNA ( final concentration ) , with 1% eDNA maintained inside dialysis tubing , or 10 mM Tris pH 7 . 4 alone as a negative control . For histone killing experiments , 1 × 107 CFU mid-log growth phase P . aeruginosa PAO1 , E . coli and S . aureus were washed into 10 mM Tris pH 7 . 4 and subsequently coincubated directly with 1 . 5 µg/mL calf thymus histones ( Roche ) . Killing experiments were performed at RT in 96-well microplates and bacterial survival was assessed by colony counts ( CFU/ml ) every hour . All survival values shown are the mean from 4–8 individual replicates and each experiment was performed at least 3 times . Differences in bacterial survival were statistically analyzed by two-tailed student t-test . PAO1::OM-lipoChFP was used as an indicator for outer membrane damage . This strain of PAO1 expresses a synthetic Cherry fluorescent lipoprotein ( CSFPOmlA-ChFP ) anchored to the outer membrane encoded on plasmid pCHAP6656 [38] . PAO1::OM-lipoChFP was exposed to a lethal concentration of 2% w/v ( 20 mg/ml ) extracellular sperm DNA ( USB ) or 2 mM EDTA and red fluorescence of untreated and DNA killed cells were monitored as described above . Fluorescent outer membrane vesicles ( OMVs ) were counted in 6 fields of view by ImageJ quantification using a manually controlled threshold cutoff . Overnight cultures were grown in LB medium , diluted 1/100 ( approximately 1 × 107 CFU ) into 100 µl of HBSS medium lacking cations ( Life Technologies ) in 96-well black plates with a transparent bottom ( Thermo Scientific ) and overlaid with 75 µl of mineral oil ( Sigma Aldrich ) to prevent evaporation . Microplate planktonic cultures were incubated at 37°C in a Wallac Victor3 luminescence plate reader ( Perkin-Elmer ) and optical density ( growth , OD600 ) and luminescence ( gene expression , CPS ) readings were taken every 20 minutes in the presence of 0 . 2% salmon sperm DNA , 0 . 125 µg/mL polymyxin B and colistin , and 0 . 1 µg/mL calf thymus histones ( Roche ) . Mean gene expression was derived from triplicate samples at 180 minutes after initial dilution and error bars represent the standard deviation from 4 individual replicates . Differences were statistically assessed by two-tailed student t-test . Overnight cultures of S . aureus and E . coli were grown in BHI medium and P . aeruginosa in BM2 medium , normalized to an OD600 = 1 and supernatants were collected by centrifugation at 8000 rpm for 3 minutes . 15 µL of supernatant was incubated with 5 µg of P . aeruginosa genomic DNA for 1 h at 37°C . Pseudomonas aeruginosa genomic DNA was purified using the Wizard Genomic DNA purification kit ( Promega ) . DNA degradation was visualized on red safe ( FroggaBio ) stained 1% agarose gels . To test whether exposure to NETs induced DNase production , supernatants from S . aureus , E . coli and P . aeruginosa incubated in HBSS lacking cations with 106 PMA-stimulated human neutrophils ( MOI 10:1 , same method as described in the NET killing experiments section ) were collected by centrifugation at 8000 rpm for 3 minutes . 100 µL of the supernatants were then coincubated at 37°C with 5 µg salmon sperm DNA stained with 2 . 5 µM Sytox green . 90 kU of DNase I was included as a positive control . Reactions were placed in 96-well black plates with a transparent bottom and Sytox green fluorescence quantified after 1 hour in a Wallac Victor3 luminescence plate reader . To determine whether phosphatase treatment or the presence of excess Mg2+ altered NET-bound protein function , 2 × 105 human neutrophils were seeded in 96-well black plates with a transparent bottom and induced with 100 nM PMA . Immediately after PMA addition , 50 units of phosphatase ( CIAP ) and 5 mM MgSO4 were added to wells . The plate was then placed in cell culture conditions for 2 hours ( 37°C , 5% CO2 ) . 300 µM elastase substrate I was added to all wells ( Calbiochem ) , which were subsequently overlaid with 75 µL of mineral oil . The plate was then placed in a Wallac Victor3 luminescence plate reader at 37°C . Neutrophil elastase activity was monitored by measuring absorbance at OD410 nm every 20 minutes for 8 hours . Statistical analysis was performed using GraphPad Prism v4 . 0 software . One-way ANOVA with Bonferroni posts tests and two-tailed students t-tests were used to calculate significant differences for plate counts , luminescence and flow cytometry analyses . Significant differences refer to P< 0 . 05 or less , or as otherwise denoted . Human neutrophils were isolated from human blood samples with ethical approval by the University of Calgary Research Ethics Committee ( Ethics ID# 23187 ) , where all subjects provided written informed consent . All animal protocols were approved by the animal care committee of the University of Calgary under the protocol number AC12–0222 . All protocols used were in accordance with the guidelines drafted by the University of Calgary Animal Care Committee and the Canadian Council on the Use of Laboratory Animals . | Comprising the first line of the innate immune response , neutrophils combat infectious microorganisms through the release of toxic molecules , phagocytosis of invaders and the production of the recently characterized neutrophil extracellular traps ( NETs ) . The antimicrobial activity of NETs has been attributed to proteins bound to the DNA backbone . Our results demonstrate that the DNA lattice of each NET is potently antibacterial and elicits upregulation of protective surface modifications by the opportunistic bacterial pathogen Pseudomonas aeruginosa . These modifications , previously shown to protect bacteria from antimicrobial peptides , confer greater bacterial tolerance to DNA and NET-mediated antibacterial activity . Treatments that quench the cation chelating capacity of DNA restore bacterial viability and suppress the expression of surface modifications even in the presence of intact NETs . These observations highlight the dual function of DNA as an antibacterial component of NETs , but also a signal perceived by bacteria to induce broad host resistance strategies . Therefore , the ability of P . aeruginosa to sense and defend against the antibacterial activity of neutrophil extracellular traps may contribute to long-term survival in chronic infection sites including the Cystic Fibrosis lung . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | DNA Is an Antimicrobial Component of Neutrophil Extracellular Traps |
Specificity within protein kinase signaling cascades is determined by direct and indirect interactions between kinases and their substrates . While the impact of localization and recruitment on kinase–substrate targeting can be readily assessed , evaluating the relative importance of direct phosphorylation site interactions remains challenging . In this study , we examine the STE20 family of protein serine–threonine kinases to investigate basic mechanisms of substrate targeting . We used peptide arrays to define the phosphorylation site specificity for the majority of STE20 kinases and categorized them into four distinct groups . Using structure-guided mutagenesis , we identified key specificity-determining residues within the kinase catalytic cleft , including an unappreciated role for the kinase β3–αC loop region in controlling specificity . Exchanging key residues between the STE20 kinases p21-activated kinase 4 ( PAK4 ) and Mammalian sterile 20 kinase 4 ( MST4 ) largely interconverted their phosphorylation site preferences . In cells , a reprogrammed PAK4 mutant , engineered to recognize MST substrates , failed to phosphorylate PAK4 substrates or to mediate remodeling of the actin cytoskeleton . In contrast , this mutant could rescue signaling through the Hippo pathway in cells lacking multiple MST kinases . These observations formally demonstrate the importance of catalytic site specificity for directing protein kinase signal transduction pathways . Our findings further suggest that phosphorylation site specificity is both necessary and sufficient to mediate distinct signaling outputs of STE20 kinases and imply broad applicability to other kinase signaling systems .
Protein kinases are key enzymes in signal transduction networks critical to essentially all aspects of cellular regulation . The human genome encodes over 500 protein kinases that function in distinct processes yet share a structurally conserved catalytic domain [1] . The unique functions of each kinase are attributable to different modes of regulation ( signal input ) and different sets of protein substrates serving as their effectors ( signal output ) . Classically , the capacity of kinases to target unique substrates was attributed to differences in the catalytic cleft that mediate recognition of distinct phosphorylation site consensus sequences [2] . However , even kinases belonging to the same family and sharing substantial sequence similarity within their catalytic domains can phosphorylate different substrate proteins [3] . There has been a growing appreciation that noncatalytic domains , adaptor proteins , and scaffolds can have important roles in substrate recruitment [4–6] . For closely related kinases , these interactions can override catalytic site specificity in driving substrate selection [7] . However , the extent to which phosphorylation site sequence motifs are necessary or sufficient to mediate selective signaling is not clear for most kinases . It is relatively straightforward to affect substrate recruitment through deletion or substitution of noncatalytic domains and binding partners . However , a lack of general approaches to perturb catalytic site interactions while maintaining kinase activity makes it difficult to assess their relative contribution to substrate targeting . One prominent example of a closely related , yet functionally distinct , group of kinases is the STE20 family . The 30 serine–threonine kinases in this group have been classified into 10 subfamilies based on similarity within the catalytic domain and overall domain architecture ( S1 Fig ) [8] . Perhaps the best characterized are the two subfamilies of p21-activated kinases ( PAKs ) , which function in growth factor and adhesion-dependent signaling pathways regulating the actin cytoskeleton to promote cell motility and survival [9] . More than 80 direct PAK substrates have been identified that collectively contribute to these functions [10 , 11] . Most of the remaining STE20 kinases constitute multiple so-called germinal center kinase ( GCK ) subfamilies . Many GCKs transduce stress signals to promote cell cycle arrest and apoptosis , for example , by acting as upstream regulators of the c-Jun N-terminal kinase ( JNK ) and p38-mitogen–activated protein kinase ( MAPK ) cascades [12] . GCKs are also important upstream components of the tumor-suppressive Hippo pathway . Several of them , most prominently Mammalian sterile 20 kinase ( MST ) 1 and MST2 , phosphorylate the serine–threonine kinases large tumor suppressor homolog 1 and 2 ( LATS1/2 ) within a C-terminal hydrophobic motif [13–16] , promoting their activation and subsequent phosphorylation of the Yes-associated protein ( YAP ) and Transcriptional co-activator with PDZ binding motif ( TAZ ) transcription factors . Some GCKs , including thousand and one amino acid kinase 1 ( TAO1 ) and TAO3 , can also act further upstream through direct phosphorylation and activation of MST2 [17 , 18] . Prior studies have largely emphasized the contribution of subcellular localization and noncatalytic site interactions to selective substrate targeting by STE20 kinases . For example , PAKs are localized and activated through interactions with small GTPases , producing an active pool of the kinase in a spatially restricted manner that likely limits its substrate repertoire . Multiple STE20 kinases , including some PAKs and GCKs , harbor proline-rich motifs that bind to Src homology 3 ( SH3 ) domains present in substrates , activators , and/or adaptor proteins [19–22] . At least some members of the family ( PAKs , TAOs , STE20/SPS1-related proline-alanine–rich kinase [SPAK]/Oxidative stress-responsive 1 [OSR1] , and Lymphocyte-oriented kinase [LOK] ) interact with defined regions of their substrates distal from sites of phosphorylation through either catalytic or noncatalytic domains [23–26] . Finally , several STE20 kinases have stable binding partners that may act as substrate adaptors [13 , 14 , 27–33] . Despite the established importance of noncatalytic site interactions , differences in phosphorylation site sequence motifs among STE20 kinases may also be functionally important . For example , the established catalytic site motif of PAKs corresponds well to sequences found at known sites of phosphorylation in protein substrates [34] , albeit with some individual sites diverging substantially . Furthermore , subtle differences in the phosphorylation site sequence can direct different members of the PAK subfamily to unique substrates [35] . While peptide library analysis of a limited number of GCKs has revealed a specificity profile distinct from PAKs [36 , 37] , there has been no systematic analysis of the entire STE20 family , and specific determinants of specificity that distinguish members of the family are not known . In this study , we probe the catalytic site specificity of the STE20 family and discover previously unknown features within the kinase domain that help determine selective substrate targeting . Exploiting these features , we engineer kinase mutants that fully exchange phosphorylation site specificity between subfamilies . These reprogrammed kinases provide an unprecedented opportunity to dissect the unique contribution of catalytic site interactions to the signaling output of a protein kinase . We show that phosphorylation site specificity has a predominant role in mediating substrate targeting in STE20-kinase–mediated signaling networks .
To better understand how STE20 kinases mediate selective substrate targeting , we performed positional scanning peptide array ( PSPA ) screens across the entire family . We used a previously reported set of 198 peptide mixtures in which nine positions flanking a central phosphorylation site are systematically substituted with each of the 20 unmodified amino-acid residues , as well as phosphothreonine ( pThr ) and phosphotyrosine ( pTyr ) [38 , 39] . In this method , we perform parallel radiolabel kinase assays in solution to determine relative rates of phosphorylation of each arrayed peptide mixture . We have previously reported PSPA analysis of five members of the PAK subfamily [35 , 40 , 41] and four additional kinases from other STE20 subfamilies [36 , 42] . To enable efficient screening of the remaining STE20 family kinases , we modified the PSPA protocol [38 , 39] to a semiautomated platform ( see Materials and methods ) . Optimization of this protocol improved interassay reproducibility over the fully manual assay format ( S2 Fig ) , consumed less kinase , and allowed more efficient multiplexing . Using the semiautomated PSPA protocol , we profiled most of the STE20 kinases not previously analyzed . Together with previously published results , we have collected PSPA data in total for 20 of the 28 active STE20 kinases ( data for representative kinases are shown in Fig 1 , with remaining kinases in S3 Fig; quantified data for all PSPA experiments are provided in S1 Data ) . The remaining eight kinases were either unavailable for screening or provided insufficient signal above background in the PSPA; two members of the family ( STE20-related kinase adapter protein [STRAD]α and STRADβ ) are pseudokinases and were not analyzed . With the exception of the myosin-III ( MYO3 ) kinases ( GCK-VII subfamily ) , our data include at least one member of each subfamily within the STE20 group ( S1 Fig ) . Because available PSPA data for closely related members of the same subfamily were indistinguishable from each other , it is likely that the remaining kinases ( SPAK , Kinase homologous to SPS1/STE20 1 [KHS1] , NIK-related protein kinase [NRK] , TAO1 , TAO3 , and PAK3 ) have profiles similar if not identical to their closest relatives . For nine kinases , we also produced and assayed a kinase-inactive mutant form . We observed no detectable activity with any of these kinase-inactive mutants , suggesting that our results with the wild-type ( WT ) kinases were not attributable to contaminants in our purified preparations . Based on the PSPA screens , we could categorize the family into four groups based on their target motifs ( Fig 1A ) . The only common feature consistently recognized by all STE20 family kinases was a hydrophobic residue immediately downstream of the phosphorylation site ( the +1 position ) . The PAKs were notably divergent from all of the other kinases we screened , having , as previously reported , a primary preference for basic amino acids at the −2 and −3 positions and for Ser over Thr as the phosphoacceptor residue . By contrast , all other members of the STE20 family shared a strong preference for basic residues at multiple positions downstream of the phosphorylation site , selected Thr over Ser as the phosphoacceptor , and preferred an aromatic residue at the −2 position . Kinases outside of the PAK subfamily could be further subdivided into three categories based on variations within this core motif or by differences observed at other positions . The largest category , comprising 11 of the kinases we screened across four subfamilies ( hereafter referred to as the GCKi–v group ) , strongly preferred Lys or Arg residues clustered at the +2 and +3 positions ( Fig 1 , S3 Fig , and S4 Fig ) . By contrast , the kinase OSR1 preferred a His residue at the +2 position as well as the −1 position and showed a preference for a Tyr residue at the −3 position . TAO2 preferred Asn at the +2 position and also selected phosphorylated amino acids ( pThr and pTyr ) at several positions , suggesting that its substrates may be “primed” by prior phosphorylation at nearby sites ( Fig 1 and S1 Data ) . These categories aligned well with the phylogenetic relationships within the STE20 group ( S1 Fig and S4 Fig ) . The residues most strongly selected by most STE20 kinases as determined by PSPA were over-represented in previously reported phosphorylation sites in protein substrates ( Fig 1 and S3 Fig ) with some exceptions . For example , while Trp was selected most strongly at the +1 position by all STE20 kinases , authentic protein substrates reflected the general preference for hydrophobic residues at this position seen by PSPA analysis . Furthermore , basic residues were less common at positions downstream of the phosphorylation site in MST4 ( and MST1/2; S3 Fig ) protein substrates than would be expected from their strong enhancement of peptide substrate phosphorylation . This observation could reflect redundancy in which the presence of basic residues at all three downstream positions is not required for maximal activity . While the phosphorylation site motif of PAKs has been explored in depth [35 , 45] , there has been comparatively little characterization of other STE20 kinases . To more quantitatively assess the contribution of specific residues to phosphorylation by kinases outside the PAK subfamilies , we designed a peptide substrate ( MSTtide , NKGYNTLRRKK ) incorporating residues selected by these kinases but specifically optimized for the GCKi–v group that includes MST4 ( S3 Fig ) . As anticipated , MST4 phosphorylated this peptide robustly ( Km = 25 ± 5 μM , kcat = 3 . 7 ± 0 . 3 s−1 , Fig 2A ) . Amino-acid substitutions at several positions within the peptide reduced the phosphorylation rate between 2- and 10-fold , with the most substantial decreases observed with substitutions at the +1 and +3 positions ( Fig 2B ) . We previously reported a similar reduction in MST4 peptide kinase activity upon exchanging a Ser residue for a Thr phosphoacceptor [46] . MST4 did not detectably phosphorylate a consensus PAK peptide substrate ( PAKtide , RKRRNSLAYKK ) , indicating that combined substitutions at multiple positions further decrease the phosphorylation rate . To confirm trends in specificity observed across the entire STE20 family , we compared the phosphorylation rate of MSTtide to PAKtide for an additional 19 kinases , including the two MYO3 kinases that did not provide a robust PSPA signal . While PAKs detectably phosphorylated only PAKtide , all other kinases preferred MSTtide over PAKtide , albeit to varying degrees ( Fig 2C ) . Overall , these assays confirm the importance of key substrate residues selected by multiple STE20 kinase subfamilies . Kinases target specific phosphorylation site sequences through complementary interactions within the catalytic cleft . While some insight into the structural basis for kinase–substrate recognition has been obtained from crystallographic studies and site-directed mutagenesis , only a few bona fide specificity-determining residues have been experimentally validated [45–48] . To identify specific kinase residues responsible for our observed substrate selectivity , we analyzed the x-ray crystal structures of PAK4-peptide complexes [46] . In the crystal structures , the guanidino headgroup of the −2 Arg residue in the peptide occupies an acidic pocket comprising two acidic residues ( Asp444 and Glu507 ) and a polar Ser residue ( Ser443 ) ( Fig 3A ) . This pocket has previously been implicated in mediating specificity at the −2 position for PAKs and other kinases , including cAMP-dependent protein kinase ( PKA ) [45] . Glu507 is found in helix αF of the kinase domain , and an acidic residue is found at the analogous position in all STE20 kinases . All STE20 kinases also have a Trp residue ( Trp481 in PAK4 ) located two residues upstream of the conserved Asp-Pro-Glu motif within the kinase activation loop ( the APE-2 position ) that orients Glu507 through a direct hydrogen bonding interaction ( Fig 3A ) . Thus , while these residues are likely required for selection of Arg at the −2 position , they cannot function to determine specificity within this group . The other two residues in this pocket , Ser443 and Asp444 , are located within a conserved KxxN sequence in the kinase catalytic loop . Almost all STE20 family kinases other than the PAKs have small residues , Gly or Ala , at both positions , potentially creating a cavity that could accommodate the larger aromatic residues observed by PSPA analysis ( Fig 3B ) . Furthermore , these residues are highly conserved across animals , fungi , and protists: the KxxN residues of the closest PAK4 homologs from representative species were invariably Ser–Asp and were either Ala–Ala or Ala–Ser in all MST4 orthologs ( S5 Fig ) . To assess their importance in determining specificity , we generated PAK4 and MST4 mutants in which the KxxN residues were exchanged between the two kinases . We conducted both PSPA analysis and assays with a −2 Tyr substituted PAKtide . In both settings , PAK4S443A/D444A lost its strong preference for Arg and subsequently selected Tyr at the −2 position ( Fig 4A ) . Similarly , the corresponding MST4 mutant ( MST4A147S/A148D ) displayed increased activity on peptides with a −2 Arg residue . We note that mutation of the −2 interaction pocket of both kinases was accompanied by a decrease in overall catalytic rate . This effect was particularly pronounced for PAK4 , resulting in an approximately 1 , 000-fold decrease in the rate of phosphorylation of the parental PAKtide . While PAK4S443A/D444A did phosphorylate the PAKtide −2Y variant faster than did WT PAK4 , a large component of the change in selectivity appears to be conferred by loss of activity on the parental peptide . Overall , these assays confirm that the KxxN residues largely confer specificity at the −2 position in STE20 kinases . A common feature of the GCKi–v group is strong selectivity for basic residues at multiple positions downstream of the phosphorylation site . We noted that in the PAK4–peptide structure , these residues were situated proximal to the loop connecting the β3 strand and the αC helix in the kinase N-lobe ( Fig 3A ) . Among STE20 kinases , the net charge of this loop and the N-terminal region of helix αC correlated with selectivity at positions downstream of the phosphorylation site as seen by PSPA analysis ( Fig 3B ) . For example , in GCKi–v kinases , this region is rich in acidic residues , suggesting complementary interactions with basic residues in the substrate . Interestingly , OSR1 and TAO2 , which select basic residues at only one position , tend to have fewer acidic residues that are balanced by some basic residues in this region ( Fig 3B ) . Finally , PAKs carry a net positive charge in this region , consistent with their strong selection against basic residues at these positions . These trends are also evolutionarily conserved because all PAK orthologs examined had a net positive charge in this region , while the closest MST4 homologs each had between five and seven acidic residues ( S5 Fig ) . We hypothesized that attractive or repulsive electrostatic interactions serve to drive substrate specificity at the +2 to +4 positions . Exchanging the MST4 β3–αC loop with that of PAK4 ( MST458–66→RKQQRREL ) led to loss of selectivity for basic residues at these positions in the PSPA and caused a 70-fold decrease in the rate of phosphorylation of MSTtide . Furthermore , the activity of this mutant was not impaired by neutralizing the +2 and +3 basic residues of MSTtide ( Fig 4A ) . The corresponding PAK4 loop exchange mutant , PAK4355–362→EEAEDEIED , gained strong preferences for basic residues at positions +2 to +4 ( Fig 4A ) . This mutant no longer detectably phosphorylated PAKtide , but substituting its +2 and +3 residues with Arg recovered activity to within 2-fold of the WT kinase on the parental peptide . These results suggest that basic residues at positions downstream of the phosphoacceptor greatly enhance activity of STE20 kinases and establish the β3–αC loop region as a strong determinant of substrate specificity within the STE20 kinase family . The above analyses of PAK4 specificity were performed on peptide substrates , which cannot recapitulate noncatalytic site interactions found with full-length protein substrates . Furthermore , these experiments were performed under steady-state conditions that may not simulate a cellular context in which kinases may be present in excess of their substrates and in which phosphorylation often progresses to full stoichiometry . To extend these results to a true protein substrate , we performed kinase assays using the established PAK4 substrate β-catenin . The major PAK4 phosphorylation site on β-catenin , Ser675 , is found in a sequence context ( KKRLSVELT ) that conforms well to our defined PAK4 consensus motif , including basic residues at the −3 and −2 positions . To assess the importance of PAK4 catalytic site interactions , we mutated this phosphorylation site to a sequence preferred by GCKi–v kinases ( KGYNTVRRK ) . Similar to our analysis of peptide substrates , in steady-state kinetic assays , we found that this mutation reduced β-catenin phosphorylation by PAK4 to background levels ( Fig 4B ) . To examine phosphorylation under non-steady–state conditions , we performed short time-course experiments at high PAK4 concentration to capture single turnover events . Under these conditions , we observed burst kinetics in which the first turnover occurred faster than the steady-state rate ( Fig 4C ) . This observation suggests that product release is at least partially rate limiting for PAK4 phosphorylation of β-catenin . Importantly , we found that mutation of the sequence surrounding the phosphorylation site dramatically decreased the single turnover rate in addition to the steady-state rate . We conclude that catalytic site interactions accelerate a step in the kinase reaction prior to product release , either substrate association or phosphate transfer . Consequently , in the presence of a nonoptimal phosphorylation site , product release is no longer the rate-limiting step . While the sequence surrounding Ser675 appears to be nearly optimal for PAK4 , PSPA analysis suggested that changing the residue at the +1 position to a bulkier Trp residue would increase the rate of phosphorylation . Contrary to anticipation , we found PAK4 to phosphorylate the V676W mutant approximately 2-fold more slowly than WT β-catenin ( S6A Fig ) . This observation could mean that the presence of a bulky hydrophobic residue near the phosphorylation site may hinder substrate dissociation . However , we found in kinetic burst experiments that the rate of the first turnover was also decreased with this mutant ( S6B Fig ) . It may be that the optimal residue at the +1 position is context-dependent , such that the β-catenin sequence confers a preference for Val rather than Trp . Like GCKi–v kinases , all members of the protein kinase C ( PKC ) family select basic residues downstream of the phosphorylation site , primarily at the +2 position . Though PKC isozymes have low sequence similarity to STE20 kinases , their β3–αC loops are also rich in acidic residues . To determine whether this region might also confer selectivity for basic residues to PKCs , we examined a PKCβ mutant in which three acidic residues in the loop were mutated to Ala ( PKCβ3A ) . PSPA analysis showed that this mutant had significantly reduced basic preferences at multiple positions and preferred aromatic hydrophobic residues at the +2 position ( Fig 5 ) . We conclude that the β3–αC loop region can act generally as a determinant of specificity among disparate kinase groups . Reprogrammed kinase mutants can provide tools to assess how particular elements of substrate specificity contribute to their signaling output . To fully re-engineer the catalytic site specificity of MST4 and PAK4 , we combined the KxxN and β3–αC loop mutations and also exchanged a residue previously shown to mediate Ser versus Thr phosphoacceptor specificity ( the residue immediately downstream of the conserved Asp-Phe-Gly sequence in the activation loop , termed the DFG+1 residue ) [46] . Enzyme engineering frequently results in loss of catalytic activity , and we found that both exchange mutants had substantially reduced activity compared to their WT counterparts as assessed on their favored peptide substrates . To improve the activity of the PAK4 mutant , we added an additional mutation reported to increase PAK4 activity ( S445N ) [51] . The resulting compound mutants ( termed PAK4M4 and MSTP4 ) largely exchanged the substrate specificity of the two kinases at the −2 , 0 , and +1 through +4 positions as judged by PSPA analysis ( Fig 4A ) . Furthermore , these mutants inverted their respective preferences for PAKtide and MSTtide . We focused on the PAK4M4 mutant because it appeared to have a more completely reprogrammed substrate specificity in comparison to MSTP4 . PAK4M4 still had lower activity than WT PAK4 , which would complicate its use in cell-based experiments because loss of function could be attributable to reduced activity rather than altered specificity . We observed that PAK4M4 less efficiently underwent autophosphorylation within its activation loop ( S7 Fig ) , a critical step in activating the kinase . Because its activation loop phosphorylation site conforms well to the PAK4 consensus sequence ( S7A Fig ) , it is likely that reduced autophosphorylation is a consequence of the reprogrammed specificity of PAK4M4 . Incorporation of a phosphomimetic mutation at the activating site ( S474E ) better normalized the activity of PAK4M4 relative to PAK4S474E ( to 63%; S7C Fig ) . Because they had similar activity on their respective favored substrates , we proceeded to compare their ability to function in PAK- and MST-dependent processes . We initially examined whether PAK4M4 was capable of phosphorylating a set of established PAK4 substrates in cultured cells , including β-catenin , Rho guanine nucleotide exchange factor H1 ( GEF-H1 ) , and the Fer/Cip4-Bin-Amphiphysin-Rvs ( F-BAR ) protein Protein kinase C and casein kinase substrate in neurons protein 1 ( Pacsin1 ) . We found that coexpression in human embryonic kidney ( HEK ) 293A cells with PAK4S474E , but not other forms of PAK4 , caused β-catenin to accumulate to higher levels than when expressed alone ( Fig 6A ) , possibly because of its stabilization by phosphorylation at Ser675 . To determine relative levels of β-catenin Ser675 phosphorylation in cells expressing various forms of PAK4 , we isolated the protein from cell lysates and analyzed equal quantities by immunoblotting . In keeping with our in vitro kinase assays , we found that PAK4S474E robustly induced β-catenin phosphorylation at Ser675 in cells ( Fig 6A ) . By contrast , we observed no increase in phosphorylation upon coexpression with PAK4M4/S474E or with kinase-inactive mutant ( KD ) PAK4 ( PAK4D440N/S474E ) . To examine substrate phosphorylation in an endogenous setting , we performed experiments in Panc1 pancreatic cancer cells that express PAK4 to high levels [52] . Silencing PAK4 expression in Panc1 cells caused only a slight reduction of endogenous GEF-H1 phosphorylation , presumably due to compensation from other kinases ( Fig 6B ) . Importantly , re-expression of PAK4S474E , but not PAK4M4/S474E or PAK4KD , led to elevated GEF-H1 phosphorylation . Because Panc1 cells do not detectably express Pacsin1 , we examined phosphorylation of ectopically expressed protein in this system . As with other substrates , PAK4S474E alone robustly enhanced phosphorylation of Pacsin1 ( Fig 6C ) . Given the high levels of substrate phosphorylation induced by PAK4S474E , the slightly lower intrinsic kinase activity of PAK4M4/S474E ( S7C Fig ) is unlikely to underlie its complete inability to phosphorylate substrates in cells . Indeed , PAK4M4/S474E failed to induce Pacsin1 phosphorylation even when expressed to higher levels than PAK4S474E ( S8 Fig ) . Collectively , these results suggest that catalytic site specificity is essential for phosphorylation of at least some authentic PAK4 protein substrates . We next examined the contribution of phosphorylation site specificity to PAK4-dependent cytoskeletal remodeling in fibroblasts . Expression of active PAK4 causes disassembly of actin stress fibers in fibroblasts and other cell types , due at least in part to direct phosphorylation and down-regulation of GEF-H1 [22 , 53] . In NIH-3T3 cells expressing PAK4S474E , there were fewer actin stress fibers relative to either empty vector or KD ( PAK4D440N/S474E ) control ( Fig 7 ) . In contrast , PAK4M4/S474E was not significantly different from the KD PAK4 in promoting reorganization of actin fibers . Taken together , these results indicate that PAK4 phosphorylation site specificity is necessary for regulation of the actin cytoskeleton , a key signaling output of the kinase . We next asked whether phosphorylation site specificity could be sufficient for signaling from STE20 kinases . Multiple STE20 family kinases have roles as upstream regulators of the tumor-suppressive Hippo pathway . This pathway integrates signals from cell–cell contact and G-protein–coupled receptors , leading to decreased cell growth and survival [13] . Canonically , MST1 and MST2 phosphorylate and activate the kinases LATS1 and LATS2 , which themselves phosphorylate the transcription factors YAP and TAZ . Phosphorylation of YAP and TAZ induces their nuclear exclusion and proteasomal degradation . In at least some contexts , other STE20 kinases can function in place of MST1 and MST2 . For example , combined ablation of genes encoding MST1/2 and six related STE20 kinases was required to fully down-regulate Hippo signaling in HEK293A cells [16 , 18] . Notably , the LATS hydrophobic motif site conforms closely to the phosphorylation site motif recognized by GCKi–v group kinases . We therefore examined whether reprogramming the phosphorylation site specificity of PAK4 could allow it to phosphorylate LATS and function in the Hippo pathway . We confirmed that HEK293A cells lacking these eight kinases ( MM-8KO ) had substantially reduced Hippo pathway signaling as judged by LATS and YAP phosphorylation , as well as nuclear exclusion of YAP ( Fig 8 ) . Expression of PAK4M4/S474E in these cells partially restored LATS phosphorylation , albeit less efficiently than did MST1 ( Fig 8A and 8B ) . As expected , active PAK4S474E mediated GEF-H1 , but not LATS , phosphorylation . Though PAK4M4/S474E did not promote complete phosphorylation of LATS , it was sufficient to effect nearly full phosphorylation of a key regulatory site on YAP , suggesting that a low threshold of LATS activity is sufficient for Hippo signaling . A chimeric protein , in which the PAK4M4/S474E catalytic domain replaced that of MST1 , was equivalent to PAK4M4/S474E in inducing LATS phosphorylation ( S9 Fig ) , suggesting that other regions of MST1 do not promote the activity of the re-engineered kinase . We also found that MST1 and PAK4M4/S474E promoted cytoplasmic retention of YAP to the same extent , while PAK4S474E was without effect ( Fig 8C and 8D ) . Though there was substantial variability between experiments , we saw no consistent effect of GCKi–v kinase deletion or expression of PAK4 mutants on signaling through other established growth control pathways in these cells ( S10 Fig ) , suggesting that effects on YAP phosphorylation and localization are unlikely to directly involve components of these other pathways . Taken together , these results suggest that the phosphorylation site motif of a STE20 kinase is sufficient for participation in the Hippo signaling cascade .
While mechanisms of substrate targeting have been explored in depth for a number of well-studied kinases , a detailed understanding of phosphorylation site specificity has been lacking for the majority of these enzymes [6 , 42] . Here , we have systematically profiled the substrate specificity of STE20 kinases and correlated their sequence preference with specific features within the kinase catalytic domain . For example , we found that mutation of two residues within the KxxN catalytic motif could exchange specificity between PAK4 and MST4 at the −2 position . This observation is consistent with x-ray crystal structures of multiple kinase-peptide complexes , including PKA , in which an Arg residue at the −2 position makes direct contact with an acidic residue within the KxxN motif [55] . Furthermore , mutation of the analogous residue in PKCθ to Ala was reported to reduce selectivity for a −2 Arg residue [45] . Less anticipated was the identification of the β3–αC loop region as a key determinant for selecting basic residues downstream of the phosphorylation site for kinases within and outside of the STE20 family . Previous studies have implicated this region in mediating substrate selectivity . For example , mutating two residues in this loop could exchange the preferences for specific hydrophobic residues at the +2/+3 positions between type I and type II PAKs [35] . In addition , the presence of basic residues in region is at least partly responsible for substrate specificity of casein kinase 2 ( CK2 ) and glycogen synthase kinase-3 ( GSK3 ) , which select acidic or phosphorylated amino acids downstream of the phosphorylation site [48 , 56] . In addition to being involved in interactions with substrates , the positioning and dynamics of helix αC strongly influences protein kinase catalytic activity [57] . The capacity of this region to act as a general hub for substrate recognition suggests a mechanism by which adopting a more active conformation is dependent on interactions with the bound substrate . Many kinases , including PAKs , have a conserved basic residue at the N-terminus of helix αC that makes direct contact with a phosphorylated Ser or Thr residue in the kinase activation loop to promote kinase activity . By contrast , GCKi–v and PKC isozymes that have a highly acidic β3–αC loop lack this basic residue . The intermolecular interaction between basic residues in the substrate and the acidic β3–αC loop region would appear to be effectively a charge reversal that substitutes for an intramolecular contact . In keeping with this idea , the presence of basic residues C-terminal to the phosphorylation site affected the kcat rather than Km value for phosphorylation of a peptide substrate by PKCα , suggesting that engagement of the β3–αC loop by these residues indeed serves to promote catalysis [58] . We note that the residues we chose to exchange between PAK4 and MST4 were identified on the basis proximity to the substrate , as seen in crystal structures of PAK4–peptide complexes . Because crystal structures can only reveal static snapshots of these interactions , we cannot rule out additional contacts that contribute to substrate selectivity . Molecular dynamics simulations , for example , revealed a key specificity-determining residue in nonreceptor tyrosine kinases that was not evident from cocrystal structures with peptides [59] . Furthermore , computational analyses have suggested that residues not directly contacting the substrate can contribute to specificity [60 , 61] . Because our mutants often had substantially decreased catalytic activity , these other residues may function to maintain activity in the presence of particular residues in the catalytic cleft . While the largest distinction we observed in phosphorylation site specificity was between the PAK and GCK subfamilies , we could further categorize the GCK kinases into three groups . The most divergent kinase was TAO2 , which we unexpectedly found to select phosphorylated residues at several positions . TAO2 thus appears to recognize substrates that are “primed” by prior phosphorylation at nearby residues , as has been established for kinases such as CK2 and GSK3β . This phenomenon likely explains in part the sequential phosphorylation by TAO2 of two residues within the activation loops of three MAPK kinases ( MKK3 , 4 , and 6 ) . TAO2 initially phosphorylates the more downstream residue , followed by a second residue located four positions upstream [62] . This order of phosphorylation was previously rationalized based on selectivity of TAO2 for an acidic residue at the −5 position , which is mediated by a pair of basic residues located in the kinase αF–αG loop . Our PSPA analysis suggests that this initial phosphorylation event is also guided by a Thr phosphoacceptor , a hydrophobic residue at the +1 position , and an acidic residue at the +2 position . Phosphorylation at this site then primes for phosphorylation at the more upstream site by placing a phosphorylated residue at the +4 position . Such “self-priming” is frequently observed with kinases that prefer phosphorylated amino-acid residues [2] and may generally apply to TAO2 substrates . Notably , a group of direct TAO2 phosphorylation sites identified through chemical genetics [43] all possessed Ser or Thr residues at downstream positions , many of which have been observed to be phosphorylated in phosphoproteomic studies [11] . Comparative analysis of the entire STE20 family allowed us to rationally design mutants that exchange specificity between kinases with divergent phosphorylation site preferences . Previous efforts to re-engineer kinase specificity have typically focused on individual residues that determine specificity at a single position near the phosphorylation site [45 , 47 , 63 , 64] . These studies have provided insight into the structural basis of kinase specificity as well as the impact of kinase mutations occurring during evolution or in human tumors . Here , we have completely re-engineered PAK4 and MST4 to effect more radical changes in specificity . The resulting PAK4 mutant harboring the specificity profile of a GCKi–v kinase was used to investigate the contribution of phosphorylation site specificity to substrate targeting within the STE20 group . Perhaps not surprisingly , this mutant failed to phosphorylate known PAK4 targets and to induce changes to the actin cytoskeleton characteristic of the WT kinase . These results are in keeping with previous observations that kinase mutations causing more subtle perturbations in phosphorylation site specificity can also cause loss of function [46 , 47 , 65] . PAK4 targets GEF-H1 in part through a noncatalytic domain [53] . Our results suggest that this interaction alone is insufficient to effect GEF-H1 phosphorylation at Ser886 , a site that conforms very closely to the PAK4 target motif . Likewise , our kinetic analysis of β-catenin phosphorylation suggests potential docking interactions with PAK4 , which could not override the requirement for complementary catalytic site interactions . We note that in many cases , kinase phosphorylation sites match poorly to their target consensus sequences [66 , 67] . For example , while some features of the motifs we determined by PSPA are recapitulated in protein substrates ( Fig 4B ) , there are notable exceptions , including a lack of Trp residues at the +1 position for all kinases examined . In these cases , other interactions , either occurring directly to the kinase itself or through scaffold and adaptor proteins , may have a more critical role by effecting substrate recruitment . Alternatively , it may be that selectivity for particular residues is dependent on the surrounding sequence context . Indeed , we found that substitution of the residue at the +1 position in β-catenin with Trp impaired , rather than improved , its phosphorylation rate by PAK4 . In many cases it is likely that the phosphorylation site sequence has been selected to be suboptimal . Suboptimal phosphorylation site sequences confer sensitivity to perturbation , which may facilitate selective phosphorylation of more efficient kinase substrates in particular contexts [68–70] . While it is therefore possible that PAK4 substrates exist that are phosphorylated independently of its phosphorylation site motif , the inability of mutant PAK4 to reorganize the actin cytoskeleton suggests that these other substrates are insufficient to mediate at least one major signaling output of this kinase . We also note that in the system we employed , actin stress fiber disassembly required ectopic expression of constitutively active kinase , and we cannot rule out a requirement for catalytic site specificity in other functions dependent on endogenous PAK4 . By contrast to its inability to mediate PAK4 function , we found that the PAK4M4/S474E mutant was able to function in place of GCKs in the Hippo signaling pathway . This observation is consistent with the ability of GCKs from distinct subfamilies to redundantly act in the Hippo pathway despite having different domain architectures and interacting proteins . For example , the canonical Hippo kinases MST1 and MST2 are characterized by a C-terminal coiled-coil region termed the Sav-Rassf-Hpo ( SARAH ) domain , which associates with Ras-association-domain–containing protein ( RASSF ) family tumor suppressor proteins and the adaptor protein Salvador ( SAV ) [14] . By contrast , members of the GCK-I subfamily , which includes multiple additional LATS kinases , have distinct interaction partners including SH3-domain–containing adaptor proteins and bind to some substrates and regulators through C-terminal citron homology domains of unknown structure . The only similarity between the two subfamilies is that they share a kinase domain having identical specificity , suggesting that catalytic site interactions are sufficient to mediate at least some level of signaling through the pathway . We did observe that LATS was phosphorylated weakly by PAK4M4 in comparison to authentic Hippo kinases , suggesting that catalytic site interactions alone do not provide maximal activity . We note that PAK4M4 had equivalent LATS kinase activity as a chimeric protein in which the MST1 catalytic domain was replaced with that of PAK4M4 . This result argues against the involvement of noncatalytic domains and interaction partners such as SAV in promoting LATS phosphorylation . The low level of LATS phosphorylation by PAK4M4 could reflect the intrinsically low catalytic activity of PAK4 in comparison with GCKi–v kinases [46] . Another potential contributing factor is that MST kinases are reported to autophosphorylate at sites outside the catalytic domain to induce interaction with the LATS adaptor protein Mps1 binder kinase activator-like 1 ( MOB1 ) [71 , 72] . In this case , inefficient autophosphorylation by PAK4M4 or the absence of the critical MOB1 binding sites could attenuate its ability to phosphorylate LATS . We cannot exclude a role for protein localization in promoting LATS phosphorylation , and we note that both PAK and MST kinases are found at least in part at the plasma membrane , which may be important for signaling through the pathway . In summary , we have shown that catalytic site is at least to some extent both necessary and sufficient for the signaling output of some kinases in the STE20 family . These observations may seem at odds with the wealth of prior data suggesting important roles for docking and adaptor protein interactions . However , in other systems , specific elements of kinase–substrate interactions appear to confer robustness to perturbation rather than serve as a binary switch for substrate selection [68 , 69 , 73] . Furthermore , it is likely that an important role for noncatalytic site interactions is to restrict the specificity of kinases , preventing potentially deleterious phosphorylation of irrelevant proteins . We suggest that these principles of kinase–substrate recognition are thus likely to have more general relevance to other eukaryotic kinase groups outside of the STE20 family .
Primers for cloning and mutagenesis are listed in S1 Table . Gateway donor vectors for PAK4 ( Uniprot O96013-1 ) and TAO2 catalytic domain ( Uniprot Q9UL54-2 , residues 1–320 ) were made by PCR amplification of their respective cDNAs and BP recombination into pDONR221 . Donor vectors for the following kinases were obtained from the human ORFeome collection [74]: GCK ( Uniprot Q12851-2 ) , HGK ( Uniprot O95819-5 ) , HPK1 ( Uniprot Q92918-2 ) , KHS2 ( Uniprot Q8IVH8-3 ) , MINK ( Uniprot Q8N4C8-1 ) , MST2 ( Uniprot Q13188-1 ) , MST3 ( Uniprot Q9Y6E0-2 ) , LOK ( Uniprot O94804 ) , and PKCβ ( Uniprot P05771-2 ) . Mammalian transient expression constructs for WT and mutant kinases were made by LR recombination into the Gateway destination vector ( pV1900 ) derived from pCMV-Sp6 and encoding a C-terminal 3× FLAG epitope tag , with the exception of PAK4 , which was made by Gibson assembly into pcDNA3-FLAG , and MST1 ( pcDNA3-Flag-mMST1 , Addgene #1965; Watertown , MA , USA ) , which was generated by the laboratory of Joseph Avruch . The mammalian expression vector pcDNA3-FLAG-β-catenin was generated by the laboratory of Eric Fearon and obtained from Addgene ( #16828 ) , and the expression vector for Myc-tagged Pacsin1 was from the laboratory of Jeffrey Peterson . PAK4 lentiviral expression constructs were made by Gateway recombination from pDONR221 into pLX304 ( Addgene #25890 , laboratory of David Root ) . Bacterial expression constructs expressing 6× His-tagged PAK4 and MST4 catalytic domains were previously described [22 , 46] . WNK1 ( 1–661 ) , OSR1 , and MO25α in the pGEX-6P-1 backbone were from the Division of Signal Transduction Therapy , University of Dundee ( Dundee , Scotland ) . For bacterial coexpression , the Ser382 codon of WNK1 was mutated to TAG ( WNK1-SP382 ) , and the ampicillin resistance marker of the OSR1 was replaced with a zeocin resistance marker cassette . The bacterial expression construct for Yeast Sps1/Ste20-related Kinase 1 ( YSK1 ) catalytic domain ( residues 2–293 ) was generated by PCR-based subcloning of the full ORF into pCDF , followed by introduction of an appropriately placed stop codon . The bacterial expression vector producing N-terminally hexahistidine-tagged mouse β-catenin was generated by subcloning the full-length ORF into a modified pET32 plasmid . Point mutations were introduced using the QuikChange protocol ( Stratagene , San Diego , CA , USA ) . All kinase-inactive mutant controls substituted the catalytic Asp in the HRD×KxxN motif with Asn . PAK4 specificity altering mutations were combinations of the following: β3–αC loop region , R355–L362→EEAEDEIED; KxxN , S443A/D444A; DFG+1 , F461V; activating mutation , S445N . MST4 mutants were combinations of: β3–αC loop region , E58–D66→RKQQRREL; KxxN , A147S/A148D; DFG+1 , V165F . The PAK4–MST1 chimera constructs comprise residues 109–426 of PAK4 ( Uniprot O96013-2 ) followed by 322–487 of MST1 ( Uniprot Q9JI11-1 ) and were constructed by overlap extension PCR , followed by Gateway recombination into pV1900 . PAK4 constructs for expression in shPAK4-expressing cells were rendered shRNA resistant by incorporating three silent point mutations at the target site . β-Catenin plasmids for bacterial expression included two point mutations to remove minor PAK4 phosphorylation sites ( S552A/T556A ) and combinations of the following: +1W ( V676W ) , MST motif ( K672–S680→GYNTVRRKK ) , and phosphorylation resistant controls ( S675A or T675A ) . The catalytic domains of PAK4 [46] , MST4 [46] , SLK [36] , TNIK [42] , YSK1 , PAK2 [35] , PAK6 [40] , and CDC42 [22] were expressed in bacteria as previously reported , and the YSK1 catalytic domain was purified as described for MST4 . MYO3A and MYO3B , containing the kinase motor and two calmodulin binding sites as well as PAK1 , were expressed in Sf9 insect cells as previously reported [35 , 75] . Active preparations of OSR1 were prepared by coexpressing WNK1-SP382 with WT OSR1 in EcAR7 [76] cells containing SepOTS [77] , following general procedures as described previously [78] . For purification of OSR1 and MO25α expressed as GST fusion proteins , induced bacterial cell pellets from 100 mL cultures were resuspended in 5 mL of bacterial lysis buffer ( 50 mM Tris-HCl [pH 7 . 4] , 500 mM NaCl , 0 . 5 mM EDTA , 0 . 5 mM EGTA , 5 mM DTT , 1 mg/mL lysozyme , 50 mM NaF , 1 mM NaVO4 , 10% glycerol , Roche protease inhibitor tablet ) , incubated on ice for 30 min , and sonicated . Lysates were clarified by two rounds of centrifugation at 22 , 000 × g for 15 min at 4°C . The clarified lysate was transferred to 200 μL bed volume of Glutathione Hi-Cap Matrix ( Qiagen , Valencia , CA , USA ) pre-equilibrated in lysis buffer and rotated for 1 hour at 4°C . The slurry was centrifuged at 500 × g , 5 min 4°C , and the pellet was resuspended and transferred to a column and washed with 6 mL of lysis buffer without lysozyme or protease inhibitors . Proteins were eluted by rotation with 200 μL of the same buffer containing 20 U PreScission protease ( GE Healthcare , Pittsburgh , PA , USA ) at 4°C overnight with agitation , followed by washing with an additional 400 μL of lysis buffer . Eluted fractions were pooled , concentrated , and buffer exchanged into a storage buffer ( 50 mM Tris/HCl [pH 7 . 4] , 150 mM NaCl , 1 mM DTT , 20% glycerol ) using a 0 . 5 mL Amicon ultra centrifugal filter ( Millipore , Billerica , MA , USA ) , and the protein was stored at −20°C . Protein concentrations of OSR1 , MO25α , PAK4 , and MST4 were determined using BSA standards by SDS-PAGE and Coomassie staining . FLAG-epitope–tagged kinases were produced by polyethyleneimine ( PEI ) transfection of HEK293T cells and purified through batch FLAG affinity chromatography . Low-passage cells were seeded into 2 × 10 cm plates ( 9 × 105 cells/plate ) and incubated overnight . Each plate was transfected with 15 μg plasmid DNA and 45 μL of 1 mg/mL PEI ( PEI ) as previously described [79] . After incubating for 40 hours , cells were washed twice with ice-cold PBS and 1 mL mammalian cell lysis buffer ( 150 mM NaCl , 20 mM Tris [pH 7 . 5] , 1 mM EGTA , 1 mM EDTA , 1% Triton X100 , 2 . 5 mM sodium pyrophosphate , 1 mM Na3VO4 , 1 mM DTT , 1 mM PMSF , 1 mM β-glycerophosphate , 10 μg/mL leupeptin , 2 μg/mL pepstatin , 10 μg/mL aprotinin ) was added to each plate . Lysates were scraped into 1 . 5-mL tubes and incubated on ice for 10 min . After clarification in a 4°C microfuge , the supernatant was mixed with 75-μL anti-FLAG M2 beads ( Sigma Aldrich , St . Louis , MO , USA ) and rotated for 2 hours at 4°C . Beads were centrifuged and washed twice with cell lysis buffer , twice with wash buffer ( 50 mM HEPES [pH 7 . 4] , 100 mM NaCl , 1 mM DTT , 5 mM β-glycerophosphate , 0 . 1 mM Na3VO4 , 0 . 01% Igepal CA630 , 10% glycerol ) . Protein was eluted into 250 μL wash buffer containing 0 . 5 mg/mL 3× FLAG Peptide ( APExBIO , A6001; Houston , TX , USA ) at 4°C for 1 hour . The supernatant was filtered to remove beads , and kinase concentration was determined by SDS-PAGE and Coomassie staining using a BSA standard curve . For production of full-length β-catenin , 200 mL bacterial cultures were induced and grown overnight at 16°C . Pellets were resuspended in 5 mL β-catenin lysis buffer ( 20 mM Tris [pH 8 . 8] , 140 mM NaCl , 10 mM imidazole [pH 7 . 4] , 10 μg/mL pepstatin A , 10 μg/mL leupeptin , 3 mM β-mercaptoethanol , 0 . 4% Igepal CA630 , 13 mM MgCl2 , 1 mM PMSF , 200 μg/mL lysozyme ) . Cell suspensions were sonicated , DNAse I was added to 0 . 03 U/μL , and lysates were rotated at 4°C for 30 min . After clarification , the lysates were rotated with immobilized metal affinity resin ( Talon , Takara , Kusatsu , Japan ) , at 4°C for 30 min . Beads were transferred to a column and washed twice with 5 mL PBS/0 . 5% Igepal CA630 , once with 4 mL Talon wash buffer ( 20 mM Tris , 500 mM NaCl , 10 mM imidazole [pH 7 . 4] ) , and eluted in 3 mL elution buffer ( 20 mM Tris , 140 mM NaCl , 250 mM imidazole [pH 7 . 4] ) with gravity flow . Concentrated fractions were pooled and dialyzed into dialysis buffer ( 10 mM HEPES [pH 7 . 4] , 100 mM NaCl , 10% glycerol , 1 mM DTT ) overnight at 4°C . Protein concentration and purity were determined by staining of an SDS-PAGE gel with Coomassie brilliant blue and comparison to BSA standards . The PSPA ( Kinase Substrates Library , Groups I and II , Anaspec , Fremont , CA , USA ) consists of 198 peptide mixtures having the general sequence Y-A-X-X-X-X-X-S/T-X-X-X-X-G-A-K-K ( biotin ) , in which S/T is an equimolar mixture of Ser and Thr and eight of the nine X positions are a degenerate mixture of 17 residues ( all but S , T , and C ) . In each peptide mixture , one X position is fixed as one of the standard 20 amino acids , pThr , or pTyr . The library also contains three peptide mixtures that are degenerate at all X positions but have either Ser , Thr , or Tyr fixed at the central position to determine phosphoacceptor preferences . The reported PSPA assay [38 , 39] was modified as follows . Buffer ( recipes below ) was added to 1 , 536 well reaction plates ( 2 μL per well ) using a Mantis nanodispenser ( Formulatrix , Bedford , MA , USA ) . Aqueous peptides ( 250 nL ) were transferred from 384-well stock plates to the reaction plates using a Mosquito liquid handler ( TTP Labtech , Melbourn , UK ) to a final concentration of 51 μM . Kinase and ATP in reaction buffer ( 200 nL per well ) were added by Mantis to a final ATP concentration of 45 μM with 0 . 027 μCi/μL [γ-33P] ATP . The reaction plate was sealed , centrifuged briefly , and incubated at 30°C for 2 h . Reactions were spotted onto a streptavidin membrane ( Promega , Madison , WI , USA ) , which was washed , dried , and imaged as described previously [38 , 39] . Spot intensities were quantified using Quantity One software ( Bio-Rad , Hercules , CA , USA ) , and normalized by dividing by the average intensity of all spots in the same sequence position . Normalized values of replicates were averaged , converted to a log2 scale , and presented as heat maps ( created in Excel ) or enoLOGOS [44] . PSPA reactions for most kinases were run in universal kinase buffer ( 50 mM Tris [pH 7 . 5] , 10 mM MgCl2 , 2 mM MnCl2 , 1 mM DTT , 0 . 1 mM EGTA , in 0 . 1% Tween 20 ) including 820 nM PKI ( Sigma Aldrich ) . The following buffers including 0 . 1% Tween 20 were used for the indicated kinases: HGK and OSR1 ( 50 mM Tris [pH 7 . 5] , 10 mM MgCl2 , 1 mM DTT , 0 . 1 mM EGTA ) , MINK ( 50 mM HEPES [pH 7 . 4] , 10 mM MgCl2 , 1 mM DTT , 0 . 1 mM EGTA ) , MST4 ( 50 mM Tris [pH 7 . 5] , 10 mM MgCl2 , 1 mM DTT ) , PAK4 ( 50 mM HEPES [pH 7 . 4] , 10 mM MgCl2 , 12 . 5 mM NaCl , 1 mM MnCl2 , 1 mM DTT , 0 . 1 mM EGTA ) , PKCβ ( 50 mM Tris [pH 7 . 4] , 10 mM MgCl2 , 1 mM DTT with 20% v/v PKC lipid activator ) . Kinase concentrations ranged from 10–37 nM with the exception of MST4 ( 320–660 nM ) , PAK4 ( 190 nM–2 . 1 μM ) , WT PKCβ ( 6 nM ) , and PKCβ3A ( 60 nM ) . OSR1 reactions included 82 nM GST-MO25 . Each kinase was assayed twice , and correlation constants ( R2 ) for position-normalized data between replicate runs were >0 . 6 for all but MST3 , which was run two additional times . Peptide substrates were synthesized at the Tufts University Core Facility and purified by HPLC . Kinase , reaction buffer , and peptide ( 15 μM for PAK4 and 20 μM for all other kinases ) were mixed and reactions were started by addition of ATP to a final concentration of 10 μM with 0 . 05 μCi/μL [γ-33P] ATP at a final volume of 20 μL . Reactions were incubated at 30°C , and 5 μL aliquots were taken at 5–7 min intervals , spotted onto P81 phosphocellulose paper filters , and quenched in 75 mM phosphoric acid . Filters were washed 3 × 5 min with 75 mM phosphoric acid , rinsed briefly with acetone , dried , and quantified by scintillation counting . Reaction buffers were the same as used for peptide library assays except that Tween 20 was excluded . PAK1 and PAK2 were assayed in MINK buffer containing 100 nM GTP-γ-S–loaded CDC42 . MYO3A and MYO3B were assayed in ( 10 mM imidazole [pH 7 . 4] , 50 mM KCl , 1 mM EGTA , 1 mM MgCl2 , 1 mM DTT ) . Kinase reactions were performed at least three times . Full-length PAK4 and full-length β-catenin were assayed in PAK4 buffer and initiated by addition of [γ-33P] ATP ( 10 μM , 0 . 05 μCi/μL ) . For steady-state reactions , the PAK4 concentration was 10 nM , and β-catenin concentration was 5 μM . Samples were taken at 5 , 10 , and 15 min . For single turnover reactions , PAK4 concentration was 500 nM , β-catenin substrates were 1 . 5 μM , and samples were taken at 5 , 10 , 15 , 20 , 30 , 60 , 90 , 120 , 150 , 240 , and 300 seconds . For both steady-state and pre-steady–state experiments , aliquots taken at each time point were quenched with SDS-PAGE loading buffer , boiled , and fractionated by SDS-PAGE . Gels were stained with Coomassie brilliant blue , dried , and exposed overnight to a phosphor storage screen ( Kodak , Rochester , NY , USA ) . Screens were scanned using a phosphorimager ( Bio-Rad ) , and band intensities were quantified using Quantity One software ( Bio-Rad ) . Panc1 pancreatic cancer cells harboring inducible shRNAs were generated by lentiviral transduction as follows . Lentiviral particles containing inducible shRNA were produced in low-passage HEK293T cells by PEI co-transfection of dR8 . 91 , VsV-G , and PAK4-targeting pTRIPZ #395103 ( mature antisense: TGAAGAGCAGCTCGCGCCT; Dharmacon , Lafayette , CO , USA ) or luciferase-targeting control vector in a 10:1:10 ratio . Supernatants were collected at 24 and 36 hours post-transfection . Panc1 cells were infected with lentivirus at an MOI of approximately 1 in the presence of 8 μg/mL polybrene . After 24 hours , infected cells were selected with 2 μg/mL puromycin . After 64 hours selection , cells were maintained in 1 μg/mL puromycin , and knockdown was induced with 500 ng/mL doxycycline at least 7 days prior to the experiments below . To assess phosphorylation of PAK4 substrates , Panc1 cells expressing either shPAK4 or shLuc or HEK293A cells were transfected with the indicated PAK4 expression plasmids either alone or in combination with a 3-fold excess of the indicated substrate expression vector ( unless otherwise indicated ) using Lipofectamine 2000 as recommended by the manufacturer . Media were exchanged 6 hours later , and cells were incubated for an additional 20 hours ( HEK293A cells ) or 40 hours ( Panc1 cells ) . For Panc1 cells , media was then aspirated , and cells were washed once with serum-free media and incubated in media containing 0 . 1% FBS for 2 hours prior to lysis . Cells were transferred to ice , washed once with cold PBS , and extracted into mammalian cell lysis buffer ( defined above ) lacking DTT . Following incubation for 10 min on ice , lysate was cleared by microcentrifugation ( 10 min , 4 ºC ) . Cleared lysates were analyzed by BCA protein assay ( Thermo Fisher Scientific , Waltham , MA , USA ) , and equal amounts of protein were analyzed by SDS-PAGE followed by immunoblotting . FLAG-β-catenin was isolated from HEK293A cell lysates as described above for kinase purification , and relative quantities of recovered protein were determined by immunoblotting samples with FLAG antibody ( Sigma M2 , #F3165; Sigma Aldrich ) . Samples containing equal amounts were reanalyzed by immunoblotting with the indicated antibodies . Antibodies used were raised against the following: GEF-H1 pSer886 ( Cell Signaling Technology #14145; Danvers , MA , USA ) , GEF-H1 ( Cell Signaling Technology #4076 ) , Pacsin1 pSer346 ( Millipore #ABS39 ) , Myc epitope ( Cell Signaling Technology 9B11 , #2276 ) , β-catenin pSer675 ( Cell Signaling Technology #9567 ) , vinculin ( Sigma #V9131 ) , and PAK4 ( BD Pharmingen , E40-883; Franklin Lakes , NJ , USA ) . Membranes were imaged using fluorophore-conjugated secondary antibodies ( Alexa Fluor 680 goat anti-rabbit IgG and IRDye 800CW donkey anti-rabbit IgG , both at 1:10 , 000 dilution ) on an Odyssey CLx imager ( LI-COR , Lincoln , NE , USA ) . Lentivirus expressing PAK4 variants was produced by co-transfecting HEK293T cells with the corresponding pLX304 vector , as well as psPAX2 and VsV-G packaging and envelope plasmids . NIH-3T3 cells were infected by treatment with lentivirus with 8 μg/mL polybrene for 24 h . Cells were then exchanged into fresh media and selected with blasticidin ( 5 μg/mL for 8 days and in 2 . 5 μg/mL thereafter ) . To visualize actin fibers , cells were blinded and sparsely plated onto coverslips in 24-well plates ( 5 , 000 cells/well ) . After 24 hours , cells were washed twice with PBS and fixed in 4% paraformaldehyde for 15 min at room temperature ( RT ) . Cells were washed three times with PBS and permeabilized with 0 . 1% Triton X100 for 30 minutes at RT . Coverslips were blocked in blocking buffer ( 1% BSA , 50 mM NH4Cl , 0 . 1% Triton X100 in PBS ) for 1 hour and then incubated with an anti-vinculin antibody ( Sigma #V9131; Sigma Aldrich ) diluted 1:5 , 000 in PMZ buffer ( 0 . 2% BSA , 50 mM NH4Cl , 0 . 1% Triton X100 in PBS ) for 1 h . After three PBS washes , coverslips were incubated with Alexa Fluor 568 goat anti-mouse antibody ( 1:500 , Invitrogen A11004; Carlsbad , CA , USA ) and Alexa Fluor 647 phalloidin ( 1:500 , Invitrogen A22287 ) in PMZ buffer for 30 min in a dark , humidified chamber . Coverslips were washed three times with PBS and mounted onto slides using ProLong Diamond mounting media containing 5 μg/mL 4′ , 6-diamidino-2-phenylindole ( DAPI , Invitrogen ) and dried overnight in the dark . Cells were imaged at 100× and 40× magnification using a Nikon Eclipse Ti-S microscope ( Nikon , Tokyo , Japan ) , and at least 80 fields with individual cells were captured for each cell line . The entire procedure was performed a total of three times with two different infections . Images were randomly selected from each condition and image set , and blindly scored for actin disassembly using a 7-point scale , with lower values indicating a less extensive stress fiber network . The number of fibers per cell was also quantified in blinded images using SFEX for automated fiber identification [54] . Parental HEK293A and MM-8KO cells lacking MST1 , MST2 , KHS2 , GCK , HPK1 , HGK , TNIK , and MINK [16] were obtained from the laboratory of Kun-Liang Guan . Cells were transfected with vectors expressing the indicated kinases using Lipofectamine 2000 as recommended by the manufacturer . After 16 hours , cells were chilled on ice , washed once with ice-cold PBS , and then mechanically released from the plate into PBS . Cells were gently pelleted at 4°C and suspended in 100 μL of cell lysis buffer ( see above ) . Lysates were clarified by centrifugation , and 4× SDS-PAGE loading buffer was added to the supernatant . Samples were boiled , fractionated by SDS-PAGE , and transferred to PVDF membranes . Membranes were probed with primary antibodies listed above and others from Cell Signaling Technology ( LATS1 pThr1079 #8654 , YAP pSer127 #4911 , YAP #4912 , Akt pSer473 D9E #4060 , AKT #9272 , phospho-ERK #4370 , ERK #9102 , S6K1 pThr389 #9205 , S6K1 #9202 ) , followed by the appropriate secondary antibodies , and imaged on the Odyssey CLx ( LI-COR ) . The phosphorylation index was determined by dividing the phospho-signal to total signal , and samples were normalized to the signals from HEK293A parental cells transfected with empty vector . Experiments were performed at least three times . For YAP localization experiments , 2 . 5 × 105 cells were plated in each well of a 12-well plate . After 24 hours , cells were transfected in duplicate with the indicated vectors in a blinded manner using Lipofectamine 2000 . After 3–4 hours , cells were trypsinized , and 1 . 0 × 105 cells were plated onto fibronectin-coated coverslips and incubated overnight . Cells were serum starved for 1 hour before fixation with 4% paraformaldehyde in PBS for 15 min at RT . Immunofluorescence labeling was performed as for actin stress fiber visualization using anti-YAP ( Santa Cruz Biotechnology H-125 , sc-15407; Dallas , TX , USA ) and anti-FLAG ( Sigma , M2; Sigma Aldrich ) antibodies . Coverslips were imaged at 40× and 100× using a Nikon Eclipse Ti-S microscope ( Nikon ) , with at least 20 FLAG-positive fields captured for each transfection . Images were also analyzed for YAP localization using CellProfiler software [80] , using DAPI and FLAG staining to define the nucleus and whole cell , respectively . Unblinding of samples occurred subsequent to automated analysis . This experiment was performed five times . | Protein kinases , which catalyze the transfer of phosphate from ATP to substrate proteins , are important enzymes in cellular signal transduction pathways mediating responses to extracellular cues . In order to function properly in signal transmission , each kinase must phosphorylate only a limited number of proteins among the thousands present within the cell . One way that kinases choose their substrates is by recognition of amino-acid sequence motifs surrounding the site of phosphorylation , but kinases can also recruit substrates through docking interactions occurring outside of the catalytic cleft . For most kinases , the relative contribution of catalytic site and docking site interactions to substrate selection is not known . Here , we investigated the phosphorylation site specificity of a group of 30 human protein serine–threonine kinases called the STE20 family . Guided by x-ray crystal structures of kinase-peptide complexes , we re-engineered the catalytic clefts of two members of this group , PAK4 and MST4 , to exchange their phosphorylation site motifs . In cells , the re-engineered form of PAK4 was unable to phosphorylate substrates of wild-type PAK4 or to carry out its normal function in reorganizing the actin cytoskeleton . In contrast , it was able to function in place of MST4 and related kinases in the growth-controlling Hippo signaling pathway . Overall , these studies suggest that catalytic site interactions can be both necessary and sufficient for kinase function . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"phosphorylation",
"protein",
"interactions",
"molecular",
"probe",
"techniques",
"enzymes",
"enzymology",
"immunoblotting",
"plasmid",
"construction",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
"dna",
"construction",
"research",
"and",
"analysis",
"methods",
"protein",
"kinase",
"signaling",
"cascade",
"sequence",
"analysis",
"contractile",
"proteins",
"actins",
"bioinformatics",
"proteins",
"protein",
"kinases",
"molecular",
"biology",
"biochemistry",
"signal",
"transduction",
"cytoskeletal",
"proteins",
"cell",
"biology",
"post-translational",
"modification",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"cell",
"signaling",
"signaling",
"cascades"
]
| 2019 | Comprehensive profiling of the STE20 kinase family defines features essential for selective substrate targeting and signaling output |
Severe fever with thrombocytopenia syndrome ( SFTS ) is an emerging infectious disease that was recently identified in China , South Korea and Japan . The objective of the study was to evaluate the epidemiologic and clinical characteristics of SFTS in South Korea . SFTS is a reportable disease in South Korea . We included all SFTS cases reported to the Korea Centers for Disease Control and Prevention ( KCDC ) from January 2013 to December 2015 . Clinical information was gathered by reviewing medical records , and epidemiologic characteristics were analyzed using both KCDC surveillance data and patient medical records . Risk factors for mortality in patients with SFTS were assessed . A total of 172 SFTS cases were reported during the study period . SFTS occurred throughout the country , except in urban areas . Hilly areas in the eastern and southeastern regions and Jeju island ( incidence , 1 . 26 cases /105 person-years ) were the main endemic areas . The yearly incidence increased from 36 cases in 2013 to 81 cases in 2015 . Most cases occurred from May to October . The overall case fatality ratio was 32 . 6% . The clinical progression was similar to the 3 phases reported in China: fever , multi-organ dysfunction , and convalescence . Confusion , elevated C-reactive protein , and prolonged activated partial thromboplastin times were associated with mortality in patients with SFTS . Two outbreaks of nosocomial SFTS transmission were observed . SFTS is an endemic disease in South Korea , with a nationwide distribution and a high case-fatality ratio . Confusion , elevated levels of C-reactive protein , and prolonged activated partial thromboplastin times were associated with mortality in patients with SFTS .
Severe fever with thrombocytopenia syndrome ( SFTS ) is an emerging infectious disease that is caused by a novel SFTS virus ( SFTSV ) which was first reported in China in 2011 [1] . China’s neighboring two countries , South Korea and Japan , have also reported the infection [2 , 3] . The virus is transmitted to humans through tick bites , and Haemaphysalis longicornis is known to be a main vector [1 , 4] . The clinical manifestations of SFTS include fever , myalgia , vomiting , diarrhea , thrombocytopenia and leukopenia . In severe cases , multi-organ dysfunction may occur [5]; the case fatality ratio ( CFR ) has been reported to be 6 . 3–30% [6 , 7] . In South Korea , the first patient with SFTSV infection was identified in 2012 [2] . Subsequently , an epidemiologic study reported that 35 cases of SFTS , with a CFR of 45 . 7% , occurred in 2013 [8] . The objective of this study was to characterize the epidemiologic and clinical findings of all SFTS patients since the first case was reported in South Korea .
SFTS has been a reportable disease in South Korea since 2013 . We included all SFTS cases reported to the Korea Centers for Disease Control and Prevention ( KCDC ) from 2013 to 2015 [9 , 10] . Confirmatory tests of SFTSV infection were performed at KCDC by detecting M segment gene of SFTSV RNA using one-step reverse transcription polymerase chain reaction ( RT-PCR ) or antibody tests with immunofluorescence assay ( IFA ) to detect the seroconversion of paired sera for anti-SFTSV immunoglobulin G , as previously described [8 , 11] . To collect clinical data , we reviewed the medical records of patients who had available epidemiologic information , clinical manifestations and laboratory findings . Epidemiologic characteristics were also supplemented by reviewing the epidemiology investigation records provided by the KCDC . Demographic factors , date of onset , history of tick bite , presence of bite wound , and comorbidity were included . History of tick bite was self-reported and collected from medical records . The locations of possible exposure to SFTSV were determined considering the patient’s history of outdoor activities or residential scope in one month prior to the onset of illness [8] . We used the KCDC data for the location of SFTS acquisition . Each case was coded according to its geographic location at the district level and was positioned on a map of South Korea ( http://www . gadm . org ) . The national and regional incidences of SFTS per 100 , 000 person-years from 2013 to 2015 were calculated using the national census ( http://kosis . kr ) . We divided the clinical course of SFTS into three stages by week , and a comparison of clinical and laboratory features was performed for the fatal and non-fatal cases occurring in each period . Worst values were selected for the data in each patient if there were multiple measurements during the unit period . Meningoencephalitis was defined as a white blood cell count of the cerebrospinal fluid >5 cells/mm3 . Myocarditis was determined by an abnormal electrocardiography ( ECG ) , serum levels of troponin or creatine phosphokinase ( CK ) fractions and an echocardiogram . Arrhythmia was defined as either a new onset during the course of illness or a previously undiagnosed new case . Acute kidney injury was defined as serum creatinine levels ≥2 mg/dl and 2 times the baseline levels [5 , 12] . The Acute Physiology and Chronic Health Evaluation ( APACHE II ) score was also calculated [13] . Severe thrombocytopenia was defined as platelet count <50x103/mm3 in view of its implication for the critical threshold that a risk of spontaneous bleeding increases [5 , 14] . The prolongation of activated partial thromboplastin time ( aPTT ) was defined as aPTT >60 sec , indicating >50% of upper normal value ( reference value of aPTT <40 sec ) [5 , 15] . Statistical analyses were performed using Pearson’s chi-square test or Fisher’s exact test to analyze the relationships between categorical variables and severity of disease . Two-sample t-tests or Mann-Whitney U-tests were used to compare the continuous variables between fatal and non-fatal cases . Longitudinal analysis for serial clinical feature and laboratory parameters was performed using a multivariable generalized estimating equation with binomial variable and linear mixed model with continuous variable . The risk factors for mortality in patients with SFTS were analyzed by binary logistic regression . P values <0 . 05 were considered statistically significant . Variables having P values <0 . 05 in the univariate analysis were used for a multivariate stepwise logistic regression analysis . We sought sensitivity , specificity , positive predictive value ( PPV ) , and negative predictive value ( NPV ) of single variable from 1st week after the onset of illness which was significant in univariate analysis . We also analyzed sensitivity , specificity , and C-statistics of combined variable from 1st week after the onset of illness which were significant in the multivariate analysis ( SPSS 22 . 0; SPSS Inc . , Chicago , IL , USA ) . This study was approved by the institutional review board ( IRB ) of Boramae Medical Center ( #15-2015-123 ) . All the institutions participating in the clinical network also obtained approval from their IRBs . Personal information was de-identified before collection and the anonymized data were processed by different analyzers . All clinical investigations were conducted according to the principles expressed in the Declaration of Helsinki .
A total of 170 SFTS cases were reported to the KCDC during the study period of 2013–2015 [9] . Of the 170 SFTS cases , 161 were confirmed by RT-PCR and 9 patients by IFA . We also included 2 additional cases that had a negative conventional RT-PCR but a positive real-time RT-PCR . Therefore , a total of 172 cases of SFTS were included in this study . The yearly incidence of SFTS was 36 cases in 2013 ( including one case in 2012 ) , 55 cases in 2014 , and 81 cases in 2015 . The case fatality ratio was 47 . 2% ( 17/36 ) in 2013 , 32 . 7% ( 18/55 ) in 2014 , and 25 . 9% ( 21/81 ) in 2015 , with an overall CFR of 32 . 6% ( 56/172 ) . The seasonal distribution of the SFTS cases is shown in Fig 1 . Most of the cases occurred between May and October . There were no cases from December to March . The geographical distribution of the SFTS cases is shown in Fig 2 . SFTS occurred throughout South Korea , with the exception of urban areas . The incidence was relatively low in the western and southwestern rice field areas and the scarcely populated eastern mountainous areas . Hilly areas were the major endemic regions . The overall incidence was 0 . 11 cases/105 person-years . Specifically , Jeju province ( 23 cases , 1 . 26 cases/105 person-years ) showed the highest incidence rate , followed by Gangwon ( 26 cases , 0 . 54 cases/105 person-years ) , Gyeongbuk ( 33 cases , 0 . 41 cases/105 person-years ) , and Jeonnam province ( 13 cases , 0 . 23 cases/105 person-years ) ( Fig 2 ) . Of the 120 patients , 37 ( 31 . 1% ) were placed on mechanical ventilation . Acute renal failure ( 14 . 2% ) , meningoencephalitis ( 13 . 8% ) , new-onset arrhythmia ( 11 . 8% ) , and myocarditis ( 4 . 2% ) were common complications during the hospital course ( Table 1 ) . Forty-six patients died , resulting in a CFR of 38 . 3% . The median time from onset of illness to death was 9 . 5 days ( IQR , 7–15 days ) ; 41 . 3% ( 19/46 ) of the non-surviving patients were male , and the median age was 73 . 5 years ( IQR , 66–79 years ) . Of the surviving patients , 56 . 8% ( 42/74 ) were male , and the median age was 66 years ( IQR , 52–74 years ) , which was lower than that of the fatal group ( p < 0 . 001 ) ( Table 1 ) . The median time from the onset of illness to hospital discharge was 16 days ( IQR , 13–23 days ) in the non-fatal group . Among the clinical parameters in the 1st week , dyspnea , gastrointestinal bleeding , and confusion were associated with death ( Table 2 ) . In terms of the categorical clinical features , the frequency of central nervous symptoms was higher in the fatal group ( p = 0 . 025 ) ( S3 Table ) . In the univariate analysis of laboratory parameters in week 1 , severe thrombocytopenia ( <50×103/mm3 ) was more common in the fatal group ( 43 . 1% vs 70 . 6% , p = 0 . 012 ) . Anemia , increases in serum alkaline phosphatase , AST ( >400 IU/L ) , ALT ( > 200 IU/L ) and CRP ( mg/dL ) , and prolongation of prothrombin time ( PT ) ( INR ≥1 . 3 ) and aPTT ( >60 sec ) were associated with death ( Table 2 ) . APACHE II scores were higher in the fatal group than in the non-fatal group ( 15 vs 23 , p <0 . 001 ) . Changes of clinical and laboratory variables over 3 weeks between the two groups were significantly different for the variables of confusion ( p < 0 . 001 ) , respiratory and cardiovascular symptoms ( p = 0 . 022 ) , platelet ( p < 0 . 001 ) , AST ( p = 0 . 005 ) , CRP ( p < 0 . 001 ) , serum creatinine ( p < 0 . 001 ) , and LDH ( p < 0 . 001 ) ( S2 , S3 and S4 Tables & Fig 3 ) . In the multivariate regression analysis , a mental state of confusion , elevated levels of CRP , and the prolongation of aPTT were associated with mortality ( Table 2 ) . The single variable for the highest value of sensitivity ( leukopenia , 97 . 1% ) , specificity ( prolongation of PT , 94 . 3% ) , PPV ( prolongation of PT , 78 . 6% ) , and NPV ( elevated ALT , 80 . 0% ) was not uniform ( S5 Table ) . With combined variables , ‘confusion + aPTT >60 sec’ showed the maximal C-statistics value of 0 . 786 ( 95% confidence interval , 0 . 625–0 . 948 ) ( S6 Table ) .
This study showed that SFTS occurred throughout South Korea . The overall incidence of SFTS in South Korea was 0 . 11 cases per 105 person-years , which was lower than that in China ( 0 . 12–0 . 73 cases per 105 person-years ) [16] . The median age of 69 . 0 years in our patients is higher than that in China , 57 . 6 years [17] , which reflects the difference of aging populations between two countries , especially in rural areas . We also showed that summer was a period of peak transmission . A previous study suggested that H . longicornis was the predominant species ( 90 . 8% ) and was widely distributed in a nationwide surveillance of ticks [18] . Although H . longicornis is the major vector , other tick species such as Haemaphysalis flava , Amblyomma testudinarium and Ixodes nipponensis have also been reported to carry SFTSV in South Korea [19 , 20] . The peak transmission in summertime may be attributable to the seasonal life cycle of H . longicornis [18 , 21] and the increased rates of outdoor activity during this season . The western and southwestern areas of South Korea showed a relatively low incidence of SFTS , which might be related to the low SFTSV infection rate in ticks [18] . This is evidenced by the contrasting finding of a high incidence of tsutsugamushi in the same area , which is closely related to people’s outdoor activity and is one of the most common notifiable infectious diseases in South Korea [22 , 23] . The main epidemic season of tsutsugamushi in South Korea is autumn , when chigger mites begin sucking blood from mammals to meet the developmental needs of their life cycle . Several SFTS patients were found to be infected in urban areas . However , 4 cases in Seoul and 5 cases in Wonju were from nosocomial outbreaks . All other cases reported in urban areas were from major cities such as Daegu , Ulsan , Busan and Gwangju , and these cases occurred in the rural suburbs of the city border . Therefore , no endemic cases in inner city were found in South Korea . The overall CFR in our study was 32 . 6% which was compared with 12 . 2% reported recently in China [6] . Although the CFRs of consecutive three years has been decreasing , the CFRs may still be exaggerated . Since SFTS is a newly emerging infectious disease in South Korea , more inclusive screening criteria and education will continue to find more patients with milder presentation . But the patients with mild presentations may have so short or no viremic period that real-time RT-PCR for serum cannot prove the presence of SFTSV effectively . The easy accessibility to well-performing serologic method is needed to help the diagnosis for such patients group . The annual increase of SFTS incidence may be interpreted as a result of increased surveillance or awareness otherwise not detected in the past . We need to follow the trend of SFTS incidence further . The clinical course of our patients was consistent with the 3 previously reported clinical stages of fever , multi-organ dysfunction and convalescence [5] . Our study identified fever , thrombocytopenia and leukopenia in more than 90% of the patients in the 1st stage , and high fever lasted for a median of 6–11 days . Chinese data showed a high-fever period of 5–11 days [7] . In the 2nd stage , biomarkers including AST , ALT and aPTT were elevated to maximum levels . However , there were notable differences in these values between the fatal and non-fatal groups ( Fig 3 ) . During the convalescence stage , the clinical symptoms of SFTS patients began to resolve from 8 to 11 days after the onset of illness , and the laboratory parameters gradually returned to their normal ranges . In China , the convalescence stage began approximately 11–19 days after disease onset , and the biochemical measurements returned to normal within approximately 3–4 weeks in survivors [7] . The overall types of clinical parameters and their changes over time were similar in South Korea and China . In our study , the median time from the onset of illness to death was 9 . 5 days . The levels of important biomarkers including hemoglobin , platelets , PT , aPTT , serum ALP , AST , ALT and CRP and clinical parameters such as age , dyspnea , gastrointestinal bleeding and confusion all showed significant differences between the fatal and non-fatal groups . Previous studies found that age , neurologic manifestations , hemorrhagic signs , thrombocytopenia , and elevations of AST , CK or LDH were related to mortality in patients with SFTS in univariate analyses [5 , 24] . Neurologic manifestations , thrombocytopenia , prolongation of aPTT , hypoalbuminemia and hyponatremia were significant prognostic factors in multivariate analyses [6 , 21 , 25] . In our study , confusion , elevated levels of CRP , and prolonged aPTT were associated with the death of patients with SFTS in the multivariate analysis . Overall , mental status and hemorrhagic tendency seemed to be the predominant factors closely related to prognosis . We sought the sensitivity/specificity and PPV/NPV for single or combined variables which were significant in the univariate and multivariate analysis . Confusion and aPTT were important predictors for death . Elevated CRP as either continuous or categorical ( >3 mg/dL ) variable was one of significant predictors for death from several analyses in our study . Secondary infections like pneumonia or catheter related bloodstream infection might complicate the fatal patients . 31 . 1% of patients underwent mechanical ventilation during the hospital course . But we couldn’t collect precise data for the infectious complications . Two hospital outbreaks of SFTS occurred , and 9 healthcare workers ( HCWs ) were infected [11 , 26] . In one hospital , 4 of the 27 HCWs who had contacted with an index patient during cardiopulmonary resuscitation were diagnosed with SFTS via seroconversion [11] . In another hospital , 5 of the 27 HCWs who had been exposed to blood and body fluids were also diagnosed with SFTS via seroconversion [26] . All of the healthcare workers had mild or asymptomatic infections . Nosocomial outbreaks were also reported in China , and not only doctors and nurses but also family members and mortuary beauticians were infected . Therefore , standard precautions should be strictly implemented when caring for suspected SFTS patients [27–29] . Several diseases should be mentioned in the differential diagnosis in South Korea . Human granulocytic anaplasmosis ( HGA ) , which is a tick-borne disease , has a similar clinical presentation including fever , thrombocytopenia and leukopenia . The first human case of HGA in South Korea was reported in 2014 [30] . A serologic survey of the blood samples submitted for SFTS tests showed a 2 . 2% positivity for anaplasmosis [31] . A . phagocytophilum has been detected in H . longicornis , I . nipponensis and I . persulcatus ticks [32 , 33] . Human monocytotrophic ehrlichiosis also has a similar clinical presentation . Ehrlichiosis was reported in an active duty soldier [34] , and its causative agent , Ehrlichia chaffeensis , was identified in ticks in South Korea [32] . Hemophagocytic lymphohistiocytosis ( HLH ) is an aggressive and life-threatening disease which is triggered commonly by infection . Its major diagnostic criteria include fever ( ≥38 . 5°C ) , thrombocytopenia , neutropenia , hyperferritinemia ( >3 , 000 ng/mL ) , CNS symptoms , hepatitis , coagulopathy and hemophagocytosis in bone marrow [35] . These findings were not uncommon in our data ( S3 & S4 Tables ) . Although we did not observe bone marrow findings , hemophagocytosis in the bone marrow of SFTS patients has been reported as one of the key findings [36–38] . As a cause of secondary HLH , SFTS needs to be considered in endemic areas . Specific antiviral therapies are urgently needed considering the high fatality and widespread prevalence of SFTS in northeast Asian regions . This study did not analyze the effects of antiviral treatment for SFTS because of the limited number of cases . Several regimens have been tested based on individual physician’s decisions , including a combination of plasma exchange and ribavirin administration [39] , plasma exchange followed by convalescent plasma therapy [40] , and combination of intravenous immunoglobulin and corticosteroid [41] . Ribavirin showed in vitro antiviral effects against SFTSV in a dose-dependent manner [42] . However , a clinical study found that ribavirin monotherapy was not effective [43] . Recent studies suggest that favipiravir and a combination of ribavirin and interferon may be effective in treating SFTS infection [44 , 45] . This study has some limitations . As the study data were retrospectively collected from many study sites , the clinical variables affecting the risk factors for death could be incompletely assessed . But most of the SFTS patients were referred to infectious disease physicians , and usually the patients had intensive medical scrutiny with frequent laboratory evaluations . We presented the epidemiologic features of all 172 patients who had SFTS in a period of 3 years . However , in the clinical analysis , we excluded 52 patients . The two groups ( 120 vs 52 patients ) showed significant differences in age distribution and CFR . This is because the group of 52 patients included the nosocomial outbreak cases , which occurred in younger ages and had a mild presentation . In conclusion , SFTS is a prevalent endemic disease in South Korea that has a high case-fatality ratio . The clinical manifestations were similar to those reported in China . Confusion , elevated levels of C-reactive protein , and prolonged activated partial thromboplastin times were associated with death in patients with SFTS . The development of effective therapeutics for SFTSV infection is urgently needed . | Severe fever with thrombocytopenia ( SFTS ) is an emerging infectious disease that was first discovered in China in 2009 . Subsequently , SFTS has also been found in South Korea and Japan . Here , we report the epidemiologic and clinical characteristics of 172 confirmed SFTS cases in South Korea that occurred since the first case was reported in 2013 . SFTS was prevalent throughout South Korea , except for in urban areas . The incidence was relatively low in the western and southwestern rice field areas and the scarcely populated eastern mountainous area . Hilly areas were the major endemic regions . The incidence was increasing annually , and the case fatality ratio was 32 . 6% . A mental status of confusion , elevated levels of C-reactive protein , and prolongation of activated partial thromboplastin time were associated with mortality in patients with SFTS . Two outbreaks of nosocomial SFTS transmission were noted . | [
"Abstract",
"Introduction",
"Patients",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"c-reactive",
"proteins",
"clinical",
"laboratory",
"sciences",
"ixodes",
"pathology",
"and",
"laboratory",
"medicine",
"china",
"geographical",
"locations",
"animals",
"signs",
"and",
"symptoms",
"nosocomial",
"infections",
"ticks",
"infectious",
"diseases",
"thrombocytopenia",
"proteins",
"clinical",
"laboratories",
"epidemiology",
"hematology",
"disease",
"vectors",
"arthropoda",
"people",
"and",
"places",
"biochemistry",
"arachnida",
"diagnostic",
"medicine",
"asia",
"south",
"korea",
"fevers",
"biology",
"and",
"life",
"sciences",
"organisms"
]
| 2016 | Severe Fever with Thrombocytopenia Syndrome in South Korea, 2013-2015 |
The sleep onset process ( SOP ) is a dynamic process correlated with a multitude of behavioral and physiological markers . A principled analysis of the SOP can serve as a foundation for answering questions of fundamental importance in basic neuroscience and sleep medicine . Unfortunately , current methods for analyzing the SOP fail to account for the overwhelming evidence that the wake/sleep transition is governed by continuous , dynamic physiological processes . Instead , current practices coarsely discretize sleep both in terms of state , where it is viewed as a binary ( wake or sleep ) process , and in time , where it is viewed as a single time point derived from subjectively scored stages in 30-second epochs , effectively eliminating SOP dynamics from the analysis . These methods also fail to integrate information from both behavioral and physiological data . It is thus imperative to resolve the mismatch between the physiological evidence and analysis methodologies . In this paper , we develop a statistically and physiologically principled dynamic framework and empirical SOP model , combining simultaneously-recorded physiological measurements with behavioral data from a novel breathing task requiring no arousing external sensory stimuli . We fit the model using data from healthy subjects , and estimate the instantaneous probability that a subject is awake during the SOP . The model successfully tracked physiological and behavioral dynamics for individual nights , and significantly outperformed the instantaneous transition models implicit in clinical definitions of sleep onset . Our framework also provides a principled means for cross-subject data alignment as a function of wake probability , allowing us to characterize and compare SOP dynamics across different populations . This analysis enabled us to quantitatively compare the EEG of subjects showing reduced alpha power with the remaining subjects at identical response probabilities . Thus , by incorporating both physiological and behavioral dynamics into our model framework , the dynamics of our analyses can finally match those observed during the SOP .
There have been numerous ways in which scientists have attempted to measure behavioral and physiological dynamics during the SOP . For an in-depth look at the methods employed in the past , see Ogilvie's review paper [1] , which comprehensively details the many multimodal correlates of sleep onset and the experimental strategies employed in characterizing them . Ogilvie divides behavioral metrics of sleep onset into categories of active and passive behavioral measurement . Active metrics involve tasks with repeated externally-generated probes for wakefulness , each of which prompts the subject for a physical response via button press or verbal response . Additionally , response via cued respiration has been used for experimental and interventional behavioral paradigms [7] , [8] . These active probes could include subjective queries [9]–[11] , or auditory [12]–[14] and vibratory [15] stimuli . Use of a psychomotor vigilance task ( PVT ) derived metric [16] has also been proposed . Active methods are useful , as repeated trials yield multiple measurements of wakefulness across the entire SOP , which can be used to characterize SOP dynamics . Moreover , multiple measurements provide statistical power for descriptive and comparative data analyses . These active measurement schemes , however , have all required the use of external stimuli that are potentially arousing and can disrupt sleep [17]–[19] . It has therefore been a question of balancing the trade-off between stimulus salience and the degree to which the SOP is perturbed . Passive behavioral methods for measuring the SOP include actigraphy [17] , [20] , [21] , continuous pressure ( dead man's switch ) systems [22] , or a finger tapping task [19] . Actigraphy is the most prevalent form of passive measurement , and has recently been brought to popular attention through home sleep tracking applications for mobile devices [17] , [21] . Since actigraphy works under the assumption that behavioral quiescence in the absence of a task indicates sleep , it cannot distinguish between wakeful motionlessness during the SOP and actual sleep , and thus is not precise enough to describe sleep onset [1] . Passive paradigms involving the use of a “dead man's switch” or finger tapping task compress all SOP dynamics into a single data point by defining sleep onset as the moment at which behavior ceases , and thus tend to underestimate sleep latency [1] . While both active and passive behavioral metrics show general correlation with features of the SOP dynamics , neither is without issue . Therefore , an important goal is to search for a behavioral task that features multiple highly-salient trials ( as with the active metrics ) , yet minimizes arousing external stimuli ( as with the passive metrics ) . As sleep is a neural process , direct observation of brain activity has been the primary means of tracking the SOP . The most obvious changes to the EEG during the SOP are a progressive decrease in alpha ( 8–12 Hz ) power , as well as a progressive increase in slow ( <1 Hz ) , delta ( . 5–5 Hz ) , and theta ( 5–8 Hz ) power [1] , [23]–[26] . Recent intracranial recording studies suggest that this progression of EEG activity relates to changes in thalamic activity that occur prior to changes in cortical activity , the timing of which has high variability between subjects [27] . Ideally , any descriptor of sleep must account for the fact that it is a complex neural process consisting of multiple local [28]–[30] and spatiotemporally-evolving [23] , [31]–[35] factors . In practice , neural activity is characterized through through polysomnography ( PSG ) —the visual analysis of brain ( EEG ) , muscle ( EMG ) , muscle ( ECG ) , cardiac ( EOG ) , and respiration ( PTAF/Airflow ) data . In current clinical practice , sleep EEG is visually scored using the Rechtschaffen and Kales ( R&K ) system [24] , comprised of Wake , Stage 1–3 NREM sleep , and REM , defined in 30-second epochs . Researchers such as Chiappa [36] and Hori [37] found that the R&K system was too coarse to properly track the SOP dynamics , and consequently developed alternative scoring systems with many more stages , which were scored in much smaller epochs . Unfortunately , neither of these higher resolution frameworks enjoyed wide implementation , perhaps due to labor-intensive scoring rules . Additionally , existing scoring systems do not explicitly account for heterogeneity observed in normal patients as well as variability associated with age , medications , or neurological disorders [38]–[40] . Overall , current practice views the SOP in a binary semantic framework , and analyzes behavioral and physiological data independently . In this paper , we place the SOP within a physiologically and statistically principled model framework , which allows us to explicitly characterize the dynamic interaction of multiple physiological and behavioral experimental observations . Specifically , given the behavioral task and our experimental setup , we simultaneously acquired three modalities of observations related to sleep initiation: behavioral responses , EMG activity , and EEG spectral power . These observation types each contribute information across multiple time scales about different components of a subject's neural state . By combining the information from of all of these different types of observations , we can create a more robust and principled estimate of wakefulness during the process of sleep initiation that takes advantage of both behavioral and physiological data .
In order to track the course of sleep initiation , our goal is to create a continuous-valued metric of wakefulness that is based on simultaneously observed data from multiple modalities , and for which statistical confidence can be computed . To do so , we must create a task that consists of multiple objective behavioral observations related to wakefulness , which can be tracked across the sleep initiation process . Standard behavioral response tasks that have been used previously , involving external audio , visual , or tactile stimuli , are potentially arousing and may perturb sleep initiation [17]–[19] , [41] . We therefore require a paradigm free of arousing external stimuli , yet with repeated trials that can persist throughout the sleep initiation process . To solve this problem , we designed a self-regulated behavioral task centering on breathing . Subjects were given a small , 2oz , gel-filled stress ball to hold in their dominant hand . They were instructed to breathe normally with eyes closed , and to gently squeeze the ball on each inhale and release on each exhale . Thus , each breath acts as a stimulus , and each corresponding squeeze ( or lack thereof ) is the corresponding response . A correct response is defined as squeeze centered on a respiratory inhale , and an incorrect response is either a lack of response or an incorrectly timed squeeze . Subjects were instructed to start the task as soon as the lights were turned out . An additional bipolar adhesive EMG sensor recorded activity of the flexor digitorum profundus ( FDP ) responsible for finger flexion . Subjects also were fit with a glove designed with a force sensitive resistor ( FSR ) embedded in the middle finger , to measure finger flexion during the behavioral task ( Fig . 1a ) . Both the glove and FDP EMG sensors detect even gentle squeezes ( on the order of the force required for a mouse click ) , thereby allowing subjects to perform the task with minimal effort or muscle fatigue . The traces from the glove and FDP EMG were time-aligned with simultaneously recorded PSG respiratory metrics ( PTAF , airflow , and abdominal belt ) ( Fig . 1b , c ) . These recordings were then visually scored in the following manner: The apex of each respiratory inhale was considered a trial . If a squeeze ( visually scored using the EMG/glove activity ) was present during an inhale ( visually scored using the PTAF , airflow , and abdominal belt ) , the trial was scored as correct ( Fig . 1b , green regions ) . If there was no visible response or a misaligned squeeze , the trial was scored as incorrect ( Fig . 1b , red regions ) . Periods including motion artifacts , signal degradation due to temporary sensor disconnection , or any other uncertainties in the signal were left unscored and treated as missing data in subsequent analyses . Scoring was started at the first sequence of trials following lights out that began with at least 3 consecutive correct responses . Scoring was stopped 10 minutes following the last correct response . Some subjects reported difficulty performing the task while they adjusted to wearing the full EEG/EMG/PSG montage . After excluding data from four nights with poor task compliance due to difficulty habituating to the extensive sensor montage , the remaining 16 nights from 9 subjects were processed using our algorithm . A wake probability curve was generated for each night . Along with each behavioral response , we simultaneously observed the EMG activity in the FDP muscle—including the amplitude of each squeeze accompanying a correct response ( Fig . 1d ) . To measure the magnitude of the squeeze , we computed the amplitude envelope of the EMG using a Hilbert transform , then calculated the mean amplitude in a 1 second window centered around the trial time . In tracking EMG data over the course of the SOP , we see that , like a continuous measurement of the muscle activity during a dead man's switch paradigm [1] the EMG squeeze amplitudes decay until the correct responses stop entirely ( Fig . 1d , bottom panel ) . Thus , the EMG squeeze amplitudes provide a continuous-valued metric of both muscle tone and of wakefulness . Paired with the behavioral task , we simultaneously recorded EEG data from each of the subjects . For our analysis , we chose to focus on the most straightforward , continuous correlates of sleep in the EEG: the power in delta , theta , and alpha bands . The power in these bands contributes information about different neurophysiological systems in play during the SOP . With all these sources of information , we can devise a method for integrating them into a single , statistically principled model of wakefulness during the SOP . Our modeling approach centers on the idea that the EMG , EEG , and behavioral observations each provide information related to the activity of different physiological systems involved in different aspects of the SOP . By integrating the information across these systems , we can create a robust framework for tracking the dynamic changes in a subject's wakefulness as they fall asleep . In this section , we provide a non-technical summary of our modeling methodology and its rationale . We describe the mathematical formulation of our approach in detail in the Materials and Methods section , Formulation of the Wake Probability Model of the Sleep Onset Process . We model sleep onset dynamics relative to the observed behavioral , EEG , and EMG data . Our wake probability model states that as the SOP progresses from wake to sleep: •Probability of a correct behavioral response decreases •EMG squeeze amplitude decreases •Alpha power decreases •Theta power increases •Delta power increases A schematic of this model is shown in Figure 2A . In our model , we define the wake probability Pr ( Wake ) as the distribution of the posterior probability ( the probability of the model given the observed data ) that the conditions necessary for the wake state are met: the subject is responding correctly , the EMG amplitude and alpha power are at their highest , and delta power and theta power are at their lowest . Therefore , as these conditions are met , the mode of Pr ( Wake ) approaches 1 . This allows us to use Pr ( Wake ) as a metric representing the degree to which we believe the subject is awake . Moreover , we have formulated Pr ( Wake ) so that it also represents the distribution of the instantaneous probability a correct behavioral response , and thus directly interpretable in terms of standard behavioral paradigms . The wake probability model can be fit to experimental EEG , EMG , and behavioral data to track Pr ( Wake ) over time . We call the time-varying estimate of Pr ( Wake ) the wake probability curve . We describe wake probability in the Materials and Methods section , Interpreting Wake Probability . To implement this approach , we use a Bayesian state-space modeling framework [6] , [42]–[44] ( Fig . 2B ) . State-space modeling allows us to estimate something that is not directly observable ( in this case , the probability of the subject being awake ) from observations that can be directly measured ( in this case , the EEG , EMG , and behavioral data ) . We first model the observations as a function of state processes that represent , in abstract , the level of activity in each of these systems ( see Materials and Methods , State Models ) . These state processes are not directly observable , but their values can be inferred from the data given the structure of the model . We create three state processes: a motor activity process state xm , an alpha process state xα , and a delta-theta process state xΔθ . For each of the state processes , we define a state equation , which describes the way the states evolve over time . The state equations are designed to reflect the notion that states cannot change instantaneously , and that they are related to their past values . The motor activity process xm represents the degree of wakefulness estimated from the amplitude of the EMG during the behavioral task ( Fig . 1d ) . As the subject becomes drowsy , the force of the squeezes will decrease and eventually revert to the underlying muscle tone . The alpha process xα represents the degree of wakefulness estimated from the spectral power in the EEG alpha band . As the subject falls asleep , the alpha power will decrease . If the subject awakens , the alpha will return ( subjects are told to maintain eyes closed ) . In our model , the delta-theta process xΔθ represents the degree of wakefulness estimated from the spectral power in the EEG delta and theta bands . As the subject enters NREM sleep , the delta and theta will increase . If the subject awakens , the power in delta and theta will rapidly decrease . Each of the state processes can change independently , reflecting the asynchronous dynamics of the cortical and subcortical systems generating these EEG rhythms throughout the SOP . We formulate our model of wake probability to be a function of the linear combination of the three states such that xm and xα have a direct relationship to Pr ( Wake ) , while xΔθ will have an inverse relationship to Pr ( Wake ) . We next define the observation equations ( Fig . 2A ) , which describe mathematically the relationship between the experimental observations ( EMG , alpha , delta , theta , and binary responses ) and the underlying state processes ( see Materials and Methods , Observation Models ) . Each observation equation is constructed so that the value of the state process is high when the data indicates high activity , and low when the data indicates low activity . We also define an observation equation relating behavioral response to wakefulness , such that response probability is directly proportional to Pr ( Wake ) . Together , the state and observations define a framework relating our experimental observations to the underlying behavioral and physiological processes , and provide an explicit model for how the aggregate activity of these processes relates to changes in behavior . Using the state and observation equations together with the data , we simultaneously estimate the hidden states and model parameters at each time , using a particle filter , which is a Bayesian sequential importance resampling method ( see Supplementary Materials , Particle Filter ) . The particle filter evaluates all the data observations in context with model equations and computes the maximum-likelihood state and parameter values . The particle filter output is an estimate the full distribution of the posterior probability of the wake probability model , given the observed EEG , EMG , and behavioral data . In summary , our approach takes basic assumptions about the way experimental data evolves during the SOP and explicitly models them in a state-space framework . From this model , we can estimate the wake probability curve , which tracks the dynamics of the SOP by integrating simultaneously observed behavioral and physiological data . Thus , our method provides a robust , statistically-principled , and physiologically-motivated method for characterizing SOP . Since subjects fall asleep at different rates with different dynamics , comparing physiological activity between subjects and populations has been a difficult problem . As a result , previous studies have been limited to anecdotal analyses or static statistical analysis using categorical bins for data . Fortunately , the wake probability now allows us to compare the SOP of different subjects in a principled manner . This is because the value of Pr ( Wake ) provides a common point of wakefulness for the alignment of the physiological data across subjects . To characterize the population dynamics of the EEG during the SOP , we estimated the EEG spectrum of the population as a function of Pr ( Wake ) . Specifically we calculated the median spectrum over all subjects and nights at each value of Pr ( Wake ) . We considered values of Pr ( Wake ) in discrete bins of width 0 . 0025 between 0 and 1 . We then plotted this group-level spectrum as a function of Pr ( Wake ) . The result is a visualization that looks like a spectrogram , but displays median population spectral power as a function of frequency and Pr ( Wake ) , rather than frequency and time . We refer to this plot as the SOP population spectrogram . Since Pr ( Wake ) also represents response probability , this analysis therefore characterizes the average EEG spectrum dynamics during the SOP as the behavioral response probability declines during the transition from wakefulness to sleep . The SOP population spectrogram allows us to summarize an SOP phenotype for a given population of subjects . Furthermore , we can characterize the difference in the SOP phenotype of two populations by comparing their population spectrograms . To do so , we performed a bootstrap procedure [45] , [46] to compute the difference distribution for each frequency-Pr ( Wake ) bin using 10 , 000 iterations per bin . A frequence × Pr ( Wake ) bin was said to be significantly different between populations if zero fell outside the 2 . 5th and 97 . 5th percentiles of the difference distribution . The procedure for constructing an SOP population spectrogram is described in detail in the Materials and Methods section , Computing SOP Population Spectrograms . Since wake probability is a useful abstract quantity not directly observable during the SOP , standard analyses of measurement error are not possible , as there is no ground truth against which Pr ( Wake ) can be compared . Instead , we can perform a likelihood analysis to assess how well a particular model of the SOP of describes the behavioral task data . We used Bayesian Monte Carlo procedures to compute the likelihood of a given model as well as compare the likelihoods of two competing models . These procedures are described in detail in the Supplementary Materials section , Bayesian Goodness-of-Fit . Clinically , the SOP is typically characterized by hypnogram-based definitions of a single moment of sleep onset . By definition , any characterization of a “sleep onset point” cleaves SOP dynamics into a unitary wake state prior to the sleep onset point and a unitary sleep following that point . Thus , while never stated outright , any definition of a sleep onset point imposes an instantaneous transition model on the SOP . Since these models assume an instantaneous wake/sleep transition , it follows that they also assume an instantaneous change in behavioral task performance . We can therefore construct a probability curve analogous to the wake probability curve for any instantaneous transition model by conservatively assuming that the subject should respond correctly with significance ( 95% accuracy ) when deemed awake , and incorrectly with significance ( 5% accuracy ) when deemed asleep . We can then compare these curves to the wake probability curve in order to assess the relative goodness-of-fit . The specifics of the Bayesian construction are also detailed in the Models section . To perform the goodness-of-fit analysis , we computed the likelihood distributions for the wake probability model and four different instantaneous transition models using the behavioral data across all subjects for all nights . We then computed the confidence ( Bayesian credible interval ) with which the wake probability model likelihood differed from that of each instantaneous transition model . Since the wake probability model incorporates information from the behavioral data , we used the posterior distribution from the time step prior to the behavioral observation in all of the goodness-of-fit analyses to insure that use of true behavioral response in the wake probability model formulation did not unfairly affect the results . In our model of the SOP , a subject's probability of wakefulness is based the combined information from both behavioral and physiological observations . Figure 3 shows an example of the model fit to data from one of our experimental subjects . The wake probability curve ( Fig . 3B ) is estimated using information from both behavioral and physiological data ( Fig . 3A , EMG: black dots , alpha , delta , theta power: black curves ) , and therefore integrates features of both modalities . This is most clearly demonstrated by comparing the wake probability curve with the corresponding raw data ( Fig . 3A ) , EEG spectrogram ( Fig . 3C ) , and clinical hypnogram ( Fig . 3D ) . The behavioral data starts with a train of correct responses while the subject is awake with eyes closed . This period is followed by increasing numbers of incorrect responses , which coalesce into a train of incorrect responses . Correspondingly , the EEG data transitions from a pattern with a strong alpha oscillation and minimal delta or theta during high wakefulness , to a pattern with intermittent alpha and rising energy in delta and theta bands . Eventually , the EEG is dominated by delta and theta power , and the alpha oscillation disappears . During the SOP , we observe the alpha power decreasing in a sigmoidal fashion , and the delta and theta power increasing in a sigmoidal fashion . The EMG amplitude decays exponentially at first , followed by a sigmoidal trajectory . These trajectories are in line with our model observation equations , depicted in ( Fig . 2A ) . Consequently , the model estimates ( Fig . 3A , colored curves and regions ) track the raw data ( black ) closely . The structure of the wake probability curve ( Fig . 3B ) appears to successfully integrate features of the behavioral and physiological observations . For roughly the first 13 minutes of the SOP , the wake probability curve is close to 1 , with a narrow confidence interval . This agrees with large number of correct behavioral responses , strong alpha mode in the spectrogram , and hypnogram score of wake . Shortly after 13 minutes into the SOP , the probability curve fluctuates several times before settling into a low Pr ( Wake ) median at around 17 minutes . After 17 minutes , the behavioral responses are exclusively incorrect , the EEG alpha power has dropped out , there is a sharp rise in delta and theta power , and Pr ( Wake ) is low . Moreover , the rise in Pr ( Wake ) between 21 and 22 minutes aligns directly with the hypnogram , which goes from Stage 2 to Stage 1 and back during the same interval . By examining the transition period in this same subject in greater detail , we can gain further insights into how the behavioral and EEG data are combined to estimate the wake probability curve . Figure 4 shows a close up of the data from Figure 3 on a time scale of 6 minutes . This period begins with a string of 22 consecutive correct behavioral responses ( Fig . 4B ) . Since correct responses indicate wakefulness , this information pushes Pr ( Wake ) towards 1 . During the same time , however , the alpha power ( Fig . 4A ) is sporadically present , supplying support that the subject is more ambiguously awake ( i . e . , less than 1 ) . Given our model , low alpha and high delta-theta power pull Pr ( Wake ) towards 0 . While the behavioral responses are correct , loss of alpha power indicates reduced wakefulness , resulting in a lowering of Pr ( Wake ) and an increase in the uncertainty of the estimate ( as indicated by wider confidence bounds ) . As the subject transitions through the SOP , the number of incorrect responses increases , the alpha diminishes progressively , and delta and theta appear and begin to coalesce into prominent oscillations . This period is marked by an alternation between alpha and delta/theta activity [12] , [26] , [32] , and continues until the alpha is gone , the delta/theta is high , and all the responses are incorrect . Consequently , we see peaks in Pr ( Wake ) where there is high alpha power , low delta/theta power , and correct responses , and troughs in Pr ( Wake ) where the opposite is true . The magnitude of these peaks and troughs are based on the degree to which the aggregate data indicates that the subject is awake . The confidence bounds are related to the degree to which all of the data is in agreement . In comparison to the clinical hypnogram ( scored in 30 s epochs ) ( Fig . 4D ) , the wake probability curve characterizes this transitional period of the SOP with a much higher temporal resolution . Additionally , the wake probability curve describes a continuum of wakefulness , whereas the hypnogram discretizes this transitional period into three categories: Wake , Stage N1 , and Stage N2 states . The transition from wake to sleep can be fragmented—most notably in patients suffering from difficulties with sleep initiation , but also in healthy people . Figure 5 shows data from the second experimental night from the same subject shown in Figures 3 and 4 . Rather than the smooth transition seen the first experimental night , we observe that this night the subject had brief arousal period in the middle of the SOP . The wake probability curve captures both the slow transition from wake to sleep , as well as the rapid changes in wakefulness during the arousal period . As in the subject's first night , the SOP begins with trains of correct responses ( Fig . 5B ) , a strong alpha mode , and low delta and theta ( Fig . 5A ) , which results a high Pr ( Wake ) ( Fig . 5C ) . The alpha mode then becomes sporadic , which coincides with an increase in incorrect responses . Next , there is a train of consecutive incorrect responses , the alpha mode disappears , and there is a dramatic increase in the theta power and rising delta power . Consequently , the median of Pr ( Wake ) drops towards 0 . Suddenly the correct responses begin again , the alpha mode returns , and the delta and theta drop off . Given this information , Pr ( Wake ) then ascends rapidly towards 1 . After approximately 1 minute , the responses become exclusively incorrect , the alpha power disappears . The delta and theta power rapidly return to their pre-arousal levels , continuing to increase for the rest of the SOP . The wake probability curve tracks the drop in Pr ( Wake ) and the dynamics for the rest of the SOP . Again , the wake probability curve structure agrees strongly with the structure of the hypnogram ( Fig . 5D ) , but provides greater temporal resolution and finer granularity in the state estimate . In the preceding illustrative examples , there is strong agreement between the behavioral and physiological data . In practice , however , there is neurophysiological heterogeneity observed—even within healthy subjects—such that there is often a great disparity between behavioral and physiological metrics of sleep onset . In this section , we show how the wake probability curve characterizes such situations . Figure 6 shows the results from another healthy subject with a dramatically different SOP phenotype . As in the previous case , the experiment begins with a strong alpha oscillation , which eventually disappears ( Fig . 6A ) . In this case , however , the correct responses persist long after the alpha has diminished ( Fig . 6A ) . Moreover , there is a roughly 5-minute interval between the time the alpha mode declines and the time the theta and delta power increase . This SOP alpha dropout phenotype with a long interval between alpha power decline and delta/theta power rise results in disagreement between standard sleep scoring and a behavioral analysis . In this period between the loss of alpha and loss of response , the hypnogram ( Fig . 6D ) describes the subject as being predominantly in Stage N1 , with a brief period of Stage N2 , and a short period of Wake when there is a short increase in alpha . Thus a standard interpretation of the hypnogram would place sleep onset at the first epoch of Stage N1 , approximately 3 minutes into the SOP . This is in contrast to the behavioral data , which continues to indicate wakefulness for another 5 minutes past the first epoch of Stage N1 . The wake probability curve ( Fig . 6C ) , however , integrates all the data such that the estimated median of Pr ( Wake ) is still high during this period , declines slightly , and has a large uncertainty as a result of the contradicting observations . By combining both the behavioral and physiological data into the estimate of Pr ( Wake ) , we can bridge the disparity seen between metrics that exclusively rely on ether behavior or EEG alone . The result is a model that can represent deviation from the population norm as increased uncertainty . In this analysis , 2 of the 9 subjects ( Supporting Information Figures S1 and S2 ) clearly exhibited this alpha dropout phenotype , in which alpha power declined up to several minutes prior to the termination of correct responses and the increase of delta and theta power . For both subjects , this phenotype was present on both experimental nights . Three of the four nights had periods of scored Stage N1 during which there were correct behavioral responses . In none of the cases did we observe correct responses in the presence of strong delta and theta . This suggests that loss of alpha power , while necessary , is not sufficient for the loss of behavioral responses . In clinical practice , the most common definitions for the moment of sleep onset are: the first epoch of Stage N1 , the first epoch of Stage N2 , the first of any 3 consecutive NREM ( N1 or deeper ) epochs , and the first of any 10 consecutive epochs of NREM . Though not stated explicitly , any characterization of a point of sleep onset actually imposes a model on the SOP with an instantaneous sleep/wake transition , which does not agree with the continuous , dynamic transitions observed in the data . We performed a likelihood analysis comparing how well of the wake probability model and instantaneous transition models fit the behavioral data . Likelihood is a relative estimate of goodness-of-fit , and given two competing models , the one with the better fit will have a higher likelihood . The comparative likelihood analysis showed that the wake probability model significantly outperformed each of the instantaneous transition models with at least 99 . 99% confidence . These results are summarized in Figure 7 and in Table 1 . Overall , the wake probability model fit the data the best with the largest median loglikelihood ( −1589 ) , followed by , in order of goodness-of-fit , the first epoch of N1 model ( −2781 ) , the first of 3 NREM model ( −2852 ) , the first of 10 NREM model ( −3191 ) , and by the first epoch of N2 model ( −5828 ) . To illustrate the way in which the wake probability model improves upon the instantaneous transition models , we performed the goodness-of-fit analysis on a single night of data . Figure 6E and F show , respectively , the instantaneous transition model response probabilities generated from the hypnogram , and the resultant goodness-of-fit analysis for that experimental session . This clearly shows the way in which the instantaneous transition models implicitly discretize complex dynamics of the SOP into unitary “wake” and “sleep” states , thus losing the ability to capture any nuances in state throughout . Furthermore , since current EEG-based definitions of sleep onset do not include behavioral information , the assumption that Stage N1 is equitable with “sleep” can be misleading [1] , particularly for those subjects ( like this one ) in which behavior persists past the alpha dropout . Consequently , the wake probability model ( C ) fit the behavioral response data the best ( F ) with median loglikelihood of −41—significantly outperforming the instantaneous transition models with at least 99 . 99% confidence . Within the class of the instantaneous transition models ( E ) , the first of any 10 consecutive NREM epochs model performed the best in this particular case , with a median loglikelihood of −68 . In this case , the first epoch of N1 model and first of 3 consecutive NREM epochs model both provided the same response probability estimates , and each had a median loglikelihood of −113 . Finally , the first epoch of N2 model performed the worst , with a median loglikelihood of −199 . Overall , these results suggest that the wake probability model is a more mathematically and physiologically appropriate metric with which to track the SOP than are the current hypnogram-based metrics . One of the key strengths of the wake probability model is that it can characterize the EEG activity for groups of subjects across the entire SOP , rather than at a single point of alignment . Using the wake probability curves from multiple subjects , we can compute an SOP population spectrogram , which plots the median cross-subject EEG power spectrum as a function of the behavioral response probability ( see Materials and Methods: Computing SOP Population Spectrograms ) . By using these techniques , we can group the SOP data of multiple subjects on a continuum , from which we make rigorous statistical statements about the differences between populations . As an example , we quantify , for the first time , the differences in the EEG between subjects with “normal” and “alpha dropout” SOP phenotypes on a continuum of wake/response probability . We computed the SOP population spectrogram using the data from all the subjects and nights ( Fig . 8A ) . These results clearly show the dynamic transition from a strong alpha mode to increasing delta/theta power as the probability of response progresses from 1 to 0 as the subject falls asleep . As the SOP progresses , the alpha power reduces amplitude , dropping out near a response probability of 0 . 55 . The delta/theta mode emerges at around a response probability of 0 . 4 , increasing its bandwidth and amplitude as the response probability approaches 0 . We can also use the SOP population spectrogram to characterize difference the average EEG activity from different populations at moments at which their behavior is identical . As an illustrative example , we computed an SOP population spectrogram using the data from the two subjects ( 2 nights/subject , 4 nights total ) that showed a clear alpha dropout phenotype ( Fig . 8B ) . The analysis reveals a different spectral structure , with the alpha mode dropping out near a response probability of 0 . 85 , and delta/theta power emerging around a response probability of 0 . 2 . We then used a bootstrap procedure to compare the SOP population spectrograms of the subsets of subjects with normal and alpha dropout phenotypes ( Fig . 8C ) . This analysis revealed a region of 95% significant difference ( red areas ) covering the bounds of the alpha mode of the standard phenotype . These results suggest that are indeed subjects that possess significantly reduced alpha power yet can maintain behavior response levels identical to other subjects with a strong alpha oscillation . Analyses such as these therefore provide a principled mathematical framework for characterizing individual SOP phenotypes , as well as for quantifying SOP heterogeneity .
There is currently is no behavioral monitoring standard in sleep medicine . In experimental sleep studies , active behavioral monitoring requires potentially arousing auditory stimuli . Our new breathing task presents a new paradigm for behavioral monitoring free of external stimuli . Moreover , with our new paradigm , all that is required is a respiration monitor and EMG leads on the subject's forearm , both of which are already part of the standard clinical setup . There is also no need to tackle the difficult problem of determining the correct stimulus volume that best balances salience with the potential for subject arousal . Further experimentation is needed to definitively ascertain the comparative benefits of the breathing task over standard active behavioral measures . However , there is significant evidence in the literature suggesting that this paradigm has major advantages . While the breathing task is like all other active tasks in the sense that it requires repeated behavioral responses to stimuli , it is innovative in that there are no external sensory stimuli , which can cause arousal from sleep [18] , [19] , [41] , [47] . Rather , this task could be said to rely on “internal stimuli” generated from the act of breathing . The breathing task therefore acts as a bridge between active and passive behavioral measures of sleep onset—providing high temporal resolution while minimizing the effects from the stimuli . It is then a question as to whether the act of concentrating on breathing is arousing in and of itself . On the contrary , focused repeated breathing has been shown to reduce anxiety and tension [48] , to decrease heart rate and blood pressure [49] , to increase parasympathetic and decrease sympathetic activation [50] , to decrease oxygen consumption [51] , and has been correlated with reduction in EEG alpha power [52] . Since many of these effects are physiological hallmarks of the SOP , the act of attending to the breathing task would be unlikely to arouse subjects by itself , and could even potentially facilitate the wake/sleep transition . Additionally , interventional cued breathing studies have been shown to reduce the duration apnea events [7] . In our model , we compute wake probability , an estimate of the probability that the subject is awake given evidence from simultaneously observed EEG , EMG , and behavioral data . This approach improves on contemporary staging of data , where a choice needs to be made between wake and sleep . Here we produce a continuous-valued metric that tracks the full spectrum of states during the SOP . In so doing , we more accurately characterize the SOP as a dynamic system , and can therefore make more precise observations and predictions about the underlying physiology . There are several key factors in this analysis that enable this dynamic , multimodal characterization . First , we designed the wake probability model with the goal of tracking the dynamics of a gradually changing system . In his 2001 review , Ogilvie comprehensively and persuasively enumerated the preponderance of scientific evidence supporting the notion that the SOP is a gradual dynamic process , and decried the notion of characterizing a single moment of sleep onset . In the decade following , newer studies have only added more support to this argument through the further analysis of cortical and subcortical activity [27] , [30] . Moreover , our nightly experiences with falling asleep tell us that the transition from wake to sleep is not an instantaneous process . In spite of all this experimental and empirical evidence , SOP dynamics have not been embodied in previous quantitative analyses . By modeling wake probability as a continuous-valued metric , we can now characterize the SOP as a dynamic process , bridging the gap between the evidence and the analysis techniques . Second , our model incorporates data from both physiological and behavioral observations . Often , there can disagreement between EEG and behavioral metrics of in the estimation of sleep onset , since changes in the EEG and behavior are not necessarily time-locked to each other . Ogilvie observed that behavioral responses could persist well into Stage N1—far beyond the point at which many standard criteria for sleep onset would consider sleep—and went so far as to suggest that N1 not be even considered to be true sleep [1] . Additionally , visual scoring paradigms have difficulty handling the heterogeneity observed in the normal EEG population , and consequently will deem a subject to be asleep due to diminished or missing alpha oscillations . It is therefore is of vital importance to use both behavioral and physiological data in any metric that characterizes the SOP . Third , we model the SOP as a combination of multiple independent components , which can evolve on different time scales . In formulating the model , we designed the state equations so that the alpha , delta/theta , and motor states could evolve independently based on the data . This flexible setup reflects the idea that interacting systems can activate or deactivate on different timescales throughout the SOP , an idea substantiated through intracranial studies of corticothalamic activity [27] . In our model , is the superposition of these states that governs the behavioral response data and vice versa . In our model , each observation type reflects the activity of a systems-level neural component of the SOP , and the aggregate effect of all the systems governs arousal and consequently behavior . Finally , our framework is statistically principled . Since the model is Bayesian and computes the full posterior distribution of Pr ( Wake ) , we can perform many other rigorous statistical comparisons between any set of points in a night for a single subject , as well as comparisons between subjects [53] , [54] . Moreover , if a single point of alignment is indeed required , we can now take a statistically principled approach by defining it using Pr ( Wake ) . For example , one could identify the first time point at which Pr ( Wake ) was significantly less than 0 . 95 . In this method , we frame the characterization of the SOP in terms of the probability of wakefulness , rather than the probability of sleep . We do this because the SOP is a complex multifocal process [28] , [29] , [32] , which is constantly evolving . Consequently , trying to estimate the probability of sleep is the equivalent of shooting at a moving target , since “sleep onset” could refer to any point on a vast continuum of dynamic neural activity . To simplify the problem , we therefore chose to create a simple model of the waking state and track its disappearance rather than tackle a complex model of sleep and track its initiation . It should be noted that Pr ( Wake ) does not necessarily equal 1−Pr ( Sleep ) , as local sleep-related processes can co-exist with wake-related processes in the brain during the SOP [32] . Additionally , this framework lets us define wake probability as equivalent to probability of a correct response , so that Pr ( Wake ) can be discussed in terms of behavioral responsiveness , given the additional data from the EEG and EMG . It is clear that alpha , delta , and theta are not the only oscillations in play during the SOP , nor are they spatially static . Fortunately , our framework provides a straightforward means of implementing more sophisticated models of wakefulness . Future models can incorporate additional physiological observations such as slow ( <1 Hz ) , beta ( 15–30 Hz ) , sigma ( 12–15 Hz ) and gamma ( >30 Hz ) power , EEG spatial and coherence dynamics , and other biomarkers of sleep such as body temperature , heart rate , blood pressure , and more . The model may also be augmented to include other behavioral measures . This model flexibility provides several major benefits . By adding more observation modalities , we can develop a model that fully captures our current understanding of the multiple systems affected during the SOP . Furthermore , continued model adjustments will allow SOP analysis to keep pace with new findings . Finally , since the behavioral component of this framework can be adjusted to characterize any other task or removed to account for no behavioral data at all , we can therefore easily apply this analysis to data previously collected in other experiments . Future work will focus on the many practical applications of our methods . Using our statistical framework , we can build models to rigorously characterize and compare the SOP phenotypes of different clinical populations , as well as to continue to characterize the natural heterogeneity of healthy subjects . By relating model component temporal dynamics to known linkages between observations and their underlying neural systems , this sort of analysis may help to shed further light on the pathophysiology of sleep . Furthermore , we can use our likelihood analysis to assess the relative goodness-of-fit of any set of proposed models , determining which model best fits the data . In doing so , we can provide a means of assigning any newly observed data to the phenotype or pathology associated with the model with the maximum likelihood , thus creating an efficient and principled means of automated diagnosis or categorization . Additionally , by characterizing the time course of the wake probability curve itself , we can quantify differences in the rapidity of sleep onset in a principled manner . This analysis could act as a diagnostic tool for disorders of sleep onset , by comparing a subject's wake probability curve to those from population possessing a known pathology . If we adapt techniques for analyzing group behavior [57] , the time course for sets of wake probability curves could also be compared , providing a way to analyze the influence of factors such as pharmacological agents , pathology , and the first night effect on the SOP . Furthermore , wake probability could be used to dynamically track drowsiness in situations in which alertness is vital . Rather than attempt to detect the onset of sleep , it may be more important to detect the point at which wakefulness and the behavior associated with it decline .
Human studies were approved by the Human Research Committee of Massachusetts General Hospital , Boston , MA . Ten healthy right-handed subjects ( 5 men and 5 women ) with ages ranging 19–32 years ( mean: 25 . 8 , std: 5 . 09 ) and BMI <30 slept for two consecutive nights in our sleep laboratory . Subjects were screened to ensure a regular sleep schedule and no history of sleep disorder , psychiatric problem , or neurological disease , as well as to ensure no history of tobacco , or prescription/recreational drug use . We performed one night of home monitoring to exclude obstructive sleep apnea ( OSA ) screening ( using a threshold of AHI <5 , and RDI <15 ) ( WatchPAT , Itamar Medical ) . A trained technician analyzed the experimental PSG data following the first experimental night , and one subject was excluded after failing to meet the OSA criteria on the first night ) . Urine tests for drug use ( Xalex Multi Drug Kit for 10 Drugs ) occurred at screening and prior to each experimental night . Female subjects were also screened for pregnancy . Subjects were fit with a high-density ( 64-channel ) EEG cap , as well as standard clinical PSG sensors including PTAF , airflow , abdominal belt , and eye , chin , and limb electrodes . EMG data were bandpass filtered between 10 and 70 Hz with the addition of a notch filter at 60 Hz . Airflow and abdominal belt data were bandpass filtered between . 1 and 12 Hz . EEG and DC channel data were unfiltered . Multitaper spectrograms of the EEG data from 8 occipital channels were computed with 6 s temporal windows and 0 . 25 s overlap , a time-bandwidth product of 3 , and 5 tapers . Delta , theta , and alpha power were defined as the total multitaper spectral power between 0 . 5–5 Hz , 5–8 Hz , and 8–12 Hz , respectively , of the median of the 8 occipital EEG channel spectrograms . It should be noted that the frequency band definitions for alpha , delta , and theta bands are not universally standardized , and thus vary between subfields within in the electroencephalography literature . Visual staging of sleep data was performed prior to the statistical analysis by an experienced clinical sleep technician using contemporary AASM guidelines [5] , [58] . We first define the state equations , which model the temporal evolution of the state processes over time . We define the random walk process at time t as ( 1 ) where and is a constant . To estimate the value of the model coefficients that are not time-varying , we used a random walk ( 11 ) where and is small . This leaves room for some exploration of the parameter space without allowing for any large changes in the parameter value . The coefficients and priors used in this model can be found in the Supplementary Materials , Implementation Details . Given our state and observation models , we can construct θt , a vector of the parameter values at t such that ( 12 ) For each θt , we can estimate the posterior density —the probability of all the model parameters θt given the data . The posterior density is proportional tolog ( L ( θt ) ) , the joint log-likelihood of all the observations given the parameters . We sum the log-likelihoods of all the observation processes to construct log ( L ( θt ) ) . Given the binomial loglikelihood for the binary responses ( 13 ) and using Gaussian likelihoods for the continuous-valued observations , the full log-likelihood becomeswhere Ib , t , Im , t , and are indicator functions for each type of observation at time t . These indicator functions take on the value of 1 if the corresponding observation is present and 0 if it is missing . This sets up a flexible likelihood function that is able to deal with missing data for any of the observations . Furthermore , any time there is missing data for any reason ( such as a disconnected EEG or a faulty connection ) , the log-likelihood can be constructed from whatever remaining data is available . Given time-frequency observations from EEG data during the SOP , from S subjects , over discrete times , and fixed-width frequency bins centered at frequencies , we define a matrix Ys as ( 15 ) such that is the magnitude of the power spectrum for subject s at time t within the frequency bin f . We also divide wake probability space into discrete bins , which divide the interval [0 , 1] into P non-overlapping bins . We define the SOP population spectrogram Φ ( p , f ) , as: ( 16 ) where is the subset of all the for all subjects in which falls within bin p . In all cases , the median may be substituted for the expectation . | How can we tell when someone has fallen asleep ? Understanding the way we fall asleep is an important problem in sleep medicine , since sleep disorders can disrupt the process of falling asleep . In the case of insomnia , subjects may fall asleep too slowly , whereas during sleep deprivation or narcolepsy , subjects fall asleep too quickly . Current methods for tracking the wake/sleep transition are time-consuming , subjective , and simplify the sleep onset process in a way that severely limits the accuracy , power , and scope of any resulting clinical metrics . In this paper , we describe a new physiologically principled method that dynamically combines information from brainwaves , muscle activity , and a novel minimally-disruptive behavioral task , to automatically create a continuous dynamic characterization of a person's state of wakefulness . We apply this method to a cohort of healthy subjects , successfully tracking the changes in wakefulness as the subjects fall asleep . This analysis reveals and statistically quantifies a subset of subjects who still respond to behavioral stimuli even though their brain would appear to be asleep by clinical measures . By developing an automated tool to precisely track the wake/sleep transition , we can better characterize and diagnose sleep disorders , and more precisely measure the effect of sleep medications . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"physiological",
"processes",
"medicine",
"and",
"health",
"sciences",
"mathematics",
"statistics",
"(mathematics)",
"computational",
"neuroscience",
"physiology",
"sleep",
"mathematical",
"and",
"statistical",
"techniques",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"neuroscience",
"statistical",
"methods",
"research",
"and",
"analysis",
"methods",
"bayesian",
"statistics"
]
| 2014 | Tracking the Sleep Onset Process: An Empirical Model of Behavioral and Physiological Dynamics |
The effects of various dengue control measures have been investigated in previous studies . The aim of this review was to investigate the relative effectiveness ( RE ) of different educational messages embedded in a community-based approach on the incidence of Aedes aegypti larvae using entomological measures as outcomes . A systematic electronic search using Medline , Embase , Web of Science and the Cochrane Library was carried out to March 2010 . Previous systematic reviews were also assessed . Data concerning interventions , outcomes , effect size and study design were extracted . Basic meta-analyses were done for pooled effect size , heterogeneity and publication bias using Comprehensive Meta-analysis . Further analysis of heterogeneitity was done by multi-level modelling using MLwiN . 21 publications with 22 separate studies were included in this review . Meta-analysis of these 22 pooled studies showed an RE of 0 . 25 ( 95% CI 0 . 17–0 . 37 ) , but with substantial heterogeneity ( Cochran's Q = 1254 , df = 21 , p = <0 . 001 , ) . Further analysis of this heterogeneity showed that over 60% of between study variance could be explained by just two variables; whether or not studies used historic or contemporary controls and time from intervention to assessment . When analyses were restricted to those studies using contemporary control , there was a polynomial relationship between effectiveness and time to assessment . Whether or not chemicals or other control measures were used did not appear have any effect on intervention effectiveness . The results suggest that such measures do appear to be effective at reducing entomological indices . However , those studies that use historical controls almost certainly overestimate the value of interventions . There is evidence that interventions are most effective some 18 to 24 months after the intervention but then subsequently decline .
Dengue fever ( DF ) is an acute viral disease affecting all age groups . It occurs mainly in tropical and subtropical areas , the predominant vectors being the mosquitoes Aedes aegypti and albopictus , which become infected with any of the four dengue viruses and transmit the disease via a bite to humans [1] . Some 2 . 5 billion people ( two-fifths of the world's population ) are now at risk from dengue , and the WHO currently estimates that there may be 50 million dengue infections worldwide every year [2] . Depending on the year , tens to hundreds of thousands of cases of the severe and potentially fatal form of the disease , dengue haemorrhagic fever , and dengue shock syndrome ( DHF/DSS ) occur [3] . The incidence of DF has increased dramatically in recent decades . Its proliferation is influenced by many mechanisms – these include population growth with unplanned urbanisation ( and consequent overburdening of water and sanitation systems ) , increases in domestic and international travel , transportation of commodities such as tyres , lack of political will to intervene , and limited financial and human resources to implement effective control measures [4] . The disease has become endemic to more than one hundred countries in Africa , the Americas , the Eastern Mediterranean , South-East Asia and the Western Pacific . Of these , South-East Asia and the Western Pacific are the most seriously affected . There is currently no vaccination for DF , and no medications that can treat DHF or DSS directly , so at present the only way of controlling or preventing the spread of the virus is to combat vector mosquitoes directly . For many years , spraying with insecticides , such as malathion , has been the main method of control , though this has often had limited success [5] . Other interventions aimed at controlling the mosquito population , have been tested with varying success . For example , the Puerto Rican government , in response to the threat of a DHF epidemic , developed an integrated approach consisting of community-based dengue control programs to complement traditional chemical-based approaches [6] . They encouraged the public to reduce or eliminate containers in and around homes , gardens and villages . These containers , which include discarded plastic packaging , metal cans , and rubber car tyres , are capable of holding water which would then harbour larvae , and allow mosquitoes to proliferate [7] . It has become clear , from the number of projects that have been initiated in recent years that community-based programs are now regarded by both national and international health agencies as the primary long-term solution for prevention and control of DHF/DSS in Asia and the Americas [3] . A recent systematic review carried out by Erlanger and colleagues investigated the effect of different types of dengue vector control interventions , including biological , chemical , environmental and integrated vector management , on well established entomological parameters [4] . Their aim was to compare the effects of these interventions , in order to find the most efficacious . They identified 56 publications with extractable data that had compared the impact of 61 different dengue control interventions with control communities or with the same community prior to the intervention . The authors concluded that dengue interventions are effective in reducing vector populations , particularly when interventions use a community-based integrated approach . An earlier systematic review by Heintze et al . specifically looked at community-based dengue control programmes , and concluded that the evidence that such programmes were effective , either alone or in combination with other programmes was weak [7] . Yet another recent systematic review also concluded that there was little evidence to support the efficacy of mosquito abatement programs due to poor study design and lack of congruent entomological indices [8] An important criticism that can be levelled at these systematic reviews is that there was substantial heterogeneity in study design and in the size of any effect that made it difficult to draw definitive conclusions . In particular some studies used historical control periods whilst others used other contemporary communities as controls . Many of the studies included multiple interventions in combination whilst others studies were of a single intervention . Furthermore the control communities may or may not have had one of more interventions themselves . We argue that these issues make it particularly difficult to disentangle the value of educational messages embedded in a community-based approach , or identify the most successful approach . Although Erlanger and colleagues did undertake subgroup analysis around types of intervention , neither of these studies adequately investigated sources of heterogeneity in effect size ( the magnitude of any association between the outcome and predictor ) making the drawing of any definitive conclusions problematic . One of the recent trends in meta-analysis has been the increasing use of methods that aim to investigate causes of heterogeneity in effect size between published studies rather than rely on pooled effects sizes that can often be difficult to interpret [9] , [10] . By this way it is hoped that additional insights can be gleaned into how study design and context , such as use of control interventions in control groups , may affect the outcome . Such insights could give some understanding of in what situations these interventions may , or may not , have benefit . This paper reports a deeper analysis of papers that have attempted to determine the impact of educational messages embedded in a community-based approach , which we define as community based intervention that had any element where members of the public were given information or exhortations intended to change their behaviour , on entomological indicators of risk of dengue disease .
Included studies were required to: firstly refer to control of dengue fever , and secondly have studies investigating an educational intervention alongside a ‘control’ approach or standard management program . Studies also had to look at quantitative outcomes , whether these were the BI , CI , HI or C+/H . Next , these studies had to be community-based , whereby members of the community partook in the interventions or played a major role . Conversely , studies based in laboratory or semi-field settings were excluded , as were purely observational cross-sectional and qualitative studies . Studies were not limited by language of publication . The primary measure of effect size was relative effectiveness ( RE ) with 95% confidence intervals . RE is the ratio between the entomological index in the intervention group and in the control group . Consequently the more effective the intervention the lower the RE . An RE of 1 . 0 would indicate no effect . Where confidence intervals were not given , these were back-calculated from the P value . Where only the entomological index/indices were presented for each group RE and its 95% confidence intervals were estimated using Monte Carlo modelling with @Risk™ . The distributions of the indices for the intervention and control groups were taken from the papers . Then values were repeatedly sampled from each distribution and the value sampled from the intervention distribution divided by that sampled from the control sample to give the RE . From the repeat samplings the distribution of the RE was then determined to give mean and 95% confidence intervals . The review carried out by Erlanger et al . investigated a range of interventions , including entomological and community measures taken in a variety of settings [4] . The objective of this review was to systematically analyse only the publications which included an educational element to their interventions ( even if other non-educational interventions were also included ) . However , we used a rather broad definition of educational intervention to include any community based intervention that had any element where members of the public were given information or exhortations intended to change their behaviour . This was followed by a rigorous up-to-date search strategy , detailed below , which was carried out in order to retrieve references which had been produced since publishing date of the existing review ( September 2008 ) . A structured electronic search of Medline , EMBASE , Web of Science and the Cochrane Database of Systematic Reviews was carried out up to March 2010 . This was performed in the format: [dengue or dengue haemorrhagic fever or dengue virus or Aedes aegypti] AND [arthropod vectors] AND [community based] AND [intervention] . Reference lists were checked for additional publications to the ones found in the initial search , which fulfilled the inclusion criteria . From the initial search results , all titles and abstracts were assessed independently by two reviewers , with disagreements being resolved by discussion . From these , a list of papers to include was made , and full text articles obtained . Once the publications had been assessed as meeting the prescribed quality and inclusion criteria , and having considered the references used by Erlanger et al . , data was extracted systematically , using a standardised form . Data was extracted from the existing systematic review , but also updated with the most recent studies found in the search . Where follow-up occurred over several time points , the longest follow up time point results were included , as this provides the most realistic indicator of long-term effectiveness of the intervention . Data was extracted on the outcome measure , study design , time of follow-up after intervention , what other interventions were used and the nature of the educational component . Where confidence intervals were not presented in the original paper , these were derived by a process of back-calculation from the presented P value . Initial analyses were done with Comprehensive Meta-analysis ( CMA ) Version 2 . 2 . 050 [12] . All four main entomological indices were included in the analyses . If more than one entomological index were reported in the same study , then a single outcome measure was calculated as the geometric mean of the different entomological index by CMA using the within program option to combine effect sizes from different types . CMA was used to calculate heterogeneity , determine potential effects of publication bias and pooled estimates of effect size . In order to determine whether combining REs using different entomological indices was valid , Pearson correlation coefficients were calculated as was paired t tests between them . Subsequent analyses of the impact of moderator variables were done using MLwiN [13] . A basic three level model was constructed to account for studies with multiple comparisons [14] . Each of the putative modifier variables were put singly into the model and those with p<0 . 2 included in a multiple modifier model . In the multiple model , any modifier with p> = 0 . 2 was then removed and the model rerun until all modifiers in the model had p<0 . 2 . The proportion of the between study variance explained by the final model was derived from τ2 ( between-studies variance ) in the model with no modifiers and in the final model .
Searches of Medline , EMBASE , Web of Science and the Cochrane Database of Systematic Reviews identified 491 original papers for assessment . Figure 1 shows the flow diagram detailing the search process and inclusion of studies in this review . Of these 491 articles , 456 articles were excluded based on abstract alone because of inappropriateness of subject or study design . A total of 35 papers were obtained in full text . Of these , 14 full text papers were excluded , deemed to be unsuitable with regard to participants , the intervention used , outcomes of the study , or study design . This left 21 papers of which 11 were based in South America , 9 were based in South East Asia , and the remainder were based in Fiji and French Polynesia . The earliest study was published in 1967 [15] and the latest in 2009 [5] . One paper [11] had two study arms that included interventions of interest and each study arm is referred to as separate study were included , giving 22 studies in total . Studies varied with regard to types of educational component , study design and control groups . The included studies and summary of their characteristics are listed in Table S1 . The educational components included the use of print or broadcast media , public lectures , in-home training by public health staff , home visits and targeting school children . The exact mix of interventions varied between studies . Three different approaches were used in the study designs: 6 studies used an historical control period , measuring outcomes in the same village at baseline and at a later time point ( ‘historical’ control group ) , 11 studies included a control arm with no additional treatment as well as an intervention arm ( ‘no treatment’ control group ) , and 5 studies included a control arm exposed to some anti-mosquito activity , along with an intervention arm ( ‘some intervention’ control group ) . The studies also varied in terms of whether or not the intervention communities received other interventions . A total of 9 studies included some form of chemical intervention , as well as the educational component . This varied from the use of malathion spraying both in- and outdoors , to larviciding with the use of abate . Another 8 studies used various additional “other” ( i . e . not chemical ) measures in the intervention group – these ranged from covering and disposal of containers capable of holding water , to community clean-up campaigns , to the use of other species such as Mesocyclops in order to predate the Aedes spp . The Pearson correlation coefficients for the REs from the three main entomological indices showed high correlation between them BI-CI 0 . 68 , BI-HI 0 . 66 , CI-HI 0 . 97 . Furthermore , there was no significant difference in the mean RE given by the different entomological indices using a pared t test . We conclude that combining the different entomological indices was valid . The result of the meta-analysis performed on all 22 studies is shown in Figure 2 . Using the random effects model , the pooled risk ratio was 0 . 25 ( 95% CI 0 . 17–0 . 37 ) . However , there was substantial heterogeneity in the effect size ( Cochran's Q = 1254 , degrees of freedom ( df ) = 21 , p = <0 . 001 ) . There was no evidence of publication bias ( Figure 3 ) . In order to investigate the sources of heterogeneity further a series of multi level meta-regression analyses were run with potential modifier variables . The results of the initial analyses are shown in table 1 . The most significant single modifier variable was whether or not the study used a historical ( comparing the same community before and after the intervention ) or a contemporary control ( comparing the intervention community with another control community ) . Those studies that used contemporary controls had a much reduced effect size compared to historical controls ( Regression coefficient ( B ) = 2 . 08 , Standard Error ( SE ) = 0 . 65 ) . Two other predictor variables , combining a chemical with the community intervention and time at follow-up almost achieved significance . Whether or not other interventions were also used ( but not including chemicals ) did not achieve statistical significance . Two measures of time at follow-up were tested , the untransformed and log transformed months . The results were very similar between these two time measures and the untransformed used in subsequent analyses as this was marginally more significant and also easier to interpret . In addition the relationship between chemical spraying and RE was further tested as some studies included chemical spraying in both intervention and controls and others in intervention only . Perhaps not surprisingly , chemical spraying where this was applied in both intervention and control arms had almost no effect on RE whilst chemical spraying in the intervention arm but not control arm was associated with a significant improvement in RE ( −2 . 08 , SE 0 . 78 ) . All variables with p<0 . 2 in the single modifier variable analyses were included ( historic or contemporary controls , time at follow-up and chemical spraying in intervention but not control group ) in a final model as shown in table 2 . It can be seen that two modifier variables remain historical v contemporary control and study duration . In particular those studies using contemporary controls gave much smaller effect sizes than those using historical controls ( B = 2 . 21 , SE = 0 . 66 ) and effect sizes improved with longer delays till the follow-up assessments ( B = −0 . 083/month , SE = 0 . 03 ) . Using chemicals in the intervention group but not in controls was not significant . These three variables were able to explain 64% of the between study variance in the original dataset , though the remaining between study variance was still significant ( τ2 = 1 . 07 , SE = 0 . 39 , z = 2 . 77 , p = 0 . 006 ) . Excluding chemical spraying from the model was still able to explain 61% of the between study variance . The relationship between RE , choice of control and time to assessment is illustrated in figure 4 . Here the difference between RE and choice of control is very clear . It can also be seen that within each category of control the relationship between RE and time to follow-up is more complex . For those studies with historic controls there is a very steep decline with time to assessment . For those studies with contemporary controls there is still a suggestion of a decline over the first 12 months then this levels out and possibly even reverses . This is reflected in the regression equation ( table 3 ) where Log RE is predicted by the time to assessment and time to assessment squared . Indeed , the polynomial equation of time to follow-up was able to explain 44% of the between study variance in the studies with only contemporary controls . Chemical spraying in the intervention but not control arms of the study were also included ( B = 0 . 54 , SE = 0 . 44 , p = 0 . 254 ) . Although this variable was not significant and subsequently dropped from the final model it is notable that the chemical spraying was if anything associated with reduced effectiveness of the community intervention .
The pooled results of the 22 studies in this meta-analysis suggest an important impact of educational messages embedded in a community-based approach on reducing larval indices . However , there was substantial heterogeneity in effect size between the different studies . This large heterogeneity in effect size reflects the very different study designs in the included studies . As discussed above the studies may or may not have included interventions additional to the community components , they may have used historic or contemporary controls , the controls may or may not have had some form of non-educational intervention . Consequently interpretation of the pooled effect size is difficult . However , the majority of the heterogeneity was explainable by just two variables , the choice of control and the time from intervention to assessment . The impact of choice of control was particularly marked with studies using historical controls finding much stronger effect sizes than those using contemporary controls ( table 2 ) . After adjusting for the time to assessment , anyone basing their judgement of the effectiveness of educational interventions based on historical controls would over-estimate the value of educational interventions by more than 10 fold compared to studies that used contemporary controls ( RE = 13 . 2 , 95%CI 4 . 1–42 . 5 ) . If the impact of one aspect of the study design is so great , it begs the question which is the correct study design to use . It could be argued both ways . In favour of the use of historic controls is the argument that at least the populations being compared are geographically the same . The arguments in favour of the contemporary controls include the fact that entomological indices may change from one time to another for reasons totally unrelated to the intervention . Indeed it could be argued that as interventions are usually implemented when the risk of dengue fever is particularly high , it is very likely that entomological indices will improve substantially whatever the intervention . The problems with historical controls are well known to medical researchers [16] , [17] . Indeed , the comment has been made that “most historical control groups are compromised for some reason” [18] . In studies with historic controls it has also been argued that the biases are worse the longer the time between the control and intervention . Our analysis would certainly support this suggestion for entomological control measures . Consequently we would argue that studies using historic controls be excluded from any assessment of the effectiveness of dengue vector control programmes . The polynomial relationship between RE and time to assessment is interesting . From the regression model of only those studies with contemporary controls , the highest effectiveness is seen somewhere around 18 months after the educational intervention . This is consistent with the suggestion that people may need time to learn , but after that their effort and good intentions may slip without reinforcement . As regards the value of non community interventions in addition to the community interventions , we found little evidence of any effect . In our single modifier analyses additional chemical application did appear of value when the control group did not receive chemical applications . However , in the model with control type and time to assessment , it was not significant . In the model including only studies with contemporary controls there was no evidence that chemical application provided any additional value over that achieved by education alone . However , there were only two studies that included chemical application in the intervention arm and not the control arm [19] , [20] . One of these studies supplied sand abate to the villagers , and the other used water treatment with temephos and outdoor spraying from the ground with malathion . Clearly , one cannot draw any definitive conclusions based on two studies in a meta-regression analysis . However , in this regard the study of Espinoza-Gomez and colleauges deserves special mention [11] . In this , well conducted study the authors randomly allocated houses to one of four intervention groups: no intervention , indoor chemical spraying only , education only and combined education and chemical spraying . For our meta-analysis , initialy compared education with no intervention and then compared education plus chemical with chemical only . In their original analysis using a two way ANOVA , the authors found that only education was effective at reducing larval indices and that chemical spraying gave no benefit either alone or in combination with education . In addition , a recent systematic review of the value of the effectiveness of peridomestic space spray also found little benefit of chemical spraying [21] . Why peridomestic spraying has uncertain benefit is unclear . With regard to comparison of educational interventions against one another , results showed that no single intervention modality ( such as the use of print or broadcast media , lectures , training by public health staff , home visits or targeting school children ) nor the number of different modalities used together was found to improve the RE significantly ( data not shown ) . However , few if any studies were designed to compare different educational intervention modalities and so we would argue that this issue remains unanswered until specific studies are designed to address the relative effectiveness of different educational modalities . Although meta-analyses of experimental studies such as randomised controlled trials are usually taken to provide high quality evidence of cause and effect , meta-regression analyses as presented here have the evidential status of observational studies . One of the other issues with meta-analyses of public health interventions is that often it is impossible to adequately blind the study participants or the study investigators . For example this has been raised as a major issue for studies of the effectiveness of household water treatment [10] , [22] . Wood et al . , in their research into evidence of bias associated with different study designs , found that lack of blinding could be associated with an apparent effect of about 30% for subjective outcomes [23] . For objective outcomes they found no evidence of such bias . Clearly the studies included in the analysis were not blinded . Whether or not the entomological indices are subjective or objective measures are open to debate . We would argue that these indices are semi-subjective and are potentially open to some form or bias due to lack of blinding of the assessors . However , even accounting for this the RE is much greater than could be explained purely by observer bias . It has been underlined by Erlanger et al . , that ‘relative effectiveness’ as numerical evidence of reduction of entomological measures , does not necessarily equate directly to reduction in pathogen transmission' [4] . Other factors such as villages sharing water supplies and garbage disposal may also enable disease transmission even if control within the village was good [11] . Gubler and Clark , in their review of dengue interventions as a whole , state that it must be kept in mind that in the types of community approaches assessed in this review , with these types of community approaches and strategies , it is expected that ordinary members of a community assume responsibility for activities that have historically been conducted by governmental bodies [3] . They suggest that for this reason , it would be very optimistic to expect immediate changes . In their systematic review of the literature , Erlanger et al . concluded that dengue vector control is effective in reducing vector populations [4] . Since that review three additional systematic reviews have been published none of which came to the same conclusion as Erlanger [7] , [8] , [21] . These later studies basically came to the conclusion that the quality of the published evidence , something that Erlanger et al . did not adequately address , was too poor to form a definitive conclusion . We would generally agree with the three later studies . In particular we have shown that a major problem is that studies with historical controls strongly overestimate RE compared to those with contemporary controls . Nevertheless , after accounting for the use of historic controls we still found evidence that supports the value of educational messages embedded in a community-based approach in reducing entomological indices of risk . We also showed that there is some evidence that the value of such interventions may decline after 18 to 24 months . With the evidence currently available it is not possible to say what types of educational modalities are most effective . There is a need to reassess whether other interventions add any further value to educational interventions . Finally , the issue also remains whether entomological indices alone is always translated into disease reduction . | Dengue fever is a mosquito-borne viral infection that is widespread in the tropics . Each year there are an estimated 50 million infections worldwide . Preventing infection relies on controlling the mosquitoes that spread disease . Unfortunately it is still not clear what does and does not work in the control of the mosquito vector . There have been several systematic reviews into control of dengue fever but still no consensus has been reached . This lack of consensus reflects the substantial heterogeneity in published effectiveness of studies of dengue control interventions . Prior systematic reviews have not adequately addressed this heterogeneity . We used multi-level modelling meta regression to investigate what variables modify the effectiveness of studies of educational messages embedded in a community-based approach . Most of the between study variation was explained by two variables , study design and time from intervention to assessment . In particular , studies using historic controls substantially overestimated the effectiveness of the intervention compared to those studies using contemporary controls . When the analysis was restricted to just those studies using contemporary controls , effectiveness was highest about 12 to 24 months after intervention but then declined . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"infectious",
"diseases",
"mosquitoes",
"health",
"screening",
"disease",
"ecology",
"public",
"health",
"and",
"epidemiology",
"global",
"health",
"dengue",
"fever",
"neglected",
"tropical",
"diseases",
"vectors",
"and",
"hosts",
"public",
"health"
]
| 2011 | The Value of Educational Messages Embedded in a Community-Based Approach to Combat Dengue Fever: A Systematic Review and Meta Regression Analysis |
Genome and exome sequencing in large cohorts enables characterization of the role of rare variation in complex diseases . Success in this endeavor , however , requires investigators to test a diverse array of genetic hypotheses which differ in the number , frequency and effect sizes of underlying causal variants . In this study , we evaluated the power of gene-based association methods to interrogate such hypotheses , and examined the implications for study design . We developed a flexible simulation approach , using 1000 Genomes data , to ( a ) generate sequence variation at human genes in up to 10K case-control samples , and ( b ) quantify the statistical power of a panel of widely used gene-based association tests under a variety of allelic architectures , locus effect sizes , and significance thresholds . For loci explaining ~1% of phenotypic variance underlying a common dichotomous trait , we find that all methods have low absolute power to achieve exome-wide significance ( ~5-20% power at α=2 . 5×10-6 ) in 3K individuals; even in 10K samples , power is modest ( ~60% ) . The combined application of multiple methods increases sensitivity , but does so at the expense of a higher false positive rate . MiST , SKAT-O , and KBAC have the highest individual mean power across simulated datasets , but we observe wide architecture-dependent variability in the individual loci detected by each test , suggesting that inferences about disease architecture from analysis of sequencing studies can differ depending on which methods are used . Our results imply that tens of thousands of individuals , extensive functional annotation , or highly targeted hypothesis testing will be required to confidently detect or exclude rare variant signals at complex disease loci .
To assess whether a single variant at a locus contributes to disease risk , the statistical analysis framework is relatively straightforward: compare the frequencies of alleles or genotypes at the site in relation to phenotype . To assess whether multiple variants in the same gene contribute to disease , a much larger array of potential genetic models must be considered . If the causal alleles are rare ( defined here as MAF<1% ) , then power to detect each variant’s effect individually is limited . For example , power to detect a variant with MAF = 0 . 5% and relative risk ( RR ) = 3 in 3K case-control samples ( 1 . 5K cases and 1 . 5K controls ) at α = 5×10-8 is ~5% [1] . Variants that are private to individuals , as some deleterious mutations are hypothesized to be , present greater challenges yet . As a result , numerous statistical methods have been developed in recent years to test aggregate groups of rare variants for association to disease [2–4] . Re-sequencing experiments have identified a handful of rare variants which modulate risk for common , complex diseases . Examples include variants in NOD2 for Crohn’s disease ( 4 variants with MAF 0 . 1–0 . 8% , ORs 1 . 4–4 . 0 , detected by single variant association ) [5] , PCSK9 for coronary heart disease ( 2 variants with MAF 0 . 8 and 1 . 8% , OR ~0 . 1 , detected by single variant association ) [6] , LPL for hypertriglyceridemia ( 154 missense variants with MAF<1% , present in cases , detected using the T1 gene-based association method ) [7] , and MTNR1B for type 2 diabetes ( 13 functionally-screened variants with MAF<0 . 1% , collective OR ~5 . 5 , detected using the KBAC gene-based method ) [8] . Each of these disease loci is characterized by different numbers , frequencies , and effect sizes of rare variants , but in each of these examples , the estimated proportion of phenotypic variance explained per locus is ~0 . 5–1 . 5% . As large-scale ( e . g . genome-wide or exome-wide ) studies are now being conducted in hundreds and thousands of individuals , several questions emerge . If loci similar to LPL or MTNR1B exist undiscovered across the genome , what is the power of different gene-based methods to detect them ? What effect sizes are studies of a given sample size well-powered to detect ? To what extent does power depend on the underlying architecture of causal allelic variation , and how should researchers navigate through the ensemble of available gene-based tests ? To interpret the results of gene-based association methods in sequencing studies , it is critical to quantify the power of each method to detect signals under a range of hypothesized locus architectures . Although the introduction of each novel gene-based association test has typically been accompanied by evaluation of the test’s performance alongside alternatives , each such analysis has compared different subsets of tests , made different assumptions about locus architecture and study design , and employed different simulation approaches . Comparative studies on the relative power of different methods [9–11] , while informative , have used small sample sizes , simulated limited locus architectures ( e . g . , with fixed numbers of causal variants ) that may not be representative of complex diseases , and considered only nominal levels of significance ( α>0 . 01 ) . Thus , further work is required to determine how different gene-based tests perform under different genetic models of complex disease . In this study , we systematically explore the power of eleven currently available and widely-used gene-based association methods to detect rare variant signals drawn from a range of principled genetic architectures of disease , in sample sizes consistent with those of ongoing re-sequencing studies . We assess the impact of locus architecture , effect size , and functional variant filters on the power of each method at stringent levels of significance . By evaluating all tests together at loci simulated under a range of continuous frequency-effect size distributions , we characterize each method’s success and failure modes , and describe genetic hypotheses for which particular methods may be better powered than others .
A key question for re-sequencing studies is: what is the power of gene-based association methods to detect causal loci at stringent levels of significance ? To address this , we ran each gene-based test at simulated loci explaining 1% of the variance in T2D liability [26 , 27] ( see Methods ) in 1500 cases and 1500 controls ( sample size comparable to several recent or ongoing complex trait sequencing studies [28 , 29] ) . Each gene-based test was run on all exonic variants ( causal and non-causal ) with MAF<1% , unless otherwise stated . The power of each test is shown as a function of significance threshold ( α ) and architecture in Figs 2 , S6 , and S7 . In the context of an exome-wide sequencing study , where an appropriate threshold may be α = 2 . 5×10-6 ( α = 0 . 05 , after Bonferroni correction for ~20K genes ) , we found that power is very low ( <20% ) across all architectures and tests considered . At a less stringent threshold of α = 10-4 , which might be used to nominate loci for follow-up ( under the null , only ~2 such genes would be expected exome-wide ) , power of the best performing tests across AR1-AR5 remained low ( 10–50% ) . This was true irrespective of the allele frequency threshold used for variant inclusion; results for a MAF threshold of 0 . 5% and 5% are shown in S8 Fig . We noted that at a nominal level of significance ( α = 0 . 05 ) , many methods had high power ( ~75%-95% ) to detect loci at which deleterious variants ( AR1-AR5 ) explain ~1% of phenotypic variance ( Figs 2 and S7 ) . KBAC was consistently the most sensitive method to detect deleterious effects at less stringent levels of significance ( up to 95% power at α = 0 . 05 , under AR4 ) . This high sensitivity could be useful in identifying putative signals when only a small number of hypotheses are being tested ( e . g . sequencing across only a few targeted loci ) , or to exclude rare variant models at candidate loci . Next , we asked whether any of the gene-based methods appear to be uniformly more powerful than others , across the various locus architectures we considered . Under simulated architectures where causal variants all have unidirectional ( deleterious ) effects ( Fig 2A , 2B , 2C , 2D , and 2E ) , we found that MiST , SKAT-O , and KBAC consistently achieve highest power , while UNIQ is least-powered . However , we did observe differential behavior of these tests depending on the significance threshold: MiST and SKAT-O retained greater power than unidirectional alternatives at stringent thresholds ( α<10-5 ) , while at less conservative thresholds ( α>10-3 ) , KBAC was more sensitive ( Figs 2A , 2B , 2C , 2D , 2E , 2F and S7 ) . We next sought to understand how power is influenced by locus architecture . Unsurprisingly , we found that power is higher when the majority of the locus’ total phenotypic effect is due to rare variants included in the association test ( e . g . those with MAF<1% ) . This is evidenced by the gain in power under models with a greater contribution of rare variants: the power of MiST , for example , increased from AR3 ( 10% at α = 10-4 in 3K individuals ) to AR2 ( 23% ) to AR1 ( 36% ) . Power was higher still under architectures where variants with MAF<1% ( i . e . those variants tested ) contributed all of the locus’ effect ( AR4 and AR5 ) : here , the power of MiST was ~50% at α = 10-4 . Power also depends on the direction of causal effects at a locus: under AR6 ( where both risk and protective effects are present ) , the variance-component tests ( SKAT and C-ALPHA ) and combined tests ( MiST and SKAT-O ) were least affected ( by design ) [21–24] and outperformed all the other methods , retaining ~10% power at α = 10-4 , while that of unidirectional tests was reduced to <5% ( Figs 2F and S7 ) . Finally , we find that power is inversely related to the degree of linkage disequilibrium between causal variants at a locus ( S9 Fig ) . We next queried the overlap between signals detected by gene-based methods versus those detected by single variant association . In direct contrast to gene-based methods , the power of single variant association decreased as the contribution of rare variants increased: power at a genome-wide threshold of α = 5×10-8 for single variants was ~20% , ~10% , and ~7% under AR3 , AR2 , and AR1 , respectively ( blue bars in Fig 3A , 3C , and 3E ) . However , in all cases , the combined application of gene-based and single variant methods yielded greater sensitivity than single variant association alone ( yellow bars in Figs 3A , 3C , 3E , and S10 ) . This occurred because the association tests detect distinct subsets of loci: gene-based methods uniquely identified loci where the signal was driven by groups of rare variants for which single variant association test statistics were not individually significant ( pink loci in Fig 3B , 3D , and 3F ) . As expected , the comparative advantage of gene-based tests was most evident under architectures where there is strong purifying selection against causal alleles ( under AR4 , for example , the power of single-variant tests at α = 5×10-8 was <5% , while gene-based tests achieved ~50% power at α = 10-4 , and ~20% power even at α = 2 . 5×10-6; S10A and S10B Fig ) . Under both AR2 and AR3 ( where limited purifying selection made causal alleles more common ) , the power of single variant association ( ~20% at α = 5×10-8 under AR3 ) exceeded that of the best gene-based test ( <5% at α = 2 . 5×10-6 under AR3 ) , though each method detected unique loci . These results confirm that single variant and gene-based association methods should be jointly employed for maximal power across divergent locus architectures . To characterize the impact of locus effect size on the power of gene-based tests , we simulated loci where the phenotypic variance explained ( VE ) by genetic variants is 0 . 5% , 1% ( as in Figs 2 and 3 ) , and 2% ( all under AR2 ) . At loci where VE = 2% , power increased to nearly 40% ( at α = 10-4 ) , as compared to ~23% when VE = 1% ( Figs 4A , S11A , and S11B ) . When VE = 0 . 5% , power was extremely low ( <8% at α = 10-4 in 3K individuals ) , indicating that exome-wide sequencing studies of this size are substantially under-powered to interrogate genes for weaker effects ( S11A Fig ) . The relatively modest power of gene-based tests at stringent levels of significance across the architectures considered here presents challenges to investigators seeking to discover novel disease-associated loci in studies of this size . Thus , we next investigated the extent to which power could be improved by a ) increasing sample size , or b ) excluding neutral variation at a locus . We found that gene-based methods exhibit differential gains in power as sample size increases from 3K to 10K individuals ( Fig 4B ) . The median power of MiST , for example , increased from ~23% to ~60% ( at α = 10-4 , under AR2 ) in 10K samples and was largely retained ( ~50% ) even at α = 2 . 5×10-6 ( S11C Fig ) . However , the increase in power was not uniform across methods . This occurred , in part , because ( unlike for single variant tests ) the relationship between sample size and power is not straightforward for gene-based tests: as sample size increases , causal alleles are observed more times , but the number of ( rare ) non-causal alleles also grows sharply . Thus , methods that up-weight all rare alleles regardless of their observed effect ( e . g . , FRQWGT ) may benefit least from increases in sample size ( S11–S13 Fig ) . As the number of observations of rare alleles increases with sample size , the performance of single variant association tests will certainly improve , but our analysis suggests that gene-based tests will still uniquely identify loci at which the aggregate signal is driven by variants too rare to be individually detected . When the top single variant in our simulated datasets had MAF < = 0 . 4% , the locus was rarely detected by single variant association in a sample of 3K individuals ( Fig 3B , 3D , and 3F ) . Single variant tests would have <80% power to detect an effect at a variant of that frequency ( at α = 5×10-8 ) even in 10K samples , unless the RR of that variant was over 3 . Moreover , as sample sizes increase , the threshold required to assess significance for gene-based methods will remain the same ( as the number of independent tests performed will not change ) , while that for single variant association tests will need to become more stringent as more novel variants are discovered . Hence , we expect the joint application of single variant and gene-based methods to remain beneficial even as sample sizes increase . Our study also confirmed that gene-based tests are highly sensitive to the fraction of neutral variation at a locus ( Figs 4C and S13 ) , as has been previously described [10 , 11 , 23] . We additionally found that unidirectional burden tests exhibit the sharpest increases in power as the fraction of neutral variation decreases . Under AR2 in 3K individuals , KBAC power at α = 10-4 exceeded 50% when only disease-causing variants were included ( increasing from ~22% prior to variant filtering ) . These tests may therefore be most powerful for testing targeted hypotheses at loci where rich functional annotation enables exclusion of a subset of neutral variants . Conversely , variance-component tests ( C-ALPHA , SKAT ) as well as combined methods ( MiST , SKAT-O ) are characterized by a relative immunity to neutral variation . This latter group of methods , then , are attractive options for jointly testing large numbers of less strictly filtered variants ( e . g . in a pathway-based analysis ) . We next investigated the degree of overlap between signals detected by each gene-based method . For each pair of association methods , we computed Pearson’s correlation coefficients between their reported p-values on a logarithmic scale ( Figs 5A , 5B , and S14 ) . We found that tests with similar design characteristics ( e . g . , SKAT and C-ALPHA , R2 = 0 . 99 ) exhibit very high correlation , as expected ( Fig 5C ) . Some methods were highly correlated , but there was variability in the p-values reported ( e . g . , MiST and SKAT-O , R2 = 0 . 92 ) , while others were much less related or even uncorrelated ( e . g . , SKAT-O and UNIQ , R2 = 0 . 02 ) . While in this latter case low correlation was driven by the lower mean power of UNIQ relative to SKAT-O , it is worth noting that there did exist a set of true causal loci ( where many case-private singletons segregate ) at which UNIQ reported p<10-4 , but SKAT-O reported p>0 . 01 ( Fig 5C ) . Other methods , such as SKAT and SKAT-O , showed asymmetric concordance ( R2 = 0 . 78 ) : SKAT-O detected a set of causal loci entirely undetected by SKAT , but was more conservative on the whole , reporting p-values up to an order of magnitude higher than those reported by SKAT at the majority of loci tested . These correlations were also architecture-dependent: under AR2 ( where there are only deleterious effects ) , for example , SKAT-O exhibited high concordance with KBAC ( R2 = 0 . 86 ) , while under AR6 ( where bidirectional effects are present ) , SKAT-O was most concordant with C-ALPHA and SKAT ( R2 = 0 . 93 ) . MiST shared this behavior , reflecting the ‘unified’ design of these tests as combinations of a unidirectional burden test and a bidirectional variance-based method [23 , 24] . To understand the drivers of such differences and identify scenarios where certain tests may be more powerful than others , we conducted pairwise comparisons between KBAC ( one of the highest performing methods at α = 10-4 across AR1-AR5 ) and the other gene-based methods . We focused here on loci where VE = 1% , simulated under AR2 . For each comparison , we characterized the properties of loci at which KBAC ( but not the other method ) reports p<0 . 01 , and vice-versa . In the comparison between KBAC and C-ALPHA ( Fig 6A ) , we found that loci at which only KBAC detected signal were characterized by a higher aggregate skew in case to control counts ( often driven by singletons , which do not contribute to the variance component tests’ dispersion statistic ) . Loci at which only C-ALPHA detected signal , on the other hand , were characterized by a relatively common single variant of large effect ( in the background of many variants with balanced case to control counts ) . For loci where the ratio of aggregate case to control counts is high , but no individual variants/genotypes show any substantial skew , the BURDEN test may be more powerful than KBAC ( Fig 6B ) . This makes sense: KBAC adaptively weights multi-site genotype counts by their observed case-bias , and if all variants have low weights , the maximum achievable KBAC statistic is low , whereas BURDEN quantifies the significance of the observed signal in aggregate . Finally , UNIQ ( unsurprisingly ) more readily detected loci at which signal is driven by either many rare variants private to cases , or by a single relatively frequent case-unique ( or control-unique ) variant ( Fig 6C ) . Taken together , these data indicate that although a given method may exhibit high mean power across divergent architectures , it may not be optimal for testing specific genetic hypotheses . Given the observation that different methods capture different signals , we wondered whether a strategy in which subsets of methods are collectively applied to a locus might be informative in an exome-wide setting ( e . g . , to test multiple hypotheses about locus architecture at once ) . To test this , we employed a stepwise forward selection approach , starting with each of the three best-performing gene-based methods across architectures ( MiST , SKAT-O and KBAC ) and using the degree of difference ( in orders of magnitude ) between additional methods’ reported p-values as the inclusion criterion ( see Methods , S1 Text ) . In 3K individuals , under AR2 ( where MiST power is ~23% at α = 10-4 ) , we found that particular combinations of tests ( e . g . , KBAC+MiST+VT+UNIQ+FRQWGT ) could jointly achieve ~31% sensitivity at α = 10-4 ( using the single minimum p-value reported across all three tests ) . However , this gain came at the cost of a higher false positive rate ( FPR ) : after adjusting the p-value significance threshold to correct for the increase in FPR , we found negligible gains in power compared to the application of a single test ( S3 Table ) . Joint application of gene-based tests may still be useful , however , in settings where a higher FPR is tolerable , e . g . , to increase sensitivity in a ‘discovery’ exome-wide sequencing scan which precedes large-scale targeted follow-up .
Given the wide array of aggregate rare variant association methods now available for application in re-sequencing or genotyping studies of complex traits [30] , it is critical to characterize and quantify the statistical power of each method to test heterogeneous genetic hypotheses . In this study , we conducted a comparative analysis of a panel of commonly used gene-based rare variant association tests under a broad range of realistic allelic architectures , significance thresholds , locus effect sizes , sample sizes , and filters for neutral variation . In sample sizes comparable to those of many contemporary sequencing studies ( 3K case-control individuals ) , we find that while gene-based association methods augment the power of single variant tests by preferentially detecting loci at which rare variants drive the causal architecture , their absolute power is low . All gene-based methods evaluated in this study have limited power , even to detect loci explaining as much as 1% of the variance in phenotypic liability underlying a common trait such as type 2 diabetes ( mean power across architectures is ~5–20% at α = 2 . 5×10-6 ) . Even in 10K case-control samples , power remains modest ( ~60% at α = 2 . 5×10-6 ) . Based on estimates of variance explained by known rare and common variant signals ( the strongest single common variant association for T2D , mapping near TCF7L2 , explains ~1% of phenotypic variance ) , it seems probable that for any given complex trait , at best a handful of loci will have effects on this scale . The full potential of exome sequencing to provide biological insights into disease , then , will depend largely on the detection of loci of smaller aggregate effects , and will require far larger sample sizes than these . The low mean power to detect disease-associated loci prompted the question of whether some methods are better powered than others to discover novel signals under specific hypothesized locus architectures . We find that at more stringent significance thresholds ( α<10-4 ) , MiST and SKAT-O have the highest power across architectures simulated here , especially when rare variants have bidirectional effects on disease . Thus , for investigators looking to discover signals across thousands of loci ( e . g . , in exome-wide scans ) , these tests are likely to maximize sensitivity . Weighted sum methods ( and KBAC in particular ) , on the other hand , are consistently best-powered to detect rare variants of deleterious effect at less stringent levels of significance , and also show the greatest gains in power when neutral variation can be filtered out . These attributes may be useful in various scenarios: to test a small number of biological hypotheses ( e . g . at only a few loci , especially if functional annotations are available ) , to prioritize signals for further follow-up from a discovery scan , or to place bounds ( e . g . , after an exome-wide sequencing study ) on the total number of genes harboring rare variants of a given effect size that are likely to exist . In addition to MiST , SKAT-O and KBAC , we find that other methods may have individual strengths under particular scenarios ( e . g . , UNIQ to test whether a gene harbors an excess of highly penetrant rare variants , or BURDEN to detect a collection of variants each of very weak effect ) ; these methods may be optimal for testing such specific genetic hypotheses . Finally , in larger sample sizes ( n = 10K case-control individuals ) , our simulations demonstrate that the increasing number of neutral ( non-causal ) rare variants may limit gains in the power of some methods ( e . g . FRQWGT ) . Here , MiST is best-powered at stringent significance thresholds . Taken together , these results suggest that the interpretation of novel signal discovery ( or the lack thereof ) in sequencing studies may vary based on the specific gene-based methods that are used . This study has a number of limitations . It is based on simulated data ( albeit data consistent with available empirical information on genetic variation and disease epidemiology [15] ) . It does not explore the effects of properties such as demographic history , gene size , mutation rate , haplotype length , or degree of linkage disequilibrium between causal variants on the power of gene-based association methods . Moreover , it does not characterize the performance of these methods at non-coding regions , where causal variant frequencies and effect sizes may be different , and where there is likely a higher proportion of neutral variation . This simulation approach , however , enabled us to undertake a controlled , quantitative characterization of the performance of gene-based association methods under a range of scenarios . Future work should characterize these methods in study populations of different ethnicities , where different site frequency spectra and linkage disequilibrium patterns between causal variants may alter power ( S9 Fig ) . Architectures we simulated assumed a common binary trait; power to detect loci explaining phenotypic variance for less prevalent traits is likely higher , but we did not study this relationship . The tools available on our website ( http://mccarthy . well . ox . ac . uk/publications/2014/moutsianas_simulations/ ) allow the investigation of this question for any complex trait by generating simulated data using a custom , user-specified RR-by-allele frequency heat-map and disease prevalence . In summary , we find that specific gene-based association methods are best deployed in the setting of particular experimental study designs , and when testing for particular genetic models of disease . Such an approach will likely enable meaningful interpretation of both positive and negative findings in ongoing sequencing studies , and is bound to remain important even as sample sizes increase and new statistical methods for aggregate testing of rare variants are developed .
Simulated datasets were generated using HAPGEN2 [12] . HAPGEN2 generates case-control data using a haplotype reshuffling approach based on the Li & Stephens model [31] . Under this model , simulated ( unobserved ) haplotypes are assumed to be an imperfect mosaic of actual ( observed ) haplotypes and are simulated using a Hidden Markov Model with recombination and mutation rates as parameters . Case and control samples are generated by over-sampling haplotype segments which contain alleles at which phenotypic effects are introduced ( based on the relative risks assigned to them ) . A phased reference panel of haplotypes from 379 European ( 98 TSI , 89 GBR , 85 CEU , 14 IBS , and 93 FIN ) individuals from the 1000 Genomes Project ( 1000G Project Phase 1 , release 3 ) [13] was augmented to 12 , 514 individuals by iteratively simulating haplotypes ( with no phenotypic effects ) and adding them to the original reference panel , in increments of 300 individuals per iteration . An excess of rare variation was introduced to the data using an empirically selected value of θ = 0 . 08 for the mutation parameter in HAPGEN2 , so as to match the singleton count observed in empirical re-sequencing data in a sample of this size . We used the SFS reported by Nelson et al [14] , which was based on sequencing 351kb of coding sequence in 12 , 514 samples of European descent . The resulting dataset was subsequently thinned using a rejection sampling approach , to match the full site frequency spectrum observed in real data . This two-step approach ( matching for singletons , and then thinning the dataset ) was necessary to model the excess in rarer variation observed in whole exome sequencing datasets while preserving the LD structure of the reference panel . In order to validate that this approach led to a realistic SFS when sub-sampled to smaller sizes , we compared the SFS observed in the simulated , thinned panel , in subsets of 2 , 738 individuals , to that of empirical exome-wide sequencing data on the same number of individuals , from the GoT2D project ( dark and light blue lines , Fig 1 ) . Forward population genetic simulations of global complex disease architecture ( specifically , for type 2 diabetes , a disease of prevalence 8% and heritability ~45% ) were conducted across a range of disease models varying in mutational target size and coupling to purifying selection [15] . By varying only these two parameters , a wide range of continuous joint frequency and effect size distributions were generated; under models with strong coupling to selection , rare variants explain the bulk of heritability and have large effects , while under models with weak coupling to selection , common variants explain the bulk of heritability and rare variants have weaker effects . For the HAPGEN2 simulations conducted here , we sampled variant effect sizes from the distributions observed in the forward simulated datasets at loci explaining ~1% of phenotypic variance underlying T2D ( S3 Fig ) . Variant effects were selected from the frequency-effect size distributions described above . We simulated these effects at randomly selected exonic variants across each gene . We used variant frequencies measured in the augmented reference panel of 12 , 514 individuals . In unidirectional architectures , all rare variants were assumed to increase risk of disease ( RR>1 ) . In bidirectional architectures , protective effects were sampled in the same way , but the relative risks were inverted . Variant effects were sampled until the cumulative variance explained ( VE ) on the liability scale by each locus reached the desired threshold ( e . g . VE = 0 . 5% , 1% , or 2% ) . The following procedure was followed for introducing variation at each locus: Pick an exonic variant at random Introduce an effect by sampling from the frequency-RR distribution of the respective architecture If the cumulative variance explained ( on the liability scale , %VE ) by variants at the locus is below of the specified threshold , go to step ( i ) and repeat If the variance is above the specified threshold , remove one of the introduced effects ( at random ) and go to step ( i ) If the cumulative variance explained is close enough to the specified threshold ( 0 . 95*VE , 1 . 05*VE ) , then If the number of introduced variants is over 35 , quit and restart , else: Accept the sampling and simulate data using the variants and effect sizes chosen , using HAPGEN2 . The upper bound of 35 on the total number of causal variants introduced per locus was imposed due to instability in HAPGEN2 behavior above this threshold; this limit was rarely reached in 3K samples , but it did restrict architectures simulated in 10K samples ( S11C and S12 Figs ) . The calculation of variance explained at each locus was conducted using the method described by So et al . , which is available online as an R script [26] . This calculation requires three parameters as input ( per variant ) : the prevalence of the trait ( in this case assumed to be 8% , to model type 2 diabetes ) , the population frequency of the risk allele , and the genotype relative risk . We assumed independence between risk variants at a given locus , and thus estimated the total percentage of variance explained as the sum of the variance explained by each individual variant . The latest releases of the PLINK/SEQ ( v0 . 09 ) [18] and EPACTS ( v3 . 2 . 3 ) [25] software packages were used to run ten of the gene-based methods evaluated in this study . MiST was run using a publicly available R package ( http://cran . r-project . org/web/packages/MiST/index . html ) [24] . All exonic variants ( causal and non-causal ) below varying minor allele frequency thresholds ( 1% for all analyses discussed in the main text , unless otherwise stated ) were included in the tests , except when the fraction of neutral variation was varied . In this case , the proportion of causal variants included in the test was fixed to 0 . 25 , 0 . 50 , 0 . 75 , or 1 ( Fig 4C ) . The subsets of tests chosen for inclusion into composite tests were selected using a stepwise forward selection approach . Starting with a single test ( three runs per architecture , each starting with one of the top three performing tests across architectures , MiST , SKAT-O and KBAC ) , the next test to be included at each step was the one which reported the greatest number of novel signals , i . e . not previously detected by the tests already included . Novel signals were defined as loci for which the p-value reported by the candidate test for inclusion was lower by a specified multiplicative “margin” ( factor ) than the lowest p-value reported by tests already included in the composite test . Three margins were used ( 100 , 10 , and 1 ) ; a margin of 100 , for example , implies that for signals to be considered novel , they p-value of the candidate test needs to be two orders of magnitude lower than the lowest of the ones already included in the composite test . All datasets discussed in this study , together with the scripts used to generate them and results of both single variant association and gene-based methods across all architectures , are available on the website http://mccarthy . well . ox . ac . uk/publications/2014/moutsianas_simulations/ . The website also contains the software used for the script generation ( a wrapper for HAPGEN2 [12] ) , which can be used to generate analogous simulated data for the genes we included in the manuscript under alternative scenarios/architectures . | Re-sequencing technologies allow for a more complete interrogation of the role of human variation in complex disease . The inadequate power of single variant methods to assess the role of less common variation has led to the development of numerous statistical methods for testing aggregate groups of variants for association with disease . Such endeavors pose substantial analytical challenges , however , due to the diverse array of genetic hypotheses that need to be considered . In this work , we systematically quantify and compare the performance of a panel of commonly used gene-based association methods under a range of allelic architectures , significance thresholds , locus effect sizes , sample sizes , and filters for neutral variation . We find that MiST , SKAT-O , and KBAC have the highest mean power across simulated datasets . Across all methods , however , the power to detect even loci of relatively large effect is very low at exome-wide significance thresholds for sample sizes comparable with those of ongoing sequencing studies; as such , the absence of signal in studies of a few thousand individuals does not exclude a role for rare variation in complex traits . Finally , we directly compare the results reported by different gene-based methods in order to identify their comparative advantages and disadvantages under distinct locus architectures . Our findings have implications for meaningful interpretation of both positive and negative findings in ongoing and future sequencing studies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| []
| 2015 | The Power of Gene-Based Rare Variant Methods to Detect Disease-Associated Variation and Test Hypotheses About Complex Disease |
Hansen’s disease ( HD ) , or leprosy , is still considered a public health risk in much of Brazil . Understanding the dynamics of the infection at a regional level can aid in identification of targets to improve control . A compartmental continuous-time model for leprosy dynamics was designed based on understanding of the biology of the infection . The transmission coefficients for the model and the rate of detection were fit for each region using Approximate Bayesian Computation applied to paucibacillary and multibacillary incidence data over the period of 2000 to 2010 , and model fit was validated on incidence data from 2011 to 2012 . Regional variation was noted in detection rate , with cases in the Midwest estimated to be infectious for 10 years prior to detection compared to 5 years for most other regions . Posterior predictions for the model estimated that elimination of leprosy as a public health risk would require , on average , 44–45 years in the three regions with the highest prevalence . The model is easily adaptable to other settings , and can be studied to determine the efficacy of improved case finding on leprosy control .
Hansen’s disease ( HD , or leprosy ) is a chronic progressive disease caused in Brazil by infection with Mycobacterium leprae . Transmission is most likely through nasal droplets [1] , and is associated with socioeconomic status [2] . While leprosy is curable through chemotherapy [3] , detection is often delayed [1] , leading to more serious sequelae ( including disfigurement and disability ) . The World Health Organization ( WHO ) has set a goal for elimination of leprosy as a public health problem , defined as a prevalence of <1/10 , 000 [4] . Several countries , including Brazil , have failed to meet that goal [5] . The Brazilian leprosy control program has been successful in decreasing the incidence of leprosy , but the prevalence remains high in 2 regions [5] . Movement towards elimination seems to have stagnated in these regions , possibly due to a downgrading of the importance of case finding [6] . Treatment of leprosy has been decentralized , so regional differences in case detection , and disease progression are to be expected [7] . Infection hotspots have also been noted in Brazil [8] , leading to regional and sub-regional differences in transmission rates [5] . These may be related to socioeconomic factors , as a systematic review has found that socioeconomic inequalities associated with leprosy were large [9] . Prediction models must take these regional differences into account in order to accurately represent these differences and identify possible control points . A number of models of leprosy have been proposed [10–21] , and 6 of the base models from these studies were recently fitted to regional data from Brazil [22] . However , only one model takes into account much of the recent research on leprosy susceptibility [10 , 18] , and it is an agent-based model that relies on specific population structures; the results of this model are quite useful on a regional level [23 , 24] , but have not been applied to national-level results . The goal of this research is to produce a compartmental model that represents current understanding of leprosy susceptibility and pathogenesis , but that is also easily adaptable to different populations . Unknown parameters for this model will be fitted to regional incidence data from Brazil and analyzed to determine differences in control efficacy and their effect on the elimination target and long-term control .
All human data was anonymized at the source before usage [25] . As these data were publicly available and fully anonymized , no institutional review board approval was required . A deterministic compartment model of leprosy ( Fig 1 ) was designed to take into account current understanding of the disease . Briefly , individuals are divided into 3 categories: resistant ( R , with probability qS = 1-ps ) , susceptible to paucibacillary infection ( SP , with probability ps*pp ) , and susceptible to multibacillary infection ( SM , with probability ps*qp where qP = 1-pp ) . Resistance , qs , is meant to convey both genetic resistance and socioeconomic protective factors [13 , 18] . Resistance to multibacillary infection , pp , is meant to convey genetic resistance [18]; this value is higher than the observed proportion of new cases that are PB ( 0 . 8 vs . 0 . 54 ) , but the discrepancy is explained by the high rate of self-cure among PB cases ( αPN ) . Resistant individuals ( R ) enter and leave the population without infection . Individuals with susceptibility to leprosy but genetic resistance to MB disease enter the paucibacillary ( PB ) track as susceptible ( SP ) . They may be exposed ( EP ) at rate λ and eventually develop symptomatic PB disease ( NP ) at rate γP . Paucibacillary disease either self-heals at rate αPN or is detected and leads to treatment ( TP ) at rate φP , either of which results in recovery ( RP ) at rate αPT . Recovered individuals may relapse to refractory disease ( AP ) at rate σP , from which they can be detected and return to treatment at rate φP . Individuals with a genetic susceptibility to MB disease enter the population as susceptible ( SM ) and may become exposed ( EM ) at rate λ . Exposed individuals develop multibacillary disease ( NM ) at rate γM and are diagnosed and entered into treatment ( TM ) at rate φM . Treated individuals recover or leave treatment ( RM ) at rate αM and may relapse to refractory disease ( AM ) at rate σM , from which they may be detected and return to treatment at rate φM . Multibacillary individuals are subject to a death rate that is proportionately higher than the general population , at νM for untreated individuals and νMT for treated individuals . The force of infection , λ , is calculated as λ=βP ( NP+θ1TP+AP ) +βM ( NM+θ2TM+AM ) N ( 1 ) for the density-dependent model , where N is the total population size , βP is the contribution of PB individuals to the force of infection , βM is the contribution of MB individuals to the force of infection , θ1 and θ2 are the proportional decreases in infectiousness due to treatment of PB and MB individuals , respectively . Total population size ( N ) is calculated as the sum of all the compartments , and varies over time as births and deaths occur . For the sake of simplification , it is assumed that treated individuals are quickly rendered non-infectious [3] and therefore that θ1 = θ2 = 0 . Thus , the force of infection becomes λ=βP ( NP+AP ) +βM ( NM+AM ) N ( 2 ) in the density-dependent model . Good estimates were available in the literature for most system parameters ( see Table 1 ) . However , estimates were unavailable for the transmission coefficients ( βP and βM ) and the true case detection rates ( φP and φM ) , as these are likely to vary by locality and can be difficult to measure directly . These parameters were therefore estimated using the Sequential Monte Carlo approximate Bayesian Computation ( SMC ABC ) algorithm [26 , 27] applied to Brazilian incidence data , as described previously [22] . Briefly , the annual incidence of PB leprosy , incobsPB ( y ) , and the annual incidence of MB leprosy , incobsMB ( y ) , were obtained for each of the 5 regions of Brazil between 2000 and 2012 [28] . For each region , an initial parameter set ( n = 100 , 000 ) was sampled from the prior distributions of the estimated parameters ( a uniform distribution with a non-informative range , Table 1 ) , and in subsequent SMC particles ( rounds ) , parameter sets were sampled from the immediately previous particle with a perturbation kernel . Each parameter set was used to simulate the incidence values between 2000 and 2010 ( allowing 2011–2012 to be used for unconstrained validation ) , where incidence was assumed to be new individuals entering the treated category ( TP or TM ) from the untreated category ( NP or NM ) . A distance function was calculated using the equation d=∑y=20002010 ( incobsPB ( y ) −incpredPB ( y ) ) 2+ ( incobsMB ( y ) −incpredMB ( y ) ) 2 ( 3 ) where incpredPB ( y ) was assumed to be equal to φPNP ( y ) , or the number of new PB cases entering treatment in year y , and incpredMB ( y ) was assumed to be equal to φMNM ( y ) , or the number of new MB cases entering treatment in year y . It was assumed that recurrent infections ( from AP or AM ) were not included in the observed incidence . A parameter set was accepted if d<τ , where τ was set equal to the 75th percentile of d in the previous particle . In each particle , the algorithm was repeated until 100 , 000 parameter sets were accepted; 10 particles were produced , with the 10th particle used to form the posterior distribution . The perturbation kernel was set to be a uniform distribution with a range limited by ± the variance of each parameter in the previous particle . Initial values in each of the compartments were determined analytically based on the parameters and observed prevalence . As the duration of treatment in MB disease is twice the duration of treatment in PB disease under MDT , the observed prevalence was assumed to be divided between TM ( 0 ) and TP ( 0 ) with the ratio 2*incidence ( MB ) incidence ( PB ) . The ratio of Ni ( 0 ) :Ti ( 0 ) , Ei ( 0 ) :Ni ( 0 ) , Ri ( 0 ) :Ti ( 0 ) , and AM ( 0 ) :NM ( 0 ) , where i ∈ {M , P} , were set empirically in a multi-step process similar to that described previously [22] . Briefly , the ratios were adjusted manually for each region such that the model , simulated with the assumed values in Table 1 , predicted the incidence of both PB and MB cases in that region with less than 10% deviation from the observed values in 2000 ( the first year of observation ) and 2002 ( the year of peak incidence in most regions ) . The model was then fitted with the initial population distribution determined by these ratios , and the median of the estimated distribution for each fitted parameter was used to predict incidence of both PB and MB cases in each region . If the predicted incidences in 2000 or 2002 deviated from observed values in any region by more than 10% , the ratios were again adjusted manually to correct the deviation and the model was re-fitted . This process repeated until the median of the estimated distribution for each fitted parameter was able to predict incidence of PB and MB cases in each region with less than 10% deviation from observed values in both 2000 and 2002 . As density-dependent transmission was assumed , but is known to be a simplification of true human contact rates [32] , the above process was repeated for a frequency-dependent transmission model . In this model , the force of infection λf becomes λf=βPf ( NP+AP ) +βMf ( NM+AM ) ( 4 ) where βPf and βMf are adjusted from the density-dependent model to account for population size . The results of the frequency and density dependent models were compared using Bayes factor analysis , in which the Bayes factor was the ratio of the summed distance in all regions , corrected for differences in regional population size , across a weighted sample of 1 , 000 posterior parameter sets . The results of the best-fitting regional model ( frequency or density dependent ) were examined for similarity between distributions , and 3 hierarchical fittings were considered: transmission parameters ( βM , βP ) shared across regions ( V1 ) , transition parameters ( φM , φP ) shared across regions ( V2 ) , all parameters ( βM , βP , φM , φP ) shared across regions ( V3 ) , and sharing no parameters ( the regional model described above , V4 ) . In the hierarchical models , the distance function was altered to d=∑r∑y=20002010 ( incobsPB ( y , r ) −incpredPB ( y , r ) ) 2+ ( incobsMB ( y , r ) −incpredMB ( y , r ) ) 2/Nr ( 5 ) where r represents the region and Nr is the population of region r in 2000 . Hierarchical models were compared to each other and the regional model using Bayes factor analysis , in which the Bayes factor was the ratio of the summed distance in all regions , corrected for differences in regional population size , across a sample of 1 , 000 posterior parameter sets weighted by the inverse of their summed regional distances ( Eq 5 ) . In order to check the consistency of the model results , data were simulated for each region using the median of the best fitted value from the preferred hierarchical model . These data were then used to repeat the full model selection and parameterization process , including hierarchical model selection and parameterization . Results were compared to the simulated input values . Posterior predictions were produced using a weighted sample of 1 , 000 parameter sets from the posterior distribution of both the preferred hierarchical model and the regional model , and outcomes of interest were predicted from this sample . Outcomes were the time to elimination in years ( telim ) and the predicted incidence overall and of MB and PB cases in the year 2050 ( i2050 , iM2050 , and iP2050 , respectively ) . All models and fitting were performed in R 3 . 0 . 3 , [33] which was accessed through the Revolution R Analytics interface ( copyright 2014 Revolution Analytics , Inc . ) .
The median and range of each parameter for the each of the model fits are shown in Table 2 . Transmission and transition parameters were similar between the density-dependent and frequency-dependent models . Bayes factor analysis identified the frequency-dependent model as having the lowest summed deviance from the observed incidence . As a result , the frequency-dependent model was used for hierarchical model fitting . Bayes factor analysis identified the regional version as having the lowest summed deviance from the observed incidence , although the preference for the regional version was not strong ( Bayes Factor of 1 . 2 to 2 . 1 , compared to the hierarchical models ) . Transmission coefficients were estimated to be similar in all regions even in the regional model , but transition rates had high variability between regions . Transition rates were lower in the Midwest for all individuals , higher in the South and Southeast for MB individuals , and higher in the North for PB individuals . The final distribution of the initial population , as determined by the empirical process , is shown in Fig 2 for each region and model fit . The number of latent and undetected individuals was the most variable across models , with the density-dependent model requiring higher numbers of latent individuals to reproduce the initial and peak incidence in each region . Posterior predictions for the preferred hierarchical model and the regional model are shown in Fig 3 and Table 3 . The results show that the fit underestimated PB incidence in the North and Midwest and MB incidence in most regions in later years . The South and Southeast reported incidences below the elimination threshold in 2001 , and this was also predicted to be possible by the model , although the average time to elimination was predicted to be 2002 and 2007 , respectively . On average , the North and Midwest were predicted to reach the elimination threshold by 2045 , while the Northeast was predicted to reach the elimination threshold by 2044 . However , the ranges of values were wide , indicating that the Northeast could require up to the year 2053 to reach the elimination threshold . Simulation results ( Table 4 ) show that the model was able to predict the simulated values in most cases . The exceptions were the values of φM and φP , which tended to overestimate the true value , and the values of βM and βP in the density-dependent model , which tended to underestimate the true value .
This study presents a compartmental model for Hansen’s Disease that takes into account the current understanding of the disease but that is computationally simple and easy to adapt . The structure of this model differs from that of Meima et al . [11] in 2 ways . First , this model assumes that the 90% of people who never develop leprosy are inherently resistant , rather than self-healing . Second , this model assumes that the 80% of susceptibles who will only develop PB disease are again inherently resistant to MB disease . That allows us to separate the relapsed cases appropriately , such that only those recovered from MB disease relapse to MB disease . These assumptions also allow for including differences of susceptibility in a compartmental model , which is less computationally intensive , easier to fit to data , and easy to adapt to other populations or across larger and more diverse regions . However , Ridley-Jopling classification allows for borderline cases which can cross between PB and MB groups during relapse . [34] While it would be advantageous to capture the full diversity of leprosy presentation in future models , the parameterization of such models will require improved reporting; at present , Brazil reports only the PB and MB classifications of detected cases . [25] Many studies at the national or regional level will report a “case detection rate” . This , however , is not the true case finding rate represented by φM and φP in this model . It is , instead , the annual observed incidence , and uses the population as the denominator , not the delay in diagnosis . This is an important difference: a high “case detection rate” could indicate an outbreak rather than fast detection , while the true case finding rate only represents the time necessary to identify a clinical case . We found that the range of potential case finding rates was fairly similar across most regions , but that cases were estimated to be infectious for 5 years in most cases before treatment was initiated . This agrees with the findings of multiple studies [35 , 36] , where people delayed seeking treatment partly from an assumption that the symptoms were not serious and potentially from a worry of the stigma attached to a diagnosis [37] . It also falls into the range assumed by previous models within Brazil for this time period [24] . In a modeling study , Fischer et al . [10] found that contact tracing was important to avoid the diagnostic delay; contact tracing was decreased in India in order to meet WHO case detection rate goals , and the result was treatment delays [38] , which would produce a long-term effect of higher incidence and , eventually , higher case detection rates . Importantly , our model predicted that cases in the Midwest were infectious for an average of 100 years before detection , an unrealistic value indicating that the true detection rate of cases is too low to estimate properly . If true , this could indicate a public health failing that should be addressed . This difference in detection rate between regions could be socio-economic in origin , as the three regions predicted to have low detection rates ( Northeast , North , and Midwest ) also have lower GDP per capita . Similar regional health disparities have been noted for ischaemic heart disease [39] and laryngeal cancer [40] mortality , suggesting regional disparities in health care [41] . The decision was made to compare density- and frequency-dependent models , despite the fact that leprosy is considered to be a disease of close contact and therefore would be classically considered to have density-dependent transmission . This is due to the limitations of a model such as this , caused by the homogeneous mixing assumption , in capturing the limited number of close contacts any individual is likely to have . Thus , while a disease may be truthfully density-dependent , it may behave mathematically as a frequency-dependent disease . The results of this study show that the density-dependent model was slightly preferred to the frequency-dependent model . This question would not arise with an individual-based model , such as SIMCOLEP [42] , but those models are not as easy to adapt as they must rely on population-specific characteristics in their design , which may require parameters that are not locally available . The goal of this study was to provide an adaptable model that was still able to capture the regional dynamics of leprosy spread . The preference of the regional model supported this decision , but the preference was not strong , indicating that some national-level models may be as informative as the regional models . Several models assume that PB individuals are non-infectious [11] . We observed that PB individuals did contribute to the force of infection , although with roughly half to two-thirds the strength of MB individuals . In other mycobacterial diseases , less infectious individuals have been found to be potentially important in maintaining the endemicity of the infection [43 , 44] . This highlights the importance of diagnosing and treating all cases; although the PB cases do not have as serious sequelae as the MB cases , they may serve to maintain the infection in a region . With regards to the intra-regional variation in transmission parameters , we found that higher transmission parameters were predicted in regions with higher incidence . This is to be expected , and highlights the ability of the model fitting to identify regional differences . The results of the scenario analysis show that all regions are well on-track to eliminate leprosy , with the South and Southeast , which have the lowest incidence , likely to eliminate leprosy first . The posterior prediction plots ( Fig 2 ) show that the fitted model estimated the observed decrease in PB incidence fairly accurately in most regions , but slightly overestimated the decrease in incidence of MB cases in all but the Midwest . It may be assumed , therefore , that these results are best-case scenarios for those regions . The predicted incidence in the Northern region is higher than has recently been predicted for Para State , which is within that region , but the time to the elimination target generally agrees between the two models [24] . All regions observed an increase in incidence up to 2003 , followed by a slow decrease . This is likely due to the slow impact of control programs , rather than a change in case detection rates; chronic diseases with long latent periods , like leprosy , will require consistent control over long periods of time to reverse incidence trends . It is important to note that the incidence of MB disease was predicted to be in the range of 13 , 983 to 14 , 913 new MB cases in Brazil in 2015 . However , the North , Northeast , and Midwest are likely to require a much longer period to reach official elimination than the South and Southeast . The best way to decrease the time to elimination would be increasing the case finding rate [45] . This would also improve the level of disability in new cases , as delay in onset of treatment is a major cause of disability . Our results , therefore , indicate that the North , Northeast , and Midwest regions of Brazil would benefit from improving the true case finding rate , which we have estimated to be slow . | Control of Hansen’s disease , or leprosy , requires understanding how quickly the infection moves through the population and how long it takes to detect the disease . These rates vary regionally , resulting in differences in the number of people detected with disease each year . We have estimated the risk of infection and the rate of detection for this disease in each of the 5 regions of Brazil . This allowed us to predict the long-term impact of Brazil’s current leprosy control program , which found that some regions of Brazil will require 44–45 years to eliminate leprosy as a public health risk , primarily due to the long delay in detection of cases . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"death",
"rates",
"pharmacologic",
"analysis",
"mycobacterium",
"leprae",
"medicine",
"and",
"health",
"sciences",
"demography",
"tropical",
"diseases",
"geographical",
"locations",
"factor",
"analysis",
"bacterial",
"diseases",
"compartment",
"models",
"mathematics",
"statistics",
"(mathematics)",
"neglected",
"tropical",
"diseases",
"pharmacology",
"population",
"biology",
"infectious",
"disease",
"control",
"bacteria",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"south",
"america",
"mathematical",
"and",
"statistical",
"techniques",
"actinobacteria",
"brazil",
"people",
"and",
"places",
"population",
"metrics",
"pharmacokinetic",
"analysis",
"leprosy",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"statistical",
"methods",
"organisms"
]
| 2016 | Proposing a Compartmental Model for Leprosy and Parameterizing Using Regional Incidence in Brazil |
The reconstruction of ancestral genome architectures and gene orders from homologies between extant species is a long-standing problem , considered by both cytogeneticists and bioinformaticians . A comparison of the two approaches was recently investigated and discussed in a series of papers , sometimes with diverging points of view regarding the performance of these two approaches . We describe a general methodological framework for reconstructing ancestral genome segments from conserved syntenies in extant genomes . We show that this problem , from a computational point of view , is naturally related to physical mapping of chromosomes and benefits from using combinatorial tools developed in this scope . We develop this framework into a new reconstruction method considering conserved gene clusters with similar gene content , mimicking principles used in most cytogenetic studies , although on a different kind of data . We implement and apply it to datasets of mammalian genomes . We perform intensive theoretical and experimental comparisons with other bioinformatics methods for ancestral genome segments reconstruction . We show that the method that we propose is stable and reliable: it gives convergent results using several kinds of data at different levels of resolution , and all predicted ancestral regions are well supported . The results come eventually very close to cytogenetics studies . It suggests that the comparison of methods for ancestral genome reconstruction should include the algorithmic aspects of the methods as well as the disciplinary differences in data aquisition .
The reconstruction of ancestral karyotypes and gene orders from homologies between extant species is a long-standing problem [1] . In the case of mammalian genomes , it has first been approached using cytogenetics methods [2]–[7] . The recent availability of sequenced and assembled genomes has led to the development of bioinformatics methods that address this problem at a much higher resolution , although with fewer available genomes . Such methods propose in general more detailed ancestral genome architectures than cytogenetics methods ( see [8]–[12] and reviews in [13]–[15] ) . The comparison of the two approaches was recently investigated and discussed in a series of papers , sometimes with diverging point of views [16]–[18] . Among the bioinformatics methods that have been applied to mammalian genomes ( previous works were limited to small genomes such as organellar genomes [19] or to bacterial genomes [20] ) , the one based on a parsimony approach in terms of evolutionary events such as reversals , translocations , fusions and fissions [8] , [11] , leads to results that are sometimes in disagreement with cytogenetics studies [16] . Recent results on this approach point out that the modeling of genome rearrangements probably needs further studies before it can be used for the reconstruction of ancestral genomes ( see [21] , or [17] , where it was suggested that inferring parsimonious rearrangement scenarios is more intended to infer evolutionary dynamics characteristics , such as rearrangement rates , than ancestral genomes ) . Another type of approach infers ancestral genome segments , called Contiguous Ancestral Regions ( CARs ) , from syntenic features that are conserved in extant species ( the terminology is borrowed from [12] ) . We call this principle model-free , following [22] , even if it is based on certain assumptions , like the absence of events inside a conserved synteny , which is a parsimony principle . But this terminology stresses the difference with rearrangement-based methods , which contraint the reconstruction by allowing prescribed operations that define then an evolution model . It is then less ambitious than the rearrangement-based approach as it does not propose evolutionary events , neither does it ensure that proposed CARs are ancestral whole chromosomes . However , when recently applied on mammalian genomes [12] it gave results more in agreement with cytogenetic methods , while exhibiting few other points of divergence [18] . We describe here a very general model-free framework for the reconstruction of CARs , that formalizes and generalizes the principles used in several computational [12] , [22] and cytogenetics [5]–[7] studies . This framework takes as input a representation of extant genomes as sequences of homologous genomic markers ( synteny blocks or orthologous genes for example ) . Then it decomposes into two main steps: we first compute a collection of possible ancestral syntenic groups ( in general small groups of genomic markers that were possibly contiguous in the ancestral genome ) , each weighted according to its conservation in the extant species; from this set of possible ancestral syntenies , we group and order the considered genomic markers into one ( or several alternative ) set ( s ) of CARs , each of these sets of CARs representing a possible ancestral genome architecture . An important feature of our framework is that we propose the set of all possible genome architectures that agree with the conserved ancestral syntenies . This framework is general in the sense that both steps can be made effective in several ways . For example , during the first phase , the signal for ancestral syntenies can be defined from extant species in terms of conserved adjacencies between homologous markers as in [12] or between chromosome segments as in [5]–[7] . We propose one possible implementation of this framework , choosing as ancestral features both conserved adjacencies and gene teams [23] , [24] , generalizing the approach of Ma et al . [12] ( where only adjacencies were considered ) , and mimicking the methods employed with cytogenetic data [5]–[7] ( conserved chromosome segments may be formalized as gene teams ) . The second step , that computes CARs and ancestral genome architectures , benefits from a combinatorial framework , centered around the Consecutive Ones Problem and an ubiquitous combinatorial structure called PQ-tree [25] , well known and used in physical mapping [26] , [27] , and recently applied in other comparative genomics problems [28] , [29]; in particular , in [22] , [30] , [31] , PQ-trees were already considered to represent ancestral genomes . In our implementation of this second step , we follow the same principle as in [12]: we extract a maximum unambiguous subset of ancestral syntenies . We apply our method on several datasets . We first consider the case of the ancestral boreoeutherian genome using a dataset obtained from the whole genome alignments available on the UCSC Genome Bioinformatics website [32]; from these alignments , we build sets of synteny blocks at different levels of resolution ( we use from 322 to 1675 homologous markers ) . Our experiments show that the results of our method are quite constant , in the sense that they are very similar , independently of the chosen resolution . This reinforces the impression that algorithmic aspects may have an important impact on the differences in the results of [11] , [12] discussed in [16]–[18] , together with the differences of data acquisition and interdisciplinarity problems [18] . Moreover , the results we obtain are very close to the ones towards which cytogenetics methods tend to converge . As these are obtained from many more species and much expertise , we take it as a validation of the framework and method we propose . We performed intensive comparisons with other computational methods , and ran our method on several published datasets . Compared to the recently published method of Ma et al . [12] , we obtain sets of CARs that are less well defined , as we propose a large set of possible ancestral boreoeutherian genome architectures , instead of only one , but better supported , as any proposed adjacency or segment is supported by at least one syntenic group that is conserved in at least two extant species whose evolutionary path in a phylogenetic tree contains the wished ancestral species . We also reconstruct an ancestral ferungulate genome architecture for the the same data as [11] . On this dataset , our method and the method of Ma et al . obtain similar results . The CARs are comparable to those of the ferungulate chromosomes from e-painting studies [33] that are ancestral boreoeutherian features , while the rearrangement-based method of [11] on the same dataset gives divergent results . In the next section , we describe the general framework and how we implemented it to design a new method for ancestral genome reconstruction . We then describe the results of our method on the considered mammalian datasets . We use our reconstruction of possible genome architectures for the boreoeutherian ancestor at several levels of resolution to assess both the internal stability of our method and the consistency of its results when compared to other published ancestral genome architectures . We compare our results to the results proposed by cytogenetic methods and by the bioinformatics method of Ma et al . [12] , that received some attention recently [18] as it was the first bioinformatics method that tended to agree well with cytogenetics . We conclude by a discussion on our results and methodology and describe several possible extensions of our framework .
We now describe more precisely the two steps of the framework , together with their implementation into an effective method for reconstructing a set of CARs . We separate the general principles from the implementation details to emphasize that there are many possible implementations: the method of Ma et al . [12] is one possibility , and we also propose a variant of our method targeted at analyzing datasets with less well defined outgroups . In this section , we first report the results of our method in reconstructing the architecture of the boreoeutherian ancestral genome from five datasets , at different levels of resolution , that we computed from whole genome alignments . Next we report results based on the original dataset used in [12] and on the ferungulate ancestral genome from the dataset of [11] . All data and results discussed in this section are available on a companion website: http://lbbe-dmz . univ-lyon1 . fr/tannier/ploscb2008_supmat/ . We computed five datasets , with parameters max_gap = 100 kb and min_len = 100 kb ( 1675 markers ) , 200 kb ( 824 markers ) , 300 kb ( 510 markers ) , 400 kb ( 406 markers ) and 500 kb ( 322 markers ) . Their coverage of the human genome goes from 2173 Mb ( min_len = 100 kb ) down to 1487 Mb ( min_len = 500 kb ) . We also analyzed the dataset of 1338 conserved segments used in [12] , downloaded from the website http://www . bx . psu . edu/miller_lab/car/ . It has the impressive property that these conserved segments span slightly more than 94% of the human genome based on alignments at a 50 kb resolution level . On the other hand , it considers less species , an unbalanced phylogeny ( one of the branch from the ancestral node contains a single species , the dog , while the other branch contains three species , human , mouse and rat ) and the segments are less well defined in the outgroups: they can be duplicated ( due to ambiguous orthology signal ) , missing or overlapping . In order to analyze this challenging dataset , we modified our method , to handle the different combinatorial nature of segments in outgroups , and we chose to define ancestral syntenies in terms of conserved adjacencies and approximate common intervals which do not require the exact same markers content and allow for duplicated markers ( see Material and Methods ) . This illustrates the generality of our framework: the way to define ancestral syntenies and the type of dataset is flexible . While we prefer to present the results with our own dataset due to its better proximity to the C1P property , we performed our method on this dataset for the method comparison to be as exhaustive as possible . The set of possible ancestral syntenies contains 2515 subsets of segments , and 208 needed to be discarded in order to clear all ambiguities and get the C1P property . This shows that by relaxing the definition of ancestral synteny by allowing inexact content , we introduced a large number ( at least 10% ) of false positives ( i . e . groups of segments which were not consecutive in the ancestral genome ) . We obtained an ancestral genome with 35 CARs , 1281 adjacencies and the following human chromosomal associations: 3-21 , 4-8 , 12-22 , 12-22 , 14-15 , to compare to 29 CARs and 1309 adjacencies and the same human chromosomal associations in [12] . Among our 1281 adjacencies , 1077 are present in the 1309 adjacencies obtained with the method of Ma et al . . As before , we define a weak adjacency as an adjacency obtained by the method of Ma et al . whose segments are not included in any of our ancestral syntenies: 8 of the 1309 adjacencies obtained in [12] are weak . Among these adjacencies are several human or rodent or dog specific adjacencies . The fact that we have significantly fewer common adjacencies while the adjacencies of Ma et al . are still well supported can be explained by the fact that some adjacencies inferred in [12] are supported by false positive ancestral syntenies , which are much more frequent with this dataset than when using or own datasets of universal markers , where we used several filters to eliminate them . For example , by assessing the support of the adjacencies in the 29 CARs obtained by Ma et al . in terms of the ancestral syntenies conserved after our second phase , which produces a C1P matrix , 21 are not supported , and the general level of support of adjacencies decreases in general . We also tested our framework on the ferungulate ancestor based on the dataset of Murphy et al . [11] . This dataset contains seven genomes , which are represented by 307 synteny blocks that cover 1343 Mb of the human genome [11] . It is hazardous to reconstruct boreoeutherian ancestors with this dataset , because there is no outgroup for the boreoeutherian clade here , but it is interesting to use this dataset to compare several methods on a dataset we did not construct . We ran both our method and the one of Ma et al . [12] on this dataset and compared the inferred genome architectures . We include in the comparison the results obtained by Murphy et al . [11] on the same dataset , and those of Kemkemer et al . [33] obtained independently by a computational method called e-painting , see Table 3 . The ancestral genome architecture we propose is based on 457 ancestral syntenies from an initial number of 461 , and here again the dataset seems to contain very little ambiguity . Some syntenies obtained belong to the boreoeutherian ancestor , and others are ferungulate specific . The synteny between human chromosomes 5 and 19 is inferred only by Murphy et al . ( where it is not marked as weak , which means that it was found in all alternative genome architectures ) but not by our method . However , it is due to an adjacency between two synteny blocks that is not found in any of the ancestral syntenies we detected in the first step of our method , and is found only in the pig genome . The synteny between human chromosomes 1 and 22 is inferred only by Murphy et al . , where it is marked as weak . It is due to an adjacency that is not found in any genome , nor supported by any of our ancestral syntenies . The same holds for the synteny between human chromosomes 2 and 20 ( which is not weak according to Murphy et al . ) , and seems to be more rodent-specific . The synteny between human chromosomes 1 and 10 was inferred by MGR and our method , and considered weak by Murphy et al . , and is supported by three of our ancestral syntenies that have significant weights . The synteny between human chromosomes 2 and 7 , which is found only by the method of Ma et al . is due to an adjacency that is found only in the pig and is not supported by any of our ancestral syntenies . We can also note that among the 250 adjacencies inferred by our method , only 196 are common with the results obtained with the methods of Ma et al . and Murphy et al . , while 240 are common with the ancestor obtained with the method of Ma et al . and 204 are common with the ancestor proposed by Murphy et al . We have only the boreoeutherian syntenies in common with Kemkemer et al . [33] , and those that are supposed to be ferungulate specific all disagree ( we don't recover the giant chromosome 1-19-3-21 , and recover 1-10 instead ) .
We perform here a comparative analysis of different methods for the reconstruction of ancestral genomes , independently of the type of data used for these reconstructions . For the boreoeutherian ancestor , Ma et al . [12] , with their own set of markers called conserved segments , recovered 29 CARs , with 8 “weak adjacencies” . Those adjacencies correspond to features that are only present in human and mouse for example , which would more account for an euarchontoglire feature , or even only in human ( as the junction of both parts of human chromosomes 10 or 16 for example ) . In contrast , at a resolution of 200 kb and with universal synteny blocks , we infer 26 CARs , which is comparable , but no such weakly supported adjacency is inferred . At the resolution of 50 kb , with Ma et al . data , we infer 35 CARs , which compares to 29 CARs plus 8 weak adjacencies . Moreover , all our chromosomal syntenies , at several resolution levels , are also supported by cytogenetic studies , but the fusion of a synteny block of human chromosome 4 with a segment of human chromosome 1 that is found only at high resolution ( min_len = 100 kb ) . The method of Ma et al . gives 31 to 37 CARs on our datasets , with a significant number of weak adjacencies , as well as some variations in terms of human chromosomal associations . The most likely explanation for the difference between the two methods lies in methodological reasons , primarily the way ancestral syntenies are defined ( adjacencies computed through a Fitch-like approach in [12] , see below for a discussion on that topic ) , rather than to the dataset itself as the way we compute synteny blocks are very similar , even if we conserve only blocks that are present in all genomes . Nevertheless , the results obtained both by our method and Ma et al . method , which both rely on model-free algorithmic principles , like cytogenetics methods but on other kind of data , strongly agree with cytogenetics results . We also tested our method on the ferungulate ancestor and compared our results with the ancestor inferred through a rearrangement-based method in Murphy et al . [11] . With the method Murphy et al . , based on a genome rearrangement model and MGR [8] , the results diverged from the cytogenetics data and provoked the discussion in [16]–[18] . Using the same synteny blocks as Murphy et al . , we found 24 CARs , all of which are chromosomes of the boreoeutherian ancestor , except a fusion of the homologs of human chromosomes 1 and 10 , which seem to be ferungulate-specific , and was also inferred by MGR . None of the other chromosomal syntenies proposed by [11] were recovered by our method , or the Ma et al . method . However , the number of common inferred ancestral adjacencies points out that our method and the method of Ma et al . compute similar ancestral genome architectures , which are different from the one proposed by MGR , despite the fact that this last one has 24 CARs , as with our method . We believe that this three-way comparison indicates that the differences discussed in [16] , [17] are partly due to the methods themselves , and more precisely to the fact that MGR is a rearrangement-based method , whereas all the others are model-free . We now summarize the main methodological features of the framework we propose , and discuss them , as well as some possible extensions . We propose to decompose the process of ancestral genome architecture inference into three steps: detection and weighting of ancestral syntenies , representation as a 0/1 matrix and a generalized PQ-tree , clearing ambiguities and representation of a set of alternative genome architectures as a PQ-tree . Although these three steps are performed independently , the implementation choices for each of them can have important consequences on the other ones , as we discuss below . We implemented this method using ( 1 ) unique and universal synteny blocks , which appear once in each genome , ( 2 ) ancestral syntenies defined as unambiguous adjacencies and maximal common intervals ( or gene teams ) which are present in at least two genomes whose evolutionary path along their phylogeny meets the considered ancestral species and ( 3 ) a combinatorial optimization approach , based on the Consecutive Ones Submatrix Problem , to clear ambiguities . The comparison of our method and the one of Ma et al . [12] through the prism of this framework highlights the important effects of some methodological choices on ancestral genome proposals . We discuss below these choices on the combinatorial nature of the considered sets of genomic markers , the definition and computation of ancestral syntenies , and the method to clear ambiguities .
We construct several datasets , by a unique method depending on two parameters , max_gap and min_len . This method , or very similar ones , are often used to construct synteny blocks from genomic alignments [40] , [41] , [59] . We first use the notion of “teams of markers” [24] . This notion relies on a parameter δ , a positive integer . In a genome , the position of a marker m , denoted by p ( m ) , is its relative rank on the its chromosome . That is , the first marker on a chromosome has rank 1 , the second has rank 2 , and so on . Two markers m1 and m2 are said to be close to each other in a genome , for the parameter δ , if they lie on the same chromosome , and |p ( m1 ) −p ( m2 ) |≤δ . A subset of markers M is said to be a team for a genome if for any two markers a , b from M , there exists a sequence S = a , a1 , …ak , b of markers from M , such that any two consecutive markers in S are close to each other . Given two genomes X and Y , a team S common to X and Y is a set of markers labels ( a subset of Σ the alphabet of markers ) that is a team in both genomes X and Y . Such a team S is maximal if no other team is common to X and Y and contains S . Maximal common intervals are maximal common teams for δ = 1 . Maximal common teams can be computed efficiently thanks to an algorithm by Beal et al . [23] and a software described in [24] . We collect a set of teams , representing possible ancestral syntenies , by computing all maximal common teams of pairs of species which evolutionary path contains the wished ancestor . In order to analyze the dataset of [12] , due to less defined markers in the two outgroup genomes , we used maximal approximate common intervals defined as follows: a subset M of markers is an approximate common interval between two genomes if there exists a genome segment in each of the two genomes whose 80% of the gene content is equal to S . An approximate common interval S is maximal if no other approximate common intervals is common to X and Y and contains the two occurrences of S in X and Y . As teams rely only on similarity in markers content , and do not involve any marker order constraints , we added to this set of ancestral syntenies the set of putative ancestral adjacencies , defined as pairs of markers that are consecutive in at least two genomes whose evolutionary path contains this ancestor and do not belong to a conflict . A conflict is defined as follows ( Figure 7 in [12] ) : an adjacency {i , j} belongs to a conflict if , in the graph G whose vertices are the markers ( V ( G ) = Σ ) and the edges are the conserved adjacencies , either i or j has degree more than 2 , or the edge {i , j} belongs to a cycle . Each of these ancestral syntenies was weighted following the same principle as in [12] . Let S be a subset of Σ that represents a possible ancestral synteny . In any leaf X of the species tree , if S is a team in X , the weight of S in X is wX ( S ) = 1 , otherwise , wX ( S ) = 0 . Then , in any internal node N of T ( other than the ancestral node A ) having two children R and L , wN ( S ) is defined recursively by the formulawhere dL and dR are respectively the length of the branch between N and L and N and R . The weight of S in A is then defined bywhere A1 , A2 and A3 are the three neighbors of the ancestral node A in T , and , and are the respective length of the branch between A and A1 , A and A2 and A and A3 . Recall is the set of homologous markers , is the set of subsets of that represent possible ancestral syntenies and the corresponding 0/1 matrix . We say that two elements Si and Sj of overlap if their intersection is not empty , but none is included in the other . Let be the family of all subsets of that do not overlap with any member of ; in other words , given X an element of , any Si of either contains all elements of X or contains no element of X . Among the subsets of , call strong the elements that do not overlap any other elements of . The inclusion tree of the strong elements of , denoted , is a tree where each strong element of corresponds to a single node and the node corresponding to a strong subset X is an ancestor of the node corresponding to a strong subset Y if and only if X contains Y as a subset . Given a node N of , we associate to it the subset of the elements of defined as all Si's that are included in N but in none of its children . The PQ-tree is defined from as follows: an internal node N such that s ( N ) = Ø is a P-node , while an internal node N such that s ( N ) ≠Ø is a Q-node if s ( N ) can be partitioned by a partition refinement process [68] and a R-node otherwise . The construction of can be achieved in optimal O ( n+m ) time where and , as described in [45] . In the last step , we want to remove the minimal amount ( in terms of weight ) of ancestral syntenies from in order that the resulting matrix is C1P . This problem , which is known as the Consecutive Ones Submatrix Problem generalizes the Minimum Path Partition ( or Path Cover ) problem used in [12] and is known to be NP-hard [69] even for sparse matrices [70] , which is the case of the matrices we obtain . However , using the structural information given by the PQ-tree , it is possible to design an efficient branch-and-bound algorithm . More precisely , it follows immediately from the definition of that ambiguous information that prevent a matrix to be C1P can only be located in the submatrices defined by the subsets s ( N ) of for the degenerate nodes of . Hence each of these subsets of can be processed independently of the remaining of . For such a subset , say , we first compute an upper bound on the maximum subset S of s ( N ) that defines a matrix that is C1P , using the same approach than in [12]: start with S = Ø and , for each element of s ( N ) , taken in decreasing order of weight , if adding to S defines a matrix that is not C1P ( which can be tested using the efficient algorithms described in [46] , [68] ) , then discard it , else leave it in S . From that upper bound , using the same principle , we use a classical branch-and-bound algorithm that looks for a better subset of s ( N ) that defines a C1P matrix . Let an adjacency in an ancestral genome architecture be defined by two markers X and Y that are adjacent in a CAR of this ancestral genome , for a given resolution ( say 100 kb ) . We say that it is conserved at a lower resolution ( say 200 kb ) if either the synteny blocks corresponding to X and Y in the human genome are both included in a single synteny block in the human genome at the lower resolution or if X and Y are contained in two blocks X′ and Y′ at the lower resolution level whose corresponding markers are adjacent in the ancestral genome inferred at this resolution . The adjacency is weakly conserved if the markers X′ and Y′ are not adjacent but present on the same CAR ( weakly conserved adjacencies point at local rearrangements resulting from changing the resolution of the considered data ) . Otherwise , if the two markers X′ and Y′ are not on the same CAR , we say that the adjacency between X and Y is not conserved . Note that we do not consider this adjacency is not conserved if at least one of the two synteny blocks corresponding to X or Y is not included in a lower resolution synteny block .
Part of this work was done while CC visited the LRI ( Université Paris-Sud , Orsay , France ) and LaBRI ( Université Bordeaux I , Talence , France ) . | No DNA molecule is preserved after a few hundred thousand years , so inferring the DNA sequence organization of ancient living organisms beyond several million years can only be achieved by computational estimations , using the similarities and differences between chromosomes of extant species . This is the scope of “paleogenomics” , and it can help to better understand how genomes have evolved until today . We propose here a computational framework to estimate contiguous segments of ancestral chromosomes , based on techniques of physical mapping that are used to infer chromosome maps of extant species when their genome is not sequenced . This framework is not guided by possible evolutionary events such as rearrangements but only proposes ancestral genome architectures . We developed a method following this framework and applied it to mammalian genomes . We inferred ancestral chromosomal regions that are stable and well supported at different levels of resolution . These ancestral chromosomal regions agree with previous cytogenetics studies and were very probably part of the genome of the common ancestor of humans , macaca , mice , dogs , and cows , living 120 million years ago . We illustrate , through comparison with other bioinformatics methods , the importance of a formal methodological background when comparing ancestral genome architecture proposals obtained from different methods . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods",
"Acknowledgments"
]
| [
"computational",
"biology/evolutionary",
"modeling",
"evolutionary",
"biology/bioinformatics",
"computational",
"biology/genomics"
]
| 2008 | A Methodological Framework for the Reconstruction of Contiguous Regions of Ancestral Genomes and Its Application to Mammalian Genomes |
There are various factors which construct the perception of stigma in both leprosy affected persons and unaffected persons . The main purpose of this study was to determine the level of perceived stigma and the risk factors contributing to it among leprosy affected person attending the Green Pastures Hospital , Pokhara municipality of western Nepal . A cross-sectional study was conducted among 135 people affected by leprosy at Green Pastures Hospital and Rehabilitation Centre . Persons above the age of 18 were interviewed using a set of questionnaire form and Explanatory Model Interview Catalogue ( EMIC ) . In addition , two sets of focused group discussions each containing 10 participants from the ward were conducted with the objectives of answering the frequently affected EMIC items . Among 135 leprosy affected persons , the median score of perceived stigma was 10 while it ranged from 0–34 . Higher perceived stigma score was found in illiterate persons ( p = 0 . 008 ) , participants whose incomes were self-described as inadequate ( p = 0 . 014 ) and who had changed their occupation due to leprosy ( p = 0 . 018 ) . Patients who lacked information on leprosy ( p = 0 . 025 ) , knowledge about the causes ( p = 0 . 02 ) and transmission of leprosy ( p = 0 . 046 ) and those who had perception that leprosy is a severe disease ( p<0 . 001 ) and is difficult to treat ( p<0 . 001 ) had higher perceived stigma score . Participants with disfigurement or deformities ( p = 0 . 014 ) , ulcers ( p = 0 . 022 ) and odorous ulcers ( p = 0 . 043 ) had higher perceived stigma score . The factors associated with higher stigma were illiteracy , perceived economical inadequacy , change of occupation due to leprosy , lack of knowledge about leprosy , perception of leprosy as a severe disease and difficult to treat . Similarly , visible deformities and ulcers were associated with higher stigma . There is an urgent need of stigma reduction strategies focused on health education and health awareness programs in addition to the necessary rehabilitation support .
Leprosy is a chronic granulomatous disease caused by Mycobacterium leprae . Besides clinical sequel followed usually after infection , the consequences of stigma associated with leprosy outweigh the burden of physical afflictions [1] . Three kinds of stigma associated with leprosy affected persons have been described . Experienced or enacted stigma refers to the real discrimination or acts experienced by leprosy affected persons while perceived stigma refers to the development of fear within an affected person where the fear may arise out of potential discrimination from family members , friends or society . As a consequence of both enacted and perceived stigma , a person over a long period of time may believe what others think and say about him , resulting to the loss of self-esteem and dignity which is referred to be a self-stigma or internalized stigma [2] . Stigma affects the psychosocial well-being of the affected person . A person may feel fear or shame which can lead to anxiety and depression . The resultant anxiety and depression may lead to decreased social participation and social exclusion [3] . Anticipation of stigma may cause affected person to conceal their condition [4] . The burden of keeping this secret , of being ever watchful and careful takes an emotional toll and adversely affects health seeking behavior [3] . Concealing the disease , avoiding the questions regarding the disease and at times even telling lie for the fear of disclosure was found to be a major concern for leprosy affected persons attending Green Pastures Hospital , Nepal [5] . Stigma has been found to be associated with misconceptions about the disease , visible deformities and the development of ulcers [4] . Disability is a broad term covering any impairment , activity limitation or participation restriction affecting a person . According to WHO , grade 0 means no disability is found . Grade I means that loss of sensation has been noted in the hand or foot while grade II means the visible damage or disability is noted [6] . Visible deformities and disabilities have been found to be the prominent contributor of stigma development in leprosy affected persons [7] while it also triggers the development of negative attitudes towards leprosy among unaffected people [8] . In a systematic review of risk factors contributing to stigma , the basis of stigma development was found to be the visibility of the disfigurements and disability augmented by the stereotypes of the society , knowledge and the status of the person in terms of economy , education and ability to participate in society [9] . In Nepal , leprosy is still a stigmatizing disease . Misconceptions about the disease have contributed to the development of negative attitudes to leprosy affected persons . In a study conducted in eastern Nepal , fear of infection and god's curse were found to be the most prevalent causes of negative behavior towards leprosy affected persons [8] . In the other study [10] conducted in eastern part of Nepal , the causes of stigma perception in leprosy affected persons were consistent with the causes of negative attitudes in unaffected community members [8] . The beliefs and perceptions about leprosy were found to be the prominent causes of stigma [10] . Fear of infection , was the most important cause of stigma different countries including China [11] and India [12] . In India , in addition to the fear of infection , false beliefs about leprosy , ignorance about the disease and lower socio-economic status were associated with stigma in leprosy [12] . Therefore , we hypothesized that there is association between the levels of perceived stigma in leprosy affected persons and the factors characterizing them ( demographic characteristics , knowledge about leprosy , natural history of disease , clinical presentation , disability grades and reaction ) While few studies are done in eastern part of Nepal , most of them are focused on the impact of the stigma , participation restriction and income generation . There has been no research so far in leprosy stigma in a view to explore the factors associated with it . The specific objective of this study was to determine the prevalence of perceived stigma and its association with factors such as socio-demographic , knowledge about leprosy and clinical presentation characterizing leprosy affected persons attending Green Pastures Hospital and Rehabilitation Centre . Green Pastures Hospital and Rehabilitation Centre , the only known leprosy referral center in western region of Nepal provides the services for leprosy patients with disability management , treatment and vocational training . Therefore , exploring the risk factors of stigma in leprosy affected persons attending GPH&RC can help to understand the leprosy stigma and therefore can direct the stigma reduction strategies and intervention programs .
Ethical permission for this research was obtained from Nepal Health Research Council and International Nepal Fellowship Research Committee . People were eligible if they were affected by leprosy , age above 18 years and willing to participate . Interviews were only conducted after the written consent was received and was conducted by principal investigator . Interviews were conducted with all leprosy affected people attending GPH&RC from February 2013 to March 2013 . Attempt was done to include equal number of participants from the ward and OPD , 5 from the ward and 3 from the OPD denied the written consent , however , there were no drop outs . The interviewer taking into the consideration the sensitivity of the subject established a friendly rapport before the interview and encouraged participants to express their views . The anonymity of the participants was secured by coding the participants' name . No incentives were offered or paid for their time .
All participants who met the eligibility criteria were recruited into the study after taking written consent . Total 135 participants were asked with the questionnaire form . Among the questions representing different aspects of perceived stigma in EMIC questionnaire , most affected areas of perceived stigma were concealment of the disease , self-esteem , disclosure concern and the shame and embarrassment due to leprosy ( Table 1 ) . Among the total participants 65 . 9% affirmed that they would conceal the disease condition as long as it is possible while 57 . 8% anticipated decreased self-esteem due to the disease condition and 40 . 7% only disclosed the disease condition to the close ones . Of the 135 leprosy affected participants , 58 . 5% of them were those who attended OPD at the hospital . Total median score of EMIC scale was higher among those leprosy patients who were in the ward compared to those who attended OPD ( p = 0 . 006 ) . There was no significant difference in mean EMIC score between different age groups ( p = 0 . 199 ) , sex ( p = 0 . 344 ) , ethnicity ( p = 0 . 934 ) , location ( p = 0 . 072 ) , marital status ( p = 0 . 477 ) and family type ( p = 0 . 356 ) . Similarly , participants were asked if they had any other member of their family affected by leprosy in past or present including if they had relatives or neighbors affected by leprosy . Neither of them had significant difference in median score of stigma ( Table 2 ) . There was a significant difference in median EMIC score ( p = 0 . 008 ) between different level of education in participants classified as illiterate ( those who could not read and write ) , those who attended primary level ( <5 years of education ) and those who attended secondary and higher education ( >5 years ) . On post hoc analysis , the illiterate and those who attended more than 5years of education had significant difference in median score ( p = 0 . 03 ) . Similarly , when EMIC scores among subjects with less than 5 years education were compared with those with more than 5 years there was a significant difference ( p = 0 . 016 ) while EMIC scores of the illiterate and those who attended <5 years of education were not significantly different ( p = 0 . 673 ) . There was no significant difference in median score between religious groups Hindu and other ( p = 0 . 309 ) , Occupation ( p = 0 . 321 ) , and amount of income ( p = 0 . 068 ) . However , on post hoc analysis two different income groups ( the highest and lowest income group ) showed significant difference ( p = 0 . 011 ) There was a significant difference in EMIC score between those who felt economic inadequacy and who did not ( p = 0 . 014 ) . Similarly , there was also significant difference in stigma score between those who had to change their occupation after being affected by leprosy and those who did not ( p = 0 . 018 ) . Knowledge and perceptions about leprosy and perceived stigma scores were analyzed in all participants . The overall stigma score for those who had knowledge about leprosy was lower than those who lacked knowledge of leprosy ( Table 3 ) . There was a significant difference in EMIC stigma score between those who had information on leprosy ( p = 0 . 025 ) , knowledge on leprosy cause ( p = 0 . 02 ) and knowledge on transmission ( p = 0 . 046 ) . Similarly , participants who did not have knowledge of leprosy signs and symptoms had lower stigma scores compared to those who knew one or more signs and symptoms of leprosy although this was statistically insignificant ( p = 0 . 344 ) . There was a difference in EMIC stigma score who perceived leprosy as a very infectious disease ( p = 0 . 127 ) . Similarly , there was a significant difference in perceived stigma score between groups who felt that leprosy is difficult to treat ( p<0 . 001 ) and a severe disease ( p<0 . 001 ) . Brief history of disease and clinical presentations were asked and assessed respectively with all the participants ( Table 4 ) . Participants' age at diagnosis ( p = 0 . 213 ) and years after diagnosis ( p = 0 . 967 ) did not show any difference in EMIC score . First sign and symptoms were categorized into skin involvement , nerve involvement and deformity development . Neither of them showed significant difference in perceived stigma score ( p = 0 . 792 ) . Similarly , there was no significant difference in EMIC between participants who sought hospital or doctor soon after development of signs and symptoms and who did not ( p = 0 . 079 ) . The majority ( 55 . 6% ) of participants received first treatment from non-medical providers such as witch doctors and traditional healers . There was no significant difference in EMIC score between groups of participants who received treatment from medical providers , non-medical providers and friends/family and others ( p = 0 . 255 ) . Similarly , there was no significant difference in EMIC score between those who completed treatment and who did not ( p = 0 . 156 ) . There was a significant difference in EMIC score in participants who had disfigurement or deformities ( p = 0 . 014 ) , ulcer ( 0 . 022 ) and odorous ulcer ( 0 . 043 ) compared to those who did not . However , there was no significant difference in EMIC between those who had reaction and who did not ( p = 0 . 331 ) . More than half ( 51 . 1% ) of the participants had grade II disabilities and higher EMIC stigma score compared to grade 0 and grade I disabilities ( p = 0 . 161 ) ( Table 5 ) . However , the difference in EMIC stigma score showed marginal significance between grade II and grade 0 combined with grade I ( p = 0 . 056 ) , not shown in table . In majorities of the leprosy affected persons as evident from EMIC profile , concealment of the disease , lowered self-esteem and the disclosure to the close ones were major aspects of the EMIC questionnaire which contributed to higher EMIC score compared to the marital problems , social exclusion acts and impacts to their family members . Focus group discussion with leprosy affected persons concluded that the discrimination and stigma attached to the disease was felt to be decreasing over the time . However , the reasons for most of the participants' intention not to disclose their disease condition were the fear of discrimination , isolation and rejection . The most often reported cause of fear was the strongly rooted stereotype attached to the disease . The most common belief leprosy affected person presumed was the fear of transmission of the disease among others . In addition to the prevalent false beliefs about the transmission , severity and myths attached with the disease , the deformities and ulcers were also reported to be the triggering factor for the disease disclosure . While most of the participants realized that ulcers and disabilities due to leprosy were affecting them physically , its psychosocial burden was the greater problem . Some patients never reported to their close ones about their causes of disabilities and ulcers . Instead they often told the causes of disabilities and ulcers to be due to some other disease . However , participants realized that keeping this secrecy was a huge burden for them .
This study was conducted in western region of Nepal , where only those people who visited hospital for treatment , rehabilitation and wound care were recruited while many other people affected by leprosy who did not have any symptoms were not included in the study which limits our finding to generalize over all leprosy affected persons . Only perceived stigma was assessed in this study while two other types of stigma were not assessed therefore , stigma in this study cannot be the whole picture of stigma . While clinical presentations of the participants were obtained from the hospital treatment card , many other questions might have encountered recall biases . The full evaluation of the data using multiple regressions was not done in this study which could have strengthened our findings . This study concludes that lower education level , perceived economic inadequacy , obligation to change the occupation due to leprosy , lack of knowledge and the wrong perceptions about leprosy were the significant factors contributing to higher levels of perceived stigma in leprosy affected persons . In addition to these socio-demographic factors , the presence of visible deformities , ulcers and disabilities also contributed to higher perceived stigma in leprosy affected persons . The major aspects of EMC stigma scale affected were the attitude to conceal the disease , and lowered self-esteem . The major causes for these have been explained by focus group discussion as the perceived fear of discrimination , rejection and the society's fear of transmission . The factors contributing to the development of stigma in leprosy affected persons from this study can direct the need of intervention programs focusing on health education . Health education which might correct the wrong perceptions and might increase understanding of leprosy and the people affected can have a significant impact in both leprosy affected persons and leprosy unaffected persons . In addition to the education and health awareness programs , empowerment of the leprosy affected persons by technical education , vocational training and social participation might be helpful to increase self-esteem and reduce perceived stigma . Ulcers and visible deformities have been found as contributing factors for the higher level of perceived stigma . Early case detection through training of health professionals and health education to the general public might prevent the delays in presentation , ulcers , and deformities which ultimately can reduce the stigma . | A total of 135 leprosy affected persons were interviewed with a questionnaire containing EMIC questions designed to assess the level of perceived stigma and the questionnaire containing variables for socio-demographic characteristics , knowledge about leprosy and the clinical presentations of the participants . Clinical presentation as disability was graded according to WHO guidelines , where grade 0 means no disability found , grade I means loss of sensation has been noted in the hand or foot while grade II means visible damage or disability . Total EMIC score was analyzed between sub-variables to see the factors associated with the higher level of perceived stigma score . Additionally , among the total participants , we included 20 of them who were admitted at hospital for various reasons . Two sets of focus group discussions were conducted with additional questions to derive the reasons behind frequently affected EMIC stigma domains . The factors associated with higher perceived stigma score were illiteracy ( those who could not read and write ) , perceived economical inadequacy , lack of knowledge on leprosy , the perceptions as difficult to treat and severe disease and presence of visible deformities and ulcers . Considering our findings pertaining to higher perceived stigma , there is an urgent need of stigma reduction strategies which should focus on health education about leprosy that can change the perceived stigma in leprosy . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"and",
"health",
"sciences",
"social",
"sciences"
]
| 2014 | Factors Affecting Perceived Stigma in Leprosy Affected Persons in Western Nepal |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.